WO2019128933A1 - 一种地图构建、导航方法及装置、系统 - Google Patents

一种地图构建、导航方法及装置、系统 Download PDF

Info

Publication number
WO2019128933A1
WO2019128933A1 PCT/CN2018/123164 CN2018123164W WO2019128933A1 WO 2019128933 A1 WO2019128933 A1 WO 2019128933A1 CN 2018123164 W CN2018123164 W CN 2018123164W WO 2019128933 A1 WO2019128933 A1 WO 2019128933A1
Authority
WO
WIPO (PCT)
Prior art keywords
obstacle
map
type
navigation
obstacles
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2018/123164
Other languages
English (en)
French (fr)
Inventor
李从兵
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
ZTE Corp
Original Assignee
ZTE Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by ZTE Corp filed Critical ZTE Corp
Priority to EP18893502.7A priority Critical patent/EP3734225A4/en
Priority to US16/958,855 priority patent/US11359930B2/en
Publication of WO2019128933A1 publication Critical patent/WO2019128933A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/38—Electronic maps specially adapted for navigation; Updating thereof
    • G01C21/3863—Structures of map data
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/20—Instruments for performing navigational calculations
    • G01C21/206—Instruments for performing navigational calculations specially adapted for indoor navigation
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/38—Electronic maps specially adapted for navigation; Updating thereof
    • G01C21/3804—Creation or updating of map data
    • G01C21/3807—Creation or updating of map data characterised by the type of data
    • G01C21/383—Indoor data
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/38—Electronic maps specially adapted for navigation; Updating thereof
    • G01C21/3804—Creation or updating of map data
    • G01C21/3833—Creation or updating of map data characterised by the source of data
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/38—Electronic maps specially adapted for navigation; Updating thereof
    • G01C21/3863—Structures of map data
    • G01C21/387—Organisation of map data, e.g. version management or database structures
    • G01C21/3878—Hierarchical structures, e.g. layering
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/0088—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots characterized by the autonomous decision making process, e.g. artificial intelligence, predefined behaviours
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/02—Control of position or course in two dimensions
    • G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
    • G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
    • G05D1/0214—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory in accordance with safety or protection criteria, e.g. avoiding hazardous areas
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/20—Control system inputs
    • G05D1/24—Arrangements for determining position or orientation
    • G05D1/246—Arrangements for determining position or orientation using environment maps, e.g. simultaneous localisation and mapping [SLAM]
    • G05D1/2464—Arrangements for determining position or orientation using environment maps, e.g. simultaneous localisation and mapping [SLAM] using an occupancy grid
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/60—Intended control result
    • G05D1/617—Safety or protection, e.g. defining protection zones around obstacles or avoiding hazards
    • G05D1/639—Resolving or avoiding being stuck or obstructed
    • G05D1/642—Resolving or avoiding being stuck or obstructed involving obstacle removal, e.g. opening doors or pushing furniture
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D2105/00—Specific applications of the controlled vehicles
    • G05D2105/80—Specific applications of the controlled vehicles for information gathering, e.g. for academic research
    • G05D2105/87—Specific applications of the controlled vehicles for information gathering, e.g. for academic research for exploration, e.g. mapping of an area
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D2107/00—Specific environments of the controlled vehicles
    • G05D2107/40—Indoor domestic environment

Definitions

  • the present application relates to, but is not limited to, the field of data processing, for example, to a map construction, navigation method, apparatus, and system.
  • a complete two-dimensional grid is used to abstract the entire indoor environment information.
  • the grids on this raster map are divided into three categories, obstacle grids, non-obstacle grids, and unknown grids.
  • the obstacle grid represents an obstacle in the corresponding position of the grid
  • the non-obstacle grid indicates that the corresponding position of the grid is free
  • the unknown grid indicates that the information on the corresponding position of the grid is unknown.
  • the traditional grid map construction method tightly couples all the environmental information on the same grid map.
  • the map of the secondary navigation is limited to the current use, and the reusability is very poor.
  • the grid map does not consider the impact of the grid where the obstacles are located on the nearby grid traffic. Since the mobile robot itself occupies a certain space, the grid of the obstacles nearby, the robot can not pass unrestricted. Otherwise, it is very likely that it will collide with an obstacle as it passes through these grids.
  • an improved solution of the related art is to introduce a hierarchical management and update mechanism of the grid map, and divide the grid map from bottom to top into a static map layer, an obstacle map layer, a sonar obstacle map layer, and Expand the map layer and so on.
  • the static map layer manages the mobile robot to walk without indoor dynamic obstacles, and is a grid representation of the overall environment constructed based on the information detected by the sensors.
  • the obstacle map layer is a grid representation of the dynamic obstacle detected by the laser sensor during the execution of the navigation task by the mobile robot.
  • the sonar obstacle map layer is a grid representation of the dynamic obstacle detected by the sonar sensor during the execution of the navigation task by the mobile robot.
  • the expanded map layer does not use external sensor data, but is a fulcrum for the obstacle grid in the static map layer, the obstacle map layer, and the sonar map layer, and the obstacle radius is measured by a given expansion radius.
  • the non-obstacle grid around the grid is expanded.
  • a non-obstacle grid within the range of the expansion radius forms an expanded map layer.
  • a non-obstacle grid on the expanded map layer, the robot can pass by ensuring that its center point does not coincide with these non-obstacle grids.
  • this method of constructing a map still has deficiencies in at least one of performance, such as utility, flexibility, security, and resource utilization.
  • the embodiment of the present application provides a map construction method, including: identifying a detected obstacle, and determining, according to the recognition result, a type of the obstacle in a plurality of obstacle types classified according to the obstacle characteristic; Map and mark the obstacle and record the type to which the obstacle belongs.
  • the embodiment of the present application further provides a map construction system, including: an obstacle recognition unit, configured to identify the detected obstacle to obtain a recognition result; and the obstacle classification unit is configured to determine the location according to the recognition result.
  • the obstacle belongs to a type belonging to a plurality of obstacle types classified according to the obstacle characteristic; the obstacle processing unit is configured to construct a map and mark the obstacle, and record the type to which the obstacle belongs.
  • the embodiment of the present application further provides a map construction apparatus, including a memory, a processor, and a computer program stored on the memory and operable on the processor, where the processor implements the program
  • a map construction apparatus including a memory, a processor, and a computer program stored on the memory and operable on the processor, where the processor implements the program
  • the map construction method described in the embodiment is applied.
  • the embodiment of the present application further provides a computer storage medium on which a computer program is stored, and the computer program is read by the processor to execute the map construction method described in the embodiment of the present application.
  • the embodiment of the present application further provides a navigation method, including: identifying a newly added obstacle detected on a travel path, and determining, according to the recognition result, the plurality of obstacles classified according to the characteristics of the obstacle.
  • the type to which the obstacle type belongs; the obstacle avoidance processing is performed according to the type to which the newly added obstacle belongs.
  • the embodiment of the present application further provides a navigation system, including a navigation module and a map construction module, wherein: the navigation module is configured to plan a travel path, and notify the map construction module to start a navigation map update task; the map construction module And being configured to identify the newly added obstacles detected on the travel path, and determine, according to the recognition result, a type of the plurality of obstacle types that the new obstacles are classified according to the obstacle characteristics; The navigation module is further configured to receive the updated navigation map, and perform obstacle avoidance processing according to the type of the newly added obstacle.
  • the embodiment of the present application further provides a navigation device, including a memory, a processor, and a computer program stored on the memory and operable on the processor, where the processor implements the program as implemented in the present application The navigation method described in the embodiment.
  • the embodiment of the present application further provides a computer storage medium on which a computer program is stored, and the computer program is read by the processor to execute the navigation method described in the embodiment of the present application.
  • FIG. 1 is a flowchart of a method for constructing a map according to Embodiment 1 of the present application
  • FIG. 2 is a block diagram of a map construction system according to Embodiment 1 of the present application.
  • FIG. 3 is a flowchart of a navigation method according to Embodiment 2 of the present application.
  • FIG. 4 is a block diagram of a navigation system according to Embodiment 2 of the present application.
  • FIG. 5 is a flowchart of a method for constructing a grid map based on indoor obstacle feature classification according to Embodiment 3 of the present application;
  • FIG. 6 is a schematic diagram of an exemplary static map layer according to Embodiment 3 of the present application.
  • FIG. 7 is a schematic diagram of an exemplary stable obstacle map layer according to Embodiment 3 of the present application.
  • FIG. 8 is a schematic diagram of an exemplary negotiated obstacle map layer according to Embodiment 3 of the present application.
  • FIG. 9 is a schematic diagram of an exemplary risk-type obstacle map layer according to Embodiment 3 of the present application.
  • FIG. 10 is a schematic diagram of an exemplary catastrophic obstacle map layer according to Embodiment 3 of the present application.
  • FIG. 11 is a schematic diagram of a main raster map obtained by an exemplary fusion of the third embodiment of the present application.
  • FIG. 12 is a schematic diagram of a static updated map layer after an exemplary update of the third embodiment of the present application.
  • FIG. 13 is a flowchart of a method for negotiating avoidance in a navigation process according to Embodiment 4 of the present application.
  • FIG. 14 is a call flow of a grid map service provided by a mobile robot service to a map construction system according to a fifth embodiment of the present application.
  • the mapping method of the related art is to roughly classify the obstacles according to the detection results of the sensors such as laser and sonar, and does not classify according to the characteristics of the identified obstacles themselves, and the semantics are weak, so that the obstacles are based on the obstacles.
  • the use of flexible composition and obstacle avoidance strategies is difficult to achieve.
  • the same marking strategy is used in the application. This strategy often lacks universality and even introduces waste of unnecessary computing and storage resources. For example, a newly added refrigerator and a family pet were found in front of the sensor. If the same discovery and marking strategy (such as tracking maintenance within 3 meters) is followed, maintenance of the family pet will be on the obstacle map layer. Bring frequent updates. The family pets are agile and erratic, and they also have certain hedging capabilities. The map layer maintained for such obstacles can be limited to a small range, saving resources while not creating safety. Hidden dangers. Conversely, if the newly added refrigerator is maintained in a smaller range, it is not conducive to the global route selection of mobile robot navigation.
  • the embodiment of the present application proposes a new map construction and navigation scheme.
  • This embodiment relates to a map construction method and system.
  • the map construction method of the embodiment is implemented by a map construction system.
  • the map construction system can be implanted in a navigation chip used by the mobile robot, and can be integrated with a visual sensor used by the mobile robot, or can be Implemented on a separate chip. In other applications, it is similar.
  • the map of the present embodiment is exemplified by an indoor grid map, but the present application is not limited thereto, and the manner of obstacle classification can also be extended to an outdoor map, a non-grid map, and the like.
  • the map construction method of this embodiment is as shown in FIG. 1 and includes step 110 and step 120.
  • step 110 the detected obstacle is identified, and the type of the obstacle in the plurality of obstacle types classified according to the obstacle characteristic is determined according to the recognition result.
  • the characteristics of the above obstacles refer to the characteristics of the obstacle itself.
  • it is necessary to identify the obstacles first such as identifying the obstacles as humans, animals, televisions, cabinets, and the like. This is different from the direct classification of the related technology according to the detection result.
  • the related technology detects whether the current obstacle moves within a certain period of time, and divides the obstacle into two categories according to the detection result, which is not an obstacle. It is also impossible to determine whether an obstacle has the ability to move autonomously.
  • the obstacle characteristic includes at least one of an autonomous movement capability, an interaction capability, a safety, and an autonomous risk avoidance capability of the obstacle.
  • the ability of the obstacle to move autonomously means that the obstacle itself can move without external force moving the obstacle.
  • people, animals, toy cars, mobile robots have the ability to move autonomously, while tables and chairs, sofas, and televisions Such furniture and electrical appliances do not have the ability to move autonomously.
  • the interactive ability of obstacles refers to the ability of human-computer interaction, including but not limited to voice interaction, gesture interaction, etc.
  • Adults or another mobile robot supporting interaction can belong to this category, and mobile robots can also be regarded as a special kind of person. ".
  • the safety of an obstacle refers to whether the obstacle is protected, such as an infant, and the user can designate other items as protected objects.
  • the autonomous risk-avoidance ability of an obstacle refers to whether an obstacle can be automatically avoided when it is under stress, such as when the mobile robot approaches.
  • the above-mentioned capabilities of the objects may overlap. For example, obstacles with autonomous risk-avoiding capabilities usually have autonomous mobility, as well as adults with human-computer interaction capabilities.
  • the plurality of obstacle types classified according to the obstacle characteristics include: at least two of a stable type, a negotiated type, a risk type, and a radical type, wherein: the stable obstacle includes at least one type that does not have autonomous movement. Obstacle of ability; the negotiated obstacle includes at least one obstacle capable of autonomous movement and human-computer interaction; the risk-type obstacle includes at least one obstacle as a protected object; the radical obstacle includes at least one Obstacles with autonomous risk aversion.
  • the above-mentioned stable obstacles include, but are not limited to, walls, doors and windows, baby carriages, sofas, tables and chairs, cabinets, televisions and the like that do not have the ability to move autonomously, and the positions of these objects are relatively stable. Once these obstacles are marked in the raster map, this information on the raster map can be reused over a longer period of time.
  • the above-mentioned negotiated obstacles include people with autonomous mobility and voice interaction capabilities other than infants and young children. They have good behavioral and interactive abilities, and can actively adjust their behavior according to external information.
  • the robot's autonomous navigation offers more options and increased flexibility.
  • the above-mentioned risk-type obstacles include the protected objects of infants and young children, and those whose autonomous behavior, voice interaction ability and risk identification ability are relatively weak. Such people are the most preventable and protected objects in the autonomous navigation of mobile robots. The safety of this group of people is more important than the efficiency of mobile robots' autonomous navigation.
  • radical obstacles include, but are not limited to, family pets and electronic devices having strong mobility, such as cats, dogs, and robots. They all have the ability to avoid risks autonomously, with good mobility and randomness of movement.
  • the current image recognition can identify more than 1000 items, and the obstacles common in the home environment can be basically recognized. It is also possible to classify obstacles that are not recognized as stable obstacles, or to classify them as separate.
  • the plurality of obstacle types classified according to obstacle characteristics include: at least two of a stable type, a negotiated type, a risk type, and a radical type, wherein: the stable obstacle includes a building member and At least one of the furniture; the negotiated obstacle includes an adult; the risky obstacle includes an infant; the radical disorder includes a family pet.
  • this other embodiment it is also classified according to obstacle characteristics, but it is not strictly required that a certain type of article has exactly one of at least one characteristic, and it is not necessary to classify all obstacles in the environment.
  • Important obstacles such as the above-mentioned building components, furniture, adults, infants and family pets can be classified.
  • the treatment of at least two of the adult, the infant, the building component, and the family pet according to different strategies indicates that the classification and recording are performed when the map is constructed, that is, the method of the other embodiment is adopted.
  • the above-mentioned "adult” refers to a healthy person who has the ability to move autonomously and interact with human-computer interaction.
  • the classification can be separately specified, especially when the family members are relatively fixed.
  • the names of the above obstacle categories may be changed.
  • the above four types may also be referred to as the first category, the second category, the third category, and the fourth category, or any other can be distinguished as different.
  • the name of the type, this name change does not constitute a difference in technical solutions.
  • the detected obstacle is identified, and the type of the obstacle in the plurality of obstacle types classified according to the obstacle characteristic is determined according to the recognition result, including: performing image information collected by the visual sensor. Obstachy detection, performing image recognition on the image of the detected obstacle, and searching for the obstacle according to the identified obstacle to find the obstacles respectively included in the plurality of obstacle types set in advance, and determining the The type to which the obstacle belongs. For example, when a cat appears in an image collected by a visual sensor, the image of the cat is extracted by obstacle detection, and image recognition is known to be a cat, and obstacles respectively included in various obstacle types set in advance are respectively included. Among them, the cat is a catastrophic obstacle, so it can be determined that the type of the obstacle is a catastrophic type.
  • the above classification can be implemented using a classifier.
  • the record of the obstacle type may be directly or indirectly, such as directly recording the detected type as a type identifier in the attribute information of the obstacle, or simply marking the obstacle on the layer corresponding to the type.
  • the type of obstacle can be determined based on the layer in which the obstacle is located.
  • the present application does not impose any limitations on the manner of recording, and only the recorded information can be used to subsequently determine the type of obstacle.
  • step 120 a map is constructed and the obstacle is marked, and the type to which the obstacle belongs is recorded.
  • the embodiment adopts a grid map, and the constructing the map and marking the obstacle includes: constructing different map layers for different types of obstacles, according to the type to which the obstacle belongs and being associated with the obstacle The obstacle is marked on the map layer corresponding to the type.
  • the map constructed above is a navigation map, and the identification, classification, and marking of the obstacles are to identify, classify, and mark the obstacles added to the obstacles detected by the relative environment map detected during navigation.
  • the present application is not limited thereto, and the map constructed by the method of the embodiment may also be used for other services than navigation, for example, real-time dynamic scene maps are constructed for real-life games, reconnaissance and the like. At this time, there is no distinction between the original obstacles and the newly added obstacles.
  • the navigation map of this embodiment is a navigation map including a static map layer, and the static map layer is marked with an obstacle existing in the environment before the navigation is performed, and the static map layer may be generated according to the imported environment map, wherein the marked obstacle is Stable obstacles; during the navigation process, identify new obstacles detected in the environment, construct a dynamic map layer and mark the newly added obstacles, and record the type of the newly added obstacles; After that, the periphery of the obstacle may be expanded, and the static map layer and the dynamic map layer are merged to obtain an updated navigation main map.
  • the navigation map can include the following map layers.
  • the static map layer is set to maintain the overall pattern of the entire indoor environment before the mobile robot navigates, and the obstacles are stable obstacles.
  • the dynamic map layer includes: a stable obstacle map, which is set to maintain a new stable obstacle that is not included in the static map layer detected by the visual sensor during the navigation process of the mobile robot.
  • the catastrophic obstacle map layer is set to maintain a catastrophic obstacle detected by the visual sensor during navigation of the mobile robot.
  • the negotiated obstacle map layer is set to maintain a negotiated obstacle detected by the visual sensor during the navigation process of the mobile robot.
  • the risk-based obstacle map layer is set to maintain the risk-type obstacles detected by the visual sensors during the navigation process of the mobile robot.
  • the present embodiment adopts at least one of the following strategies depending on the type to which the obstacle belongs.
  • the constructing a map and marking the obstacle includes: constructing a catastrophic type if the type of the obstacle belongs to a catastrophic type
  • An obstacle map layer marks the obstacle, the maintenance area of the radical obstacle map layer being smaller than a maintenance area of other map layers, the other map layer being set to mark other types of obstacles of the plurality of obstacle types .
  • the maintenance area is small and the map constructed is small.
  • the other map layers herein may be at least one of a stable obstacle map layer, a risk type obstacle map layer, and a negotiated obstacle map layer, but in other embodiments, may also be used to mark other types of obstacles.
  • the map layer may be at least one of a stable obstacle map layer, a risk type obstacle map layer, and a negotiated obstacle map layer, but in other embodiments, may also be used to mark other types of obstacles.
  • the map layer may be at least one of a stable obstacle map layer, a risk type obstacle map layer, and a negotiated obstacle map layer, but in other embodiments, may also be used to mark other types of
  • the method further includes: The periphery of the obstacle is subjected to an expansion treatment, such as the type of the obstacle belongs to a risk type, and a first expansion radius is used when performing the expansion treatment, the first expansion radius being greater than other types of the plurality of obstacle types The radius of expansion used when the periphery of the obstacle is subjected to expansion treatment.
  • the expansion process can expand the surrounding obstacles in the map according to the corresponding expansion radius before the layers are merged, and then fuse after processing. It can also be carried out after the fusion, and this application does not limit this.
  • the expansion strategy can also be used in maps that are not layered.
  • the perimeter of the risk-type obstacle grid satisfies the condition of sufficient expansion, that is, centered on the risk-type obstacle grid, with the aggressive expansion radius as the scale, in the risk type.
  • the predetermined range of expansion treatment is performed in the front, rear, left, right, upper left, upper right, lower left, and lower right directions of the obstacle.
  • the risk-type obstacle grid perimeter grid satisfies the sufficient expansion condition
  • the obstacle avoidance condition is satisfied outside the expansion radius, the mobile robot avoids the obstacle through the grid outside the expansion radius and travels to the target point; and when the risk grid obstacle surrounding grid does not satisfy the sufficient expansion condition, or the expansion radius cannot meet the obstacle avoidance
  • the condition indicates that the road is unreachable or the free space around the obstacle is not enough to ensure the safety navigation requirements, and it is necessary to re-plan the new travel route to reach the target point.
  • the method further includes: marking a new mark on the dynamic map layer An added stable obstacle is added to the static map layer.
  • constructing a stable obstacle map layer marks the newly added obstacle
  • the newly added obstacle is another type of obstacle
  • Build another dynamic map layer to mark the newly added obstacle; in this example, after the end of one navigation, the newly added obstacle marked on the stable obstacle map layer can be directly added to the Static map layer.
  • the location of stable obstacles in the environment may change. If this change is not reflected in the static map, it will affect the accuracy of environmental map modeling, which will affect Subsequent autonomous navigation.
  • the static map layer is updated and improved according to the detected stable obstacles, which avoids the repeated processing of the stable obstacle in the next navigation task and improves the efficiency. If the obstacles marked in the static map layer are in the current navigation, the static map layer can also be updated to delete the corresponding obstacles.
  • This embodiment also proposes a map construction system that can implement the above method.
  • the map construction system is embedded in a navigation chip used by the mobile robot, integrated with a visual sensor used by the mobile robot, or implemented by a separate chip.
  • the chip or vision sensor is connected to the robot control system.
  • the map construction system of the present embodiment includes an obstacle recognition unit 10, an obstacle classification unit 20, an obstacle processing unit 30, an obstacle capturing unit 40, a map layer fusion unit 50, and an obstacle updating unit 60.
  • the navigation map construction interface 70, the static map construction interface 80, the map update interface 90, and the map save interface 100 are provided.
  • the obstacle recognition unit 10 is configured to identify the detected obstacle and obtain a recognition result.
  • the obstacle classifying unit 20 is configured to determine, according to the recognition result, a type to which the obstacle belongs to the plurality of obstacle types classified according to the obstacle characteristic.
  • the obstacle processing unit 30 is configured to construct a map and mark the obstacle, and record the type to which the obstacle belongs.
  • the obstacle characteristic includes at least one of an autonomous movement capability, an interaction capability, a safety, and an autonomous risk avoidance capability of the obstacle.
  • the plurality of types classified according to obstacle characteristics include: at least two of a stable type, a negotiated type, a risk type, and a radical type, wherein: the stable obstacle includes at least the building member and the furniture.
  • the stable obstacle includes at least the building member and the furniture.
  • negotiated obstacles include adults
  • risk-type obstacles include infants
  • radical-type obstacles include family pets.
  • the constructed map is an indoor grid map
  • the obstacle processing unit is configured to construct a map and mark the obstacle by: constructing different map layers for different types of obstacles, according to the The type to which the obstacle belongs is marked on the map layer corresponding to the type of the obstacle.
  • the obstacle processing unit may be subdivided into the following subunits: a radical obstacle map layer processing unit, configured to process the radical obstacle, construct a radical obstacle map layer; and a risk obstacle map layer
  • the processing unit is configured to process the risk-type obstacles to construct a risk-type obstacle map layer
  • the negotiation-type obstacle map layer processing unit is configured to process the negotiated obstacles and construct a negotiated obstacle map layer
  • the obstacle map layer processing unit is configured to process the stable obstacle and construct a stable obstacle map layer. Obstacles that do not fall into the above three categories can also be handled by the stable obstacle map layer processing unit.
  • the different map layer processing units described above have some different processing when constructing the corresponding map layer.
  • the migratory obstacle map layer processing unit uses the smallest screening distance when screening obstacles, and the screening distance and the maintenance area size of the map layer. It is equivalent.
  • each map layer processing unit adds different type values to obstacles on this layer.
  • the system further includes: an obstacle capturing unit 40 configured to perform obstacle detection according to image information collected by the visual sensor, that is, to find an obstacle; and the obstacle recognition unit is configured to detect by using the following operation Identifying an obstacle: performing image recognition on an image of the obstacle detected by the obstacle capturing unit; the obstacle classifying unit is configured to determine, by using the following operation, the plurality of obstacles obtained by classifying the obstacle according to the obstacle characteristic Type of the obstacle type: the obstacle identified by the obstacle recognition unit searches in the information of the obstacles respectively included in the preset plurality of obstacle types, and determines the type to which the obstacle belongs .
  • an obstacle capturing unit 40 configured to perform obstacle detection according to image information collected by the visual sensor, that is, to find an obstacle
  • the obstacle recognition unit is configured to detect by using the following operation Identifying an obstacle: performing image recognition on an image of the obstacle detected by the obstacle capturing unit
  • the obstacle classifying unit is configured to determine, by using the following operation, the plurality of obstacles obtained by classifying the obstacle according to the obstacle characteristic Type of the obstacle type
  • the above obstacle capturing unit is optional, and the processing of performing obstacle detection based on the image information acquired by the visual sensor may also be performed by a visual sensor or a subsequent functional unit.
  • the obstacle processing unit constructs a map and marks the obstacle, including: if the obstacle belongs to The type is a catastrophic type, and the catastrophic obstacle map layer is constructed to mark the obstacle, the maintenance area of the catastrophic obstacle map layer is smaller than the maintenance area of other map layers, and the other map layers are set to mark the plurality of obstacles Other types of obstacles in the object type.
  • the map is a navigation map including a static map layer, and the static map layer is marked with an obstacle existing in the environment before navigation;
  • the obstacle recognition unit is configured to detect the detected obstacle by the following operations; Identifying: identifying, during the navigation process, new obstacles detected in the environment; constructing a map and marking the obstacles by the obstacle processing unit includes: constructing a dynamic map layer and marking the New obstacles.
  • the system further includes a map layer fusion unit 50 configured to perform an expansion process on the periphery of the obstacle, and fuse the static map layer and the dynamic map layer to obtain an updated navigation main map.
  • the static map layer and the dynamic map layer are merged, that is, the obstacles marked by the static map layer and the dynamic map layer are marked on the same map according to their positions, and the obstacles, the expanded grid and the map coordinates in the map are maintained.
  • Other relevant information such as the extent of the expansion, the type to which the obstacle belongs, and the like are also stored in the associated data of the map.
  • the fusion of the map layer is optional.
  • the above expansion strategy, reuse strategy, and avoidance strategy can still be used by classifying the obstacle characteristics.
  • the expansion process performed on the obstacle is handled by a separate functional unit.
  • the plurality of obstacle types include a risk type, the risk type obstacle includes an infant; the system further includes: an obstacle expansion unit configured to perform a periphery of the obstacle marked in the map An expansion process, and when the type to which the obstacle belongs is a risk type, a first expansion radius is used in performing the expansion process, the first expansion radius being greater than other types of obstacles in the plurality of obstacle types The radius of expansion used for the expansion treatment in the periphery.
  • the map is a navigation map including a static map layer
  • the plurality of obstacle types include a stable type
  • the stable obstacle includes at least one of a building member and a piece of furniture.
  • the system further includes an obstacle update unit 60 configured to add the new stable obstacle marked by the dynamic map layer to the static map layer after the navigation ends.
  • the above functional unit of the map construction system is interface-encapsulated, so that it becomes a service platform that can provide map services to services such as mobile robot services.
  • the functional units and interfaces described above may also be implemented separately using discrete devices.
  • the interface of the map construction system of the embodiment includes: a navigation map construction interface 70, configured to receive an external request for updating the navigation map, and activate the obstacle recognition unit, the obstacle classification unit, the obstacle processing unit, and the map layer fusion unit to execute The map update task is navigated and the updated navigation master map is returned to the mobile robot service.
  • a navigation map construction interface 70 configured to receive an external request for updating the navigation map, and activate the obstacle recognition unit, the obstacle classification unit, the obstacle processing unit, and the map layer fusion unit to execute The map update task is navigated and the updated navigation master map is returned to the mobile robot service.
  • the interface of the map construction system may further include: a static map construction interface 80, configured to receive an external request constructed by the static map, complete the construction of the static map of the indoor environment, and return to the requesting party.
  • a static map construction interface 80 configured to receive an external request constructed by the static map, complete the construction of the static map of the indoor environment, and return to the requesting party.
  • the static layer map layer can also be implemented by importing an environment map, which is not necessarily created by the map construction system.
  • the map update interface 90 is configured to receive an external request for the map update, and add a new stable obstacle marked by the dynamic map layer to the static map layer after the end of one navigation.
  • the map save interface 100 is set to save the built or updated static map.
  • the embodiment further provides a map construction apparatus, including a memory, a processor, and a computer program stored on the memory and operable on the processor, and the processor implements the program as implemented in the present embodiment.
  • a map construction apparatus including a memory, a processor, and a computer program stored on the memory and operable on the processor, and the processor implements the program as implemented in the present embodiment. Example method.
  • the map construction scheme of the embodiment of the present application makes the representation of the map more semantic and refined, and thus can provide more practicality, flexibility and security for the autonomous navigation of the mobile robot, and can also save resources. Ideal for indoor mobile robot navigation applications.
  • This embodiment provides a navigation method, as shown in FIG. 3, including step 210 and step 220.
  • step 210 the newly added obstacles detected on the travel path are identified, and the type of the new obstacles classified in the plurality of obstacle types classified according to the obstacle characteristics is determined according to the recognition result.
  • step 220 the obstacle avoidance process is performed according to the type to which the newly added obstacle belongs.
  • the method for classifying according to the obstacle characteristics includes a negotiation type and other types
  • the negotiation type obstacle includes an adult
  • the other types may be a stable type, a radical type, etc., but in this embodiment, the other types are not limited.
  • the obstacle avoidance processing is performed according to the type of the newly added obstacle, including: if the newly added obstacle is a negotiated obstacle, interacting with the negotiated obstacle, requesting the negotiation type Obstacle avoidance.
  • the method further includes: determining whether the negotiated obstacle performs the following within a set time An operation of at least one of: exiting the travel path and returning a response to consent avoidance; performing at least one of (i) leaving the travel path and (ii) returning a consent avoidance response based on the negotiated obstacle
  • the result of the judgment of the operation proceeds in accordance with the travel path; and the new progress path is re-planned based on the judgment result that the negotiated obstacle does not perform the operation of at least one of the travel path and the response to return the consent avoidance.
  • the navigation map used for navigation is an indoor grid map for mobile robot navigation, and may be constructed by using the map construction method in the first embodiment, but this need not be the case, as long as the negotiated obstacle is to be constructed in the constructed map. It can be distinguished from other types of obstacles.
  • the embodiment further provides a navigation system, as shown in FIG. 4, including a navigation module 101 and a map construction module 102.
  • the navigation module 101 is configured to plan a travel path, and notify the map construction module to start a navigation map update task.
  • the map construction module 102 is configured to identify the newly added obstacles detected on the travel path, and determine, according to the recognition result, the plurality of obstacle types obtained by the newly added obstacles according to the obstacle characteristics The type in which the navigation map is updated to mark the newly added obstacle and record the type to which it belongs, and return the updated navigation map to the navigation module.
  • the navigation module 101 is further configured to receive the updated navigation map, and perform obstacle avoidance processing according to the type of the newly added obstacle.
  • the above navigation module can be implemented using a navigation system such as controlling the travel of the mobile robot, and the map construction system can use the map construction system described in the first embodiment, but is not limited thereto.
  • the plurality of obstacle types include a negotiation type, and the negotiation type obstacle includes an adult.
  • the system further includes a human-computer interaction module, and the human-computer interaction module can be, for example, a voice system.
  • the navigation module is configured to perform an obstacle avoidance process according to a type to which the newly added obstacle belongs according to the following operation: if the newly added obstacle is a negotiated obstacle, the human-machine interaction module and the Negotiated obstacles interact to request the negotiated obstacle to avoid.
  • the navigation module interacts with the negotiation obstacle through the human-machine interaction module, and after requesting the negotiation-type obstacle to avoid, the navigation module is further configured to: determine the set time Whether the negotiated obstacle performs an operation of at least one of: leaving the travel path and returning a response to consent avoidance; performing a response to leave the travel path and return a consent avoidance based on the negotiated obstacle a result of the judgment of the at least one operation, keeping the travel path unchanged; re-planning new based on the judgment result that the negotiated obstacle does not perform an operation of at least one of leaving the travel path and returning the consent avoidance response The path of the proceeding.
  • the type of the obstacle is determined according to the characteristics of the obstacle when the navigation is added, and different obstacle avoidance strategies are adopted for different types of obstacles, thereby improving the obstacle avoidance effect.
  • This embodiment describes a raster map construction process based on indoor obstacle feature classification by an example.
  • the mobile robot walks around the indoor environment, and completes the static map layer construction according to the information of the external sensor. Since this phase occurs before the mobile robot navigation task, the obstacles in the static map layer are stable obstacles.
  • the flow includes steps 301 to 313.
  • step 301 a navigation task is initiated.
  • step 302 an obstacle is found.
  • the map construction system detects the image collected by the sensor and finds that a new obstacle appears in the front.
  • step 303 the obstacle image is identified.
  • the map construction system performs feature extraction and recognition according to the image information of the obstacle, such as identifying that the obstacle in front is a cat or a table.
  • step 304 the obstacles are classified.
  • the map construction system classifies according to the classification result according to the category to which the obstacle belongs.
  • step 305 it is determined whether the obstacle is a catastrophic obstacle, and if the obstacle is a catastrophic obstacle, the process proceeds to step 306; if the obstacle is not a catastrophic obstacle, the process proceeds to step 307.
  • step 306 the obstacle is marked on the catastrophic obstacle map layer, and the process proceeds to step 312.
  • the migratory obstacle map layer has a smaller maintenance area than other map layers.
  • step 307 it is determined whether the obstacle is a risk-type obstacle, and based on the judgment result that the obstacle is a risk-type obstacle, the process proceeds to step 308, based on the judgment result that the obstacle is not a risk-type obstacle. Proceed to step 309.
  • step 308 the obstacle is marked on the risky obstacle map layer and the process proceeds to step 312.
  • step 309 it is determined whether the obstacle is a negotiated obstacle, and based on the judgment result that the obstacle is a negotiated obstacle, the process proceeds to step 310, based on the judgment result that the obstacle is not a negotiated obstacle. Proceed to step 311.
  • step 310 the obstacle is marked on the negotiated obstacle map layer, and the process proceeds to step 312.
  • step 311 the obstacle is marked on the stable obstacle map layer.
  • step 312 each of the dynamic map layers and the static map layer are merged to obtain a main raster map for navigation.
  • the periphery of the obstacle Before or after the fusion, the periphery of the obstacle may be expanded, wherein the radius of expansion of the obstacle in the map layer of the risk obstacle is larger than the radius of expansion of the obstacle in the other map layer.
  • step 313 it is determined whether or not the navigation task is completed. Based on the determination result of the end of the navigation task, the current flow is ended, and based on the determination result that the navigation task has not ended, the process proceeds to step 302.
  • Figure 6 through Figure 11 show a simple example of map layer construction based on indoor obstacle feature classification.
  • Figure 6 shows an example of a static map layer.
  • the grid of the map layer preserves the general distribution of the entire indoor environment. For the sake of understanding, the indoor environment is simplified here. If the map layer is maintained from (0,0) to (19,19), only four stable obstacles are included, which are located at the coordinate points (17). , 0), (0, 4), (0, 19) and (19, 16).
  • Figure 7 shows an example of a stable obstacle map layer.
  • the raster of the map layer holds the stable obstacles found during the navigation of the mobile robot. It does not need to pay attention to the entire indoor environment space at all times, but only maintains obstacles in a certain area in front of the sensor.
  • the area covered by the map layer is a 6*6 grid.
  • the grid coordinates corresponding to the current maintenance area are from (3, 3) to (8, 8), and a newly found one is located at the coordinates.
  • Stable obstacle on point (6, 5).
  • Figure 8 shows an example of a negotiated obstacle map layer.
  • the raster of the map layer holds the negotiated obstacles found during the navigation of the mobile robot. It does not need to pay attention to the entire indoor environment space at all times, but only maintains obstacles in a certain area in front of the sensor.
  • the area covered by the map layer is a 6*6 grid.
  • the grid coordinates corresponding to the current maintenance area are from (0,0) to (5,5), and a new one is located at the coordinates.
  • Figure 9 shows an example of a risky obstacle map layer.
  • the raster of the map layer holds the risky obstacles found during the navigation process of the mobile robot. It does not need to pay attention to the entire indoor environment space at all times, but only maintains obstacles in a certain area in front of the sensor.
  • the area of the map layer responsible for maintenance is a 6*6 grid.
  • the grid coordinates corresponding to the current time maintenance area are from (7, 7) to (12, 12), and a newly found one is located at the coordinates. Risky obstacles at points (10, 9).
  • Figure 10 shows an example of a radical obstacle map layer.
  • the raster of the map layer holds the catastrophic obstacles found during the navigation of the mobile robot. It does not need to pay attention to the entire indoor environment space at all times, but only maintains obstacles in a certain area in front of the sensor.
  • the area of the map layer responsible for maintenance is a 3*3 grid.
  • the grid coordinates corresponding to the current maintenance area are from (13, 13) to (15, 15), and a newly found one is located at the coordinates.
  • Figure 11 shows an example of a main raster map after being merged by a map layer fusion unit.
  • the map layer first extracts the obstacles of each map layer and places them on the coordinates corresponding to the main raster map.
  • the various barrier perimeter grids are then expanded for different expansion treatment strategies. Specifically, the risk-type obstacle is expanded by a radical expansion radius, and the other obstacles are expanded by a conservative expansion radius.
  • the aggressive expansion radius is 2 grid lengths and the conservative expansion radius is 1 grid length.
  • the map construction system may update the obstacle data of the stable obstacle map layer to the static map layer.
  • the updated static map layer is shown in Figure 12.
  • This embodiment describes an obstacle avoidance strategy for an exemplary negotiated obstacle. As shown in FIG. 13, the flow of the negotiation avoidance method in this embodiment includes steps 401 to 407.
  • step 401 during the navigation process of the mobile robot, the original planned travel path is blocked due to the appearance of a new obstacle in front.
  • step 402 the navigation system extracts the type identifier of the obstacle on the main grid map, determines that the obstacle is a negotiated obstacle, and notifies the mobile robot voice interaction service module.
  • step 403 the voice interaction service module initiates a voice interaction process, requests a negotiated obstacle avoidance in front, and waits for a response.
  • step 404 it is determined whether the negotiated obstacle sends a positive voice response and takes evasive measures to leave the travel path within the set time; a positive voice response is issued based on the negotiated obstacle and a evasive measure is taken, leaving the travel path
  • step 405 is executed, based on the negotiation type obstacle, the positive voice response is not issued and the avoidance measure is taken, and the determination result without leaving the travel path is passed to step 406.
  • Step 405 may be performed when a positive response to the negotiated obstacle is obtained, or if it is determined that one of the two conditions of the negotiated obstacle has been avoided.
  • step 405 the navigation system does not change the travel path, instructing the robot to normally travel to the target in accordance with the travel path, ending.
  • step 406 the navigation system takes the position of the current mobile robot as a starting point, and ends the travel path with the target point as the end point. If the travel path is re-planned successfully, step 407 is performed. If it fails, it indicates that the target point has not been reached. A path that forces the navigation task to be terminated.
  • step 407 the navigation system instructs the mobile robot to reach the target point in accordance with the new travel path, and ends.
  • the grid map construction system of the embodiment is provided with a plurality of interfaces, and the map service can be provided externally.
  • This embodiment is referred to as a grid map service platform.
  • This embodiment describes the calling process of the mobile robot service to the grid map service platform, and the third-party mobile robot service may be other application scenarios based on the autonomous navigation of the indoor robot.
  • Autonomous navigation is a basic function. What is done after navigating to the target point will generate value is a problem solved by the third-party mobile robot business.
  • the flow includes the following steps.
  • Step 1 The third-party mobile robot service invokes a static map construction interface provided by the grid map service platform to start a map creation task.
  • Step 2 The grid map service platform starts the map real-time creation process, and returns the map real-time creation interface to the third-party mobile robot business.
  • Step 3 The third-party mobile robot business drives the mobile robot to move around the indoor environment.
  • Step 4 The third-party mobile robot service sends the indoor environment information scanned by the sensor to the static map construction interface on the grid map service platform.
  • Step 5 The grid map service platform “updates the map” and sends it to the third-party mobile robot service through the static map construction interface, and performs real-time update and presentation of the map interface.
  • Step 6 Repeat steps 3 through 5 until the indoor environment map is completed.
  • the third-party mobile robot service calls the map save interface of the raster map service platform to save the constructed map.
  • Step 7 The map save interface starts the save process, and saves the constructed map data as a map file in the external storage.
  • Step 8 After the map is saved, the map save interface returns the completed status to the third-party mobile robot service.
  • Step 9 The third-party mobile robot performs autonomous navigation, loads the saved map file into the navigation system, and initiates a map update task by calling a navigation map construction interface provided by the raster map service platform.
  • Step 10 Navigating the map construction interface execution flow, respectively starting a stable obstacle map layer processing unit, a negotiated obstacle map layer processing unit, a radical obstacle map layer processing unit, a risk type obstacle map layer processing unit, and a map layer
  • the fusion unit is configured to process map layer construction and main map generation corresponding to various types of obstacles in the mobile robot navigation process.
  • Step 11 The navigation map construction interface returns the obtained main raster map for real-time navigation to the third-party mobile robot service.
  • Step 12 The third-party mobile robot service performs navigation and other business tasks using the main raster map of real-time navigation.
  • Step 13 The third-party mobile robot service completes the current navigation task, and invokes the grid map service platform map update interface to update the static layer map.
  • step 14 the map update interface execution flow updates the static map layer by using the information of the stable obstacle map layer.
  • Step 15 After the map update is completed, the map update interface exports the updated map, replaces the original map file, and returns the status of the update completion to the third-party mobile robot service, and the entire process ends.
  • Such software may be distributed on a computer readable medium, which may include computer storage media (or non-transitory media) and communication media (or transitory media).
  • computer storage medium includes volatile and nonvolatile, implemented in any method or technology for storing information, such as computer readable instructions, data structures, program modules or other data. Sex, removable and non-removable media.
  • Computer storage media include, but are not limited to, Random Access Memory (RAM), Read-Only Memory (ROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Flash memory Or other memory technology, Compact Disc Read-Only Memory (CD-ROM), Digital Video Disc (DVD) or other optical disc storage, magnetic cassette, magnetic tape, disk storage or other magnetic storage device, Or any other medium that can be used to store the desired information and that can be accessed by the computer.
  • communication media typically includes computer readable instructions, data structures, program modules, or other data in a modulated data signal, such as a carrier wave or other transport mechanism, and can include any information delivery media. .

Landscapes

  • Engineering & Computer Science (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Health & Medical Sciences (AREA)
  • Business, Economics & Management (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Game Theory and Decision Science (AREA)
  • Medical Informatics (AREA)
  • Databases & Information Systems (AREA)
  • Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
  • Navigation (AREA)

Abstract

一种地图构建、导航方法及装置、系统，对检测到的障碍物进行识别，根据识别结果确定所述障碍物在根据障碍物特性分类得到的多种障碍物类型中所属的类型；构建地图并标记所述障碍物，记录所述障碍物所属的类型。在导航时，对行进路径上检测到的新增的障碍物，根据所述新增的障碍物所属的类型进行避障处理。

Description

一种地图构建、导航方法及装置、系统
本申请要求在2017年12月29日提交中国专利局、申请号为201711499191.0的中国专利申请的优先权，该申请的全部内容通过引用结合在本申请中。
技术领域
本申请涉及但不局限于数据处理领域，例如涉及一种地图构建、导航方法及装置、系统。
背景技术
在传统的栅格地图构建方法当中，采用一张完整的二维网格来抽象表示整个室内环境信息。这张栅格地图上的栅格分三类，障碍物栅格、非障碍物栅格和未知栅格。其中，障碍物栅格代表该栅格对应位置上有障碍物存在，非障碍物栅格表示该栅格对应位置是通行自由的，而未知栅格则表示该栅格对应位置上的信息未知。
传统栅格地图构建方法，将所有的环境信息紧耦合在同一张栅格地图上面，当次导航的地图仅限当次使用，复用性很差。且栅格地图并没有考虑障碍物所在栅格对附近栅格通行情况的影响。由于移动机器人本身会占据一定的空间，因此障碍物的附近栅格，机器人并不能无限制通行。否则，很可能在经过这些栅格时与障碍物发生碰撞。
针对上述情况，相关技术的一种改进方案是，引入栅格地图的分层管理和更新机制，将栅格地图由下而上划分成静态地图层、障碍物地图层、声呐障碍物地图层和膨胀地图层等。其中，静态地图层管理移动机器人在室内无动态障碍物干扰情况下行走，是根据传感器检测到的信息构建的整体环境的栅格表示。障碍物地图层是移动机器人在执行导航任务过程中，激光传感器检测到的动态障碍物的栅格表示。声呐障碍物地图层为移动机器人在执行导航任务过程中，声呐传感器检测到的动态障碍物的栅格表示。膨胀地图层并不使用外部的传感器数据，而是对静态地图层、障碍物地图层和声呐地图层当中的障碍物栅格为支点，以事先给定的膨胀半径为尺度，对这些障碍物栅格的周边非障碍物栅格 进行膨胀处理。膨胀半径范围内的非障碍物栅格，形成一个膨胀地图层。位于膨胀地图层上的非障碍物栅格，机器人可以在保证其中心点不与这些非障碍物栅格重合即可通过。
但是，这种构建地图的方法在性能上，例如实用性、灵活性、安全性和资源利用率中的至少一个方面，仍存在不足。
发明内容
以下是对本文详细描述的主题的概述。本概述并非是为了限制权利要求的保护范围。
本申请实施例提供了一种地图构建方法，包括：对检测到的障碍物进行识别，根据识别结果确定所述障碍物在根据障碍物特性分类得到的多种障碍物类型中所属的类型；构建地图并标记所述障碍物，记录所述障碍物所属的类型。
本申请实施例还提供了一种地图构建系统，包括：障碍物识别单元，设置为对检测到的障碍物进行识别，得到识别结果；障碍物分类单元，设置为根据所述识别结果，确定所述障碍物在根据障碍物特性分类得到的多种障碍物类型中所属的类型；障碍物处理单元，设置为构建地图并标记所述障碍物，记录所述障碍物所属的类型。
本申请实施例还提供了一种地图构建装置，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，所述处理器执行所述程序时实现如本申请实施例所述的地图构建方法。
本申请实施例还提供了一种计算机存储介质，其上存储有计算机程序，所述计算机程序被处理器读取后执行本申请实施例所述的地图构建方法。
本申请实施例还提供了一种导航方法，包括：对行进路径上检测到的新增的障碍物进行识别，根据识别结果确定所述新增的障碍物在根据障碍物特性分类得到的多种障碍物类型中所属的类型；根据所述新增的障碍物所属的类型进行避障处理。
本申请实施例还提供了一种导航系统，包括导航模块和地图构建模块，其中：所述导航模块，设置为规划行进路径，通知所述地图构建模块启动导航地图更新任务；所述地图构建模块，设置为对所述行进路径上检测到的新增的障碍物进行识别，根据识别结果确定所述新增的障碍物在根据障碍物特性分类得 到的多种障碍物类型中所属的类型；所述导航模块还设置为接收所述更新后的导航地图，根据所述新增的障碍物所属的类型进行避障处理。
本申请实施例还提供了一种导航装置，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，所述处理器执行所述程序时实现如本申请实施例所述的导航方法。
本申请实施例还提供了一种计算机存储介质，其上存储有计算机程序，所述计算机程序被处理器读取后执行本申请实施例所述的导航方法。
在阅读并理解了附图和详细描述后，可以明白其他方面。
附图说明
图1是本申请实施例一地图构建方法的流程图；
图2是本申请实施例一地图构建系统的模块图；
图3是本申请实施例二导航方法的流程图；
图4是本申请实施例二导航系统的模块图；
图5是本申请实施例三基于室内障碍物特征分类的栅格地图构建方法的流程图；
图6是本申请实施例三一个示例性的静态地图层的示意图；
图7是本申请实施例三一个示例性的稳定型障碍物地图层的示意图；
图8是本申请实施例三一个示例性的协商型障碍物地图层的示意图；
图9是本申请实施例三一个示例性的风险型障碍物地图层的示意图；
图10是本申请实施例三一个示例性的激变型障碍物地图层的示意图；
图11是本申请实施例三一个示例性融合得到的主栅格地图的示意图；
图12是本申请实施例三一个示例性更新后的静态地图层的示意图；
图13是本申请实施例四在导航过程中协商避让方法的流程图；
图14是本申请实施例五移动机器人业务对地图构建系统提供的栅格地图服务的调用流程。
具体实施方式
为下文中将结合附图对本申请的实施例进行详细说明。需要说明的是，在不冲突的情况下，本申请中的实施例及实施例中的特征可以相互任意组合。
相关技术的建图方法对于障碍物的处理，是根据激光、声呐等传感器的检测结果来粗略地划分类别，并未根据识别出的障碍物本身的特性来分类，语义性很弱，使得根据障碍物特性而采用灵活的构图、避障策略难以实现。
这种语义的微弱，会导致以下情况的发生：
1)由于障碍物之间没有明显的区别，在应用中都采用了相同的标记策略。这种策略往往缺乏普适性，甚至会引入无谓的计算和存储资源的浪费。举例来说，在传感器前方发现了一台新添置的冰箱和一只家庭宠物，如果按照相同的发现与标记策略(如3米以内开始跟踪维护)，对于家庭宠物的维护将在障碍物地图层上带来频繁的更新。而家庭宠物行动敏捷而又变化无常，而自身还具有一定的避险能力，对于这类障碍物维护的地图层完全可以限定在一个较小范围，在有效节省资源的同时而又不产生安全上的隐患。反过来，如果按照较小范围来维护新添置的冰箱，则不利于移动机器人导航的全局路线选择。
2)对所有的障碍物都采用同样一个膨胀半径来处理，在实际的场景中并不合理。事实上，环境空间中的一台新添的冰箱，其周边的非障碍物栅格完全可以按照较小的膨胀半径来处理，以保证机器人最大的通行效率。但对于环境空间中突然出现的一个婴幼儿，则需要以较大的膨胀半径来处理其周边的非障碍物栅格，以保证安全性。
3)一次导航任务结束时，障碍物层的所有信息都被丢弃。这种设计的考虑是，当次导航时出现在环境当中的新增障碍物，在下次导航时不一定还出现并且保持位置不变。但实际上，如果对新增障碍物加以区别的话，至少像新添置的家具这样的障碍物，其位置一旦固定就不会经常发生变化。这类障碍物完全可以在当次导航任务结束后，实时更新到静态地图层中去。这样，既保证了静态地图层的实时更新，又为以后的导航任务节省了不必要的障碍物识别标记的计算量和存储空间。
4)在避障时所有障碍物无差别对待，直接导致移动机器人避障的灵活性丧失。事实上，如果移动机器人识别出前方的成年人阻挡去路时，完全可以采取语音交互的方式变被动避障为主动避障。
为此，本申请实施例提出了新的地图构建和导航方案。
实施例一
本实施例涉及地图构建方法和系统。
本实施例的地图构建方法由地图构建系统来实现，在移动机器人导航这一应用中，地图构建系统可以植入在移动机器人使用的导航芯片中，可以与移动机器人使用的视觉传感器集成，或者可以采用单独的芯片实现。在其他应用中，也是类似的。
本实施例的地图以室内栅格地图为例，但本申请并不局限于此，其障碍物分类的方式也可扩展到室外地图，非栅格地图等。
本实施例的地图构建方法如图1所示，包括步骤110和步骤120。
在步骤110中，对检测到的障碍物进行识别，根据识别结果确定所述障碍物在根据障碍物特性分类得到的多种障碍物类型中所属的类型。
与相关技术不同的是，以上障碍物的特性是指障碍物本身的特性，在确定障碍物分类时需要先识别出障碍物，如识别出障碍物是人、动物、电视、柜子等。这与相关技术根据检测结果直接分类是不同的，例如，相关技术检测当前障碍物在一段时间内是否移动，根据检测结果将障碍物分为动态和静态两类，这种分类体现的并非障碍物本身的特性，也无法确定障碍物是否具备自主移动能力。
本实施例中，所述障碍物特性包括障碍物的自主移动能力、交互能力、安全性和自主避险能力中的至少一种。其中，障碍物的自主移动能力指在没有外力移动该障碍物的情况下，障碍物自身就可以移动，例如，人、动物、玩具车、移动机器人具有自主移动能力，而桌椅、沙发、电视等家具、电器不具有自主移动能力。障碍物的交互能力指人机交互的能力，包括但不限于语音交互、手势交互等方式，成人或另一支持交互的移动机器人可以属于此类，移动机器人也可视为一种特殊的“人”。障碍物的安全性指障碍物是否被保护对象，便如婴幼儿，用户也可以将其他物品指定为被保护对象。障碍物的自主避险能力是指障碍物在应激情况下，例如移动机器人靠近时，是否可以自动避让。在对具体的对象分类时，对象的上述能力可以有重叠的情况，例如具备自主避险能力的障碍物通常也会具有自主移动能力，而具备人机交互能力的成人也是如此。
本实施例中，根据障碍物特性分类得到的多种障碍物类型包括：稳定型、协商型、风险型和激变型中的至少二种，其中：稳定型障碍物包括至少一种不具备自主移动能力的障碍物；协商型障碍物包括至少一种具备自主移动能力和人机交互能力的障碍物；风险型障碍物包括至少一种作为被保护对象的障碍物； 激变型障碍物包括至少一种具备自主避险能力的障碍物。
上述稳定型障碍物包括但不限于墙壁、门窗、婴儿车、沙发、桌椅、橱柜、电视等不具备自主移动能力的物体，这些物体的位置相对比较稳定。一旦这些障碍物在栅格地图当中标记出来，在较长的一段时间栅格地图上的这些信息可以复用。
上述协商型障碍物包括除了婴幼儿之外的具备自主移动能力和语音交互能力的人，他们本身具备良好的行为能力和交互能力，并可根据外部的信息来主动调整自己的行为，可以为移动机器人的自主导航提供更多的选择，增强其灵活性。
上述风险型障碍物包括婴幼儿这种被保护对象，其自主行为能力、语音交互能力和风险辨识能力都比较弱的人群。此类人群是移动机器人自主导航中最应该预防和被保护的对象。与移动机器人自主导航的效率相比，这类人群的安全性更加重要。
上述激变型障碍物包括但不限于猫、狗、机器人等具备较强移动能力的家庭宠物和电子设备。它们都具有自主避险的能力，机动性好且和移动的随机性强。
目前的图像识别可以对超过1000种物品进行识别，家庭环境常见的障碍物基本都可以识别。对于识别不出来的障碍物可以归入稳定型障碍物进行处理，或者另外归为单独的一类也是可以的。
在另一实施例中，所述根据障碍物特性分类得到的多种障碍物类型包括：稳定型、协商型、风险型和激变型中的至少二种，其中：稳定型障碍物包括建筑构件和家具中的至少一种；协商型障碍物包括成人；风险型障碍物包括婴幼儿；激变型障碍物包括家庭宠物。
在该另一实施例中，也是根据障碍物特性分类，但并不严格要求某一类型的物品具有完全相同的某至少一个特性，而且并不需要对环境中的所有障碍物进行分类，对部分重要的障碍物如上述建筑构件、家具、成人、婴幼儿和家庭宠物分类即可。将成人、婴幼儿、建筑构件、家庭宠物中的至少二种按照不同的策略进行处理，则说明进行了在构建地图时进行了分类和记录，即采用了该另一实施例的方法。
另外，上述的“成人”是指具备自主移动能力和人机交互能力的健康人， 对于下肢残疾、聋哑这些特殊的成人，可以另行指定其分类，特别在家庭成员相对固定时。当然也可以不做细分，也归入协商型障碍物，在避障时得不到响应可以重新规划路线。
另外，需要说明的是，上述障碍物类别的名称是可以变化的，例如，以上四种类型也可以称为第一类、第二类、第三类和第四类或者其他任何能够区别为不同类型的名称，这种名称变化并不会构成技术方案上的区别。
本实施例中，对检测到的障碍物进行识别，根据识别结果确定所述障碍物在根据障碍物特性分类得到的多种障碍物类型中所属的类型，包括：对视觉传感器采集的图像信息进行障碍物检测，对检测到的障碍物的图像进行图像识别，根据识别出的所述障碍物在查找预先设置的所述多种障碍物类型分别包含的障碍物的信息中进行查找，确定所述障碍物所属的类型。例如，视觉传感器采集的图像中出现一只猫时，通过障碍物检测将猫的图像提取出来，做图像识别知道这是一只猫，而在预先设置的多种障碍物类型分别包含的障碍物中，猫是属于激变型障碍物，因而就可以确定该障碍物的类型为激变型。上述分类可采用分类器实现。
本申请中，对障碍物类型的记录可以采用直接或间接的方式，如直接将检测出的类型记录为障碍物的属性信息中的类型标识，或者只是将障碍物标记在与类型对应的层上，根据障碍物所在的层就可以确定障碍物的类型。因而，本申请不对记录方式做任何的局限，只需记录的信息可用于后续确定障碍物的类型即可。
在步骤120中，构建地图并标记所述障碍物，记录所述障碍物所属的类型。
本实施例采用的是栅格地图，所述构建地图并标记所述障碍物，包括：为不同类型的障碍物构建不同的地图层，根据所述障碍物所属的类型在与所述障碍物所属类型对应的地图层上标记所述障碍物。
本实施例中，上述构建的地图是导航地图，而对障碍物的识别、分类和标记是对导航时检测到的相对环境地图已有障碍物新增的障碍物进行识别、分类和标记。但本申请不局限于此，按本实施例方法构建的地图也可以用于导航之外的其他服务，例如构建的是实时动态的场景地图，用于进行实景游戏、侦察等服务。此时并没有原有障碍物、新增障碍物的区分。
本实施例的导航地图是包括静态地图层的导航地图，所述静态地图层标记 有导航进行之前环境中存在的障碍物，静态地图层可以根据导入的环境地图而生成，其中标记的障碍物为稳定型障碍物；在导航过程中，对环境中检测到的新增的障碍物进行识别，构建动态地图层并标记所述新增的障碍物，记录所述新增的障碍物所属的类型；之后，还可以对障碍物的周边进行膨胀处理，融合所述静态地图层和动态地图层，得到更新后的导航主地图。
在一个示例中，导航地图可以包括以下地图层。
静态地图层设置为维护移动机器人导航之前整个室内环境的总体格局，其中的障碍物属于稳定型障碍物。
动态地图层包括：稳定型障碍物地图，设置为维护移动机器人导航过程中视觉传感器检测到的、静态地图层当中没有包含的新增稳定型障碍物。
激变型障碍物地图层，设置为维护移动机器人导航过程中视觉传感器检测到的激变型障碍物。
协商型障碍物地图层，设置为维护移动机器人导航过程中视觉传感器检测到的协商型障碍物。
风险型障碍物地图层，设置为维护移动机器人导航过程中视觉传感器检测到的风险型障碍物。
这样，各层地图彼此独立起来，具有了各自鲜明的语义。在使用时，只需要按照顺序对个地图层的数据进行融合即可。
在记录了障碍物所属的类型之后，本实施例根据障碍物所属的类型，采用了以下至少一种策略。
标记策略
在所述多种障碍物类型包括激变型(激变型障碍物包括家庭宠物)时，所述构建地图并标记所述障碍物，包括：如所述障碍物所属的类型为激变型，构建激变型障碍物地图层标记所述障碍物，所述激变型障碍物地图层的维护区域小于其他地图层的维护区域，所述其他地图层设置为标记所述多种障碍物类型中其他类型的障碍物。维护区域小，构建的地图也小。这里的其他地图层可以是稳定型障碍物地图层、风险型障碍物地图层和协商型障碍物地图层中的至少一个，但在其他实施例中，也可以是用于标记其他类型障碍物的地图层。
不同障碍物采用不同的标记策略，可以有效避免激变型的障碍物由于频繁变动位置而导致障碍地图频繁的更新，节省了系统资源耗费。
膨胀策略
在所述多种障碍物类型包括风险型(风险型障碍物包括婴幼儿)时，构建地图并标记所述障碍物，记录所述障碍物所属的类型之后，所述方法还包括：对所述障碍物的周边进行膨胀处理，如所述障碍物所属的类型为风险型，在进行膨胀处理时使用第一膨胀半径，所述第一膨胀半径大于对所述多种障碍物类型中其他类型的障碍物的周边进行膨胀处理时使用的膨胀半径。
在地图分层时，膨胀处理可以在对各地图层融合之前，对本层地图当中的障碍物周边按照相应的膨胀半径进行膨胀处理，处理完后再进行融合。也可在融合之后再进行，本申请对此不做局限。另外，膨胀策略在不分层的地图中也可以采用。
在导航过程中发现新的风险型障碍物时，会检查风险型障碍物栅格周边是否满足充分膨胀的条件，即以风险型障碍物栅格为中心，以激进膨胀半径为尺度，在风险型障碍物前、后、左、右、左上、右上、左下、右下八个方向上都进行既定范围的膨胀处理。当风险型障碍物激进膨胀半径范围内有其他障碍物，或抵达主地图边界时，会出现膨胀不充分的现象。当风险型障碍物栅格周边栅格满足充分膨胀条件，进一步检查膨胀半径外是否满足避障条件；即当前位置和目标位置之间的行进路径上有机器人能通过的无障碍物的栅格。当膨胀半径外满足避障条件时，移动机器人经膨胀半径外的栅格避障并行进至目标点；而当风险型障碍物周边栅格不满足充分膨胀条件，或者膨胀半径外无法满足避障条件时，表明此路不通或该障碍物周边的自由空间不足以保证安全导航的要求，需要重新规划新的行进路线到达目标点。
不同障碍物采用不同的膨胀策略，将有效提高移动机器人导航路线的安全性，保证婴幼儿这种被保护对象的安全。
重用策略
在所述多种障碍物类型包括稳定型(稳定型障碍物包括建筑构件和家具中的至少一种)时，在导航结束之后，所述方法还包括：将在所述动态地图层标记的新增的稳定型障碍物添加到所述静态地图层。
在一个示例中，如所述新增的障碍物为稳定型障碍物，则构建稳定型障碍物地图层标记所述新增的障碍物，如所述新增的障碍物为其他类型的障碍物，构建其他动态地图层标记所述新增的障碍物；在该示例中，在一次导航结束之 后，就可以将所述稳定型障碍物地图层上标记的新增的障碍物直接添加到所述静态地图层。但在其他示例中，也可以根据障碍物属性中的类型信息来选择稳定型障碍物添加到静态地图层中，并不必须为新增的障碍物构建一个单独的动态地图层。
对于当次导航和后续导航，环境当中的稳定型障碍物的位置可能发生了改变，如果这种变化没有在静态地图当中体现出来，则会对环境地图建模的准确性产生影响，进而影响到后续的自主导航。根据检测到的稳定型障碍物对静态地图层进行更新和完善，避免了该稳定型障碍物在下一次导航任务中的重复处理工作，提高了效率。如果在静态地图层中标记的障碍物在当次导航，也可以对静态地图层更新以将相应的障碍物删除。
同时，不同障碍物的划分，使得在避障时有了更多的延伸策略和补救措施(如行进过程中前面遇到一个协商型障碍物，而剩余的自由空间不够时，可以尝试人机对话的方式来变被动避障变为主动避障。这将在下一实施例中进行详细说明。
本实施例还提出了一种可实现上述方法的地图构建系统，所述地图构建系统植入在移动机器人使用的导航芯片中，与所述移动机器人使用的视觉传感器集成，或者采用单独的芯片实现，该芯片或视觉传感器与机器人控制系统相连。
如图2所示，本实施例的地图构建系统包括障碍物识别单元10，障碍物分类单元20，障碍物处理单元30，障碍物捕捉单元40，地图层融合单元50，障碍物更新单元60，导航地图构建接口70，静态地图构建接口80，地图更新接口90以及地图保存接口100。
障碍物识别单元10，设置为对检测到的障碍物进行识别，得到识别结果.
障碍物分类单元20，设置为根据所述识别结果，确定所述障碍物在根据障碍物特性分类得到的多种障碍物类型中所属的类型。
障碍物处理单元30，设置为构建地图并标记所述障碍物，记录所述障碍物所属的类型。
本实施例中，所述障碍物特性包括障碍物的自主移动能力、交互能力、安全性和自主避险能力中的至少一种。
本实施例中，所述根据障碍物特性分类得到的多种类型包括：稳定型、协商型、风险型和激变型中的至少二种，其中：稳定型障碍物包括建筑构件和家 具中的至少一种；协商型障碍物包括成人；风险型障碍物包括婴幼儿；以及激变型障碍物包括家庭宠物。
本实施例中，所述构建的地图为室内栅格地图；所述障碍物处理单元设置为通过以下操作构建地图并标记所述障碍物：为不同类型的障碍物构建不同的地图层，根据所述障碍物所属的类型在与所述障碍物所属类型对应的地图层上标记所述障碍物。
此时，所述障碍物处理单元可以细分为以下子单元：激变型障碍物地图层处理单元，设置为对激变型障碍物进行处理，构建激变型障碍物地图层；风险型障碍物地图层处理单元，设置为对风险型障碍物进行处理，构建风险型障碍物地图层；协商型障碍物地图层处理单元，设置为对协商型障碍物进行处理，构建协商型障碍物地图层；稳定型障碍物地图层处理单元，设置为对稳定型障碍物进行处理，构建稳定型障碍物地图层。不属于以上三类的障碍物，也可以交由稳定型障碍物地图层处理单元处理。
上述不同的地图层处理单元在构建相应地图层时，有一些不同的处理，例如，激变型障碍物地图层处理单元在筛选障碍物时使用的筛选距离最小，筛选距离与地图层的维护区域大小是等价的。另外，各个地图层处理单元会给本层上的障碍物添加不同的类型值。
本实施例中，所述系统还包括：障碍物捕捉单元40，设置为根据视觉传感器采集的图像信息进行障碍物检测，即发现障碍物；所述障碍物识别单元设置为通过以下操作对检测到的障碍物进行识别：对所述障碍物捕捉单元检测到的障碍物的图像进行图像识别；所述障碍物分类单元设置为通过以下操作确定所述障碍物在根据障碍物特性分类得到的多种障碍物类型中所属的类型：根据所述障碍物识别单元识别出的障碍物在预设的所述多种障碍物类型分别包含的障碍物的信息中进行查找，确定所述障碍物所属的类型。
上述障碍物捕捉单元是可选地，根据视觉传感器采集的图像信息进行障碍物检测的处理也可以由视觉传感器或之后的功能单元来完成。
本实施例中，所述多种障碍物类型包括激变型(激变型障碍物包括家庭宠物)时，所述障碍物处理单元构建地图并标记所述障碍物，包括：如所述障碍物所属的类型为激变型，构建激变型障碍物地图层标记所述障碍物，所述激变型障碍物地图层的维护区域小于其他地图层的维护区域，所述其他地图层设置 为标记所述多种障碍物类型中其他类型的障碍物。
本实施例中，所述地图为包括静态地图层的导航地图，所述静态地图层标记有导航进行之前环境中存在的障碍物；所述障碍物识别单元设置为通过以下操作对检测到的障碍物进行识别：在导航过程中，对在所述环境中检测到的新增的障碍物进行识别；所述障碍物处理单元构建地图并标记所述障碍物包括：构建动态地图层并标记所述新增的障碍物。
所述系统还包括：地图层融合单元50，设置为对障碍物的周边进行膨胀处理，融合所述静态地图层和动态地图层，得到更新后的导航主地图。这里，融合所述静态地图层和动态地图层，即将静态地图层和动态地图层标记的障碍物按照其位置标记到同一张地图上，并维护该地图中障碍物、膨胀栅格与地图坐标之间的映射关系。其他相关的信息如膨胀范围、障碍物所属的类型等也保存在该地图的关联数据中。
对于本申请来说，地图层的融合是可选的，在一张不分层的地图上，通过障碍物特性分类，仍然可以使用上述膨胀策略、重用策略及避让策略等。
在另一实施例中，对障碍物进行的膨胀处理由单独的功能单元来处理。在该实施例中，所述多种障碍物类型包括风险型，风险型障碍物包括婴幼儿；所述系统还包括：障碍物膨胀单元，设置为对地图中标记的所述障碍物的周边进行膨胀处理，且在所述障碍物所属的类型为风险型时，在进行膨胀处理时使用第一膨胀半径，所述第一膨胀半径大于对所述多种障碍物类型中其他类型的障碍物的周边进行膨胀处理时使用的膨胀半径。
本实施例中，所述地图为包括静态地图层的导航地图，所述多种障碍物类型包括稳定型，稳定型障碍物包括建筑构件和家具中的至少一种。
所述系统还包括：障碍物更新单元60，设置为在导航结束之后，将所述动态地图层标记的新增的稳定型障碍物添加到所述静态地图层。
本实施例中，将地图构建系统的上述功能单元进行接口化封装，使其成为可以向移动机器人业务等业务提供地图服务的服务平台。在其他实施例中，上述功能单元和接口也可以用分立的器件分别实现。
本实施例地图构建系统的接口包括：导航地图构建接口70，设置为接收导航地图更新的外部请求，启动所述障碍物识别单元、障碍物分类单元、障碍物处理单元和地图层融合单元以执行导航地图更新任务，并将更新后的导航主地 图返回给所述移动机器人业务。
在一实施例中，地图构建系统的接口还可以包括：静态地图构建接口80，设置为接收静态地图构建的外部请求，完成室内环境静态地图的构建，并返回给请求方。但静态层地图层也可以通过导入环境地图来实现，并不一定是地图构建系统创建的。
地图更新接口90，设置为接收地图更新的外部请求，在一次导航结束后，将所述动态地图层标记的新增的稳定型障碍物添加到所述静态地图层。
地图保存接口100，设置为对构建完成的或更新后的静态地图进行保存。
如果环境地图保存在地图构建系统内部的存储卡上，则可以自行完成上述静态地图的更新和保存的处理。
本实施例还提供了一种地图构建装置，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，所述处理器执行所述程序时实现如本实施例方法。
本申请实施例地图构建方案，使得地图的表示更加语义化和精细化，进而可以为移动机器人的自主导航提供更多的实用性、灵活性和安全性，也可以节约资源。非常适用于室内移动机器人导航的应用场景。
实施例二
本实施例提供一种导航方法，如图3所示，包括步骤210和步骤220。
在步骤210中，对行进路径上检测到的新增的障碍物进行识别，根据识别结果确定所述新增的障碍物在根据障碍物特性分类得到的多种障碍物类型中所属的类型。在步骤220中，根据所述新增的障碍物所属的类型进行避障处理。
本实施例中，根据障碍物特性分类的方法包括协商型和其他类型，协商型障碍物包括成人，其他类型可是稳定型、激变型等，但在本实施例中，对其他类型并不进行局限。本实施例根据所述新增的障碍物所属的类型进行避障处理，包括：如所述新增的障碍物为协商型障碍物，与所述协商型障碍物进行交互，请求所述协商型障碍物避让。
在一个示例中，所述与所述协商型障碍物进行交互，请求所述协商型障碍物避让之后，所述方法还包括：判断在设定的时间内，所述协商型障碍物是否执行以下至少之一的操作：离开所述行进路径和返回同意避让的响应；基于所述协商型障碍物执行了(i)离开所述行进路径和(ii)返回同意避让的响应中的 至少一种的操作的判断结果，按照所述行进路径继续前进；基于所述协商型障碍物没有执行离开所述行进路径和返回同意避让的响应中至少一种的操作的判断结果，重新规划新的进行路径。
本实施例中，导航使用的导航地图为移动机器人导航用的室内栅格地图，可以采用实施例一中的地图构建方法构建，但并不必须如此，只要在构建的地图中将协商型障碍物与其他类型障碍物可以区分开来即可。
本实施例还提供了一种导航系统，如图4所示，包括导航模块101和地图构建模块102。
所述导航模块101，设置为规划行进路径，通知所述地图构建模块启动导航地图更新任务。
所述地图构建模块102，设置为对所述行进路径上检测到的新增的障碍物进行识别，根据识别结果确定所述新增的障碍物在根据障碍物特性分类得到的多种障碍物类型中所属的类型，更新所述导航地图以标记所述新增的障碍物并记录其所属的类型，并将更新后的导航地图返回给所述导航模块。
所述导航模块101还设置为接收所述更新后的导航地图，根据所述新增的障碍物所属的类型进行避障处理。
上述导航模块可以使用如控制移动机器人行进的导航系统来实现，而地图构建系统可以使用实施例一所述的地图构建系统，但不局限于此。
本实施例中，所述多种障碍物类型包括协商型，协商型障碍物包括成人。
所述系统还包括人机交互模块，人机交互模块如可以是语音系统。
所述导航模块设置为通过以下操作来根据所述新增的障碍物所属的类型进行避障处理：如所述新增的障碍物为协商型障碍物，通过所述人机交互模块与所述协商型障碍物进行交互，请求所述协商型障碍物避让。
本实施例中，所述导航模块通过所述人机交互模块与所述协商型障碍物进行交互，请求所述协商型障碍物避让之后，所述导航模块还设置为：判断在设定的时间内，所述协商型障碍物是否执行以下至少之一的操作：离开所述行进路径和返回同意避让的响应；基于所述协商型障碍物执行了离开所述行进路径和返回同意避让的响应中至少一种的操作的判断结果，保持所述行进路径不变；基于所述协商型障碍物没有执行离开所述行进路径和返回同意避让的响应中至少一种的操作的判断结果，重新规划新的进行路径。
上述实施例方案在导航时对新增障碍物要确定其根据障碍物特性划分的类型，对不同类型的障碍物采用不同的避障策略，从而提高避障效果。
实施例三
本实施例通过一个示例，描述基于室内障碍物特征分类的栅格地图构建流程，在导航任务开始前，移动机器人围绕室内环境行走，根据外部传感器的信息，完成静态地图层的构建。由于本阶段发生在移动机器人导航任务之前，因此静态地图层当中的障碍物，都属于稳定型障碍物。
如图5所示，该流程包括步骤301至步骤313。
在步骤301中，启动导航任务。
在步骤302中，发现障碍物。
本步骤中，地图构建系统对传感器采集的图像进行检测，发现新的障碍物在前方出现。
在步骤303中，障碍物图像识别。
本步骤中，地图构建系统根据障碍物的图像信息，进行特征提取和识别，如识别出前方的障碍物为一只猫，或一张桌子。
在步骤304中，障碍物分类。
本步骤中，地图构建系统根据所述识别结果，按照障碍物所属的类别进行归类。
在步骤305中，判断所述障碍物是否激变型障碍物，如果障碍物是激变型障碍物，转入步骤306；如果障碍物不是激变型障碍物，转入步骤307。
在步骤306中，在激变型障碍物地图层标记所述障碍物，转入步骤312。
该示例中，激变型障碍物地图层的维护区域比其他地图层都小。
在步骤307中，判断所述障碍物是否风险型障碍物，基于所述障碍物是风险型障碍物的判断结果，转入步骤308，基于所述障碍物不是风险型障碍物的判断结果，转入步骤309。
在步骤308中，在风险型障碍物地图层标记所述障碍物，转入步骤312。
在步骤309中，判断所述障碍物是否协商型障碍物，基于所述障碍物是协商型障碍物的判断结果，转入步骤310，基于所述障碍物不是协商型障碍物的判断结果，转入步骤311。
在步骤310中，在协商型障碍物地图层标记所述障碍物，转入步骤312。
在步骤311中，在稳定型障碍物地图层标记所述障碍物。
在步骤312中，将上述各个动态地图层和静态地图层进行融合，得到导航用的主栅格地图。
在融合之前或之后，可以对障碍物周边进行膨胀处理，其中，对风险型障碍物地图层中障碍物使用的膨胀半径比其他地图层中障碍物的膨胀半径均要大。
在步骤313中，判断导航任务是否结束，基于导航任务结束的判断结果，结束本次流程，基于导航任务没有结束的判断结果，转入步骤302。
图6至图11给出了一个基于室内障碍物特征分类的地图层构建的简单示例。
图6所示的是静态地图层的示例。该地图层的栅格保存的是整个室内环境的大致分布情况。为便于理解，这里对室内环境进行了简化处理，设若本地图层维护的栅格范围自(0，0)至(19，19)，只包括四个稳定型障碍物，分别位于坐标点(17，0)，(0，4)，(0，19)和(19，16)上。
图7所示的是稳定型障碍物地图层的示例。该地图层的栅格保存的是移动机器人导航过程中发现的稳定型障碍物。它无需时刻关注整个室内环境空间，而只是维护传感器前方一定区域内的障碍物变动情况。为便于理解，这里假定本地图层负责维护的区域范围为一个6*6的栅格，当前时刻维护区域对应的栅格坐标自(3，3)至(8，8)，新发现一个位于坐标点(6，5)上的稳定型障碍物。
图8所示的是协商型障碍物地图层的示例。该地图层的栅格保存的是移动机器人导航过程中发现的协商型障碍物。它无需时刻关注整个室内环境空间，而只是维护传感器前方一定区域内的障碍物变动情况。为便于理解，这里假定本地图层负责维护的区域范围为一个6*6的栅格，当前时刻维护区域对应的栅格坐标自(0，0)至(5，5)，新发现一个位于坐标点(2，1)上的协商型障碍物。
图9所示的是风险型障碍物地图层的示例。该地图层的栅格保存的是移动机器人导航过程中发现的风险型障碍物。它无需时刻关注整个室内环境空间，而只是维护传感器前方一定区域内的障碍物变动情况。为便于理解，这里假定本地图层负责维护的区域范围为一个6*6的栅格，当前时刻维护区域对应的栅格坐标自(7，7)至(12，12)，新发现一个位于坐标点(10，9)上的风险型障碍物。
图10所示的是激变型障碍物地图层的示例。该地图层的栅格保存的是移动机器人导航过程中发现的激变型障碍物。它无需时刻关注整个室内环境空间，而只是维护传感器前方一定区域内的障碍物变动情况。为便于理解，这里假定 本地图层负责维护的区域范围为一个3*3的栅格，当前时刻维护区域对应的栅格坐标自(13，13)至(15，15)，新发现一个位于坐标点(14，14)上的激变型障碍物。
图11所示是经地图层融合单元融合之后的主栅格地图示例。该地图层首先将各个地图层的障碍物抽取出来，放置在主栅格地图对应的坐标上。然后按照不同的膨胀处理策略对于其中的各种障碍物周边栅格进行膨胀处理。具体而言，风险型障碍物周边采用激进的膨胀半径加以膨胀处理，其他的障碍物周边采用保守的膨胀半径加以膨胀处理。为简便起见，这里假设激进的膨胀半径为2个栅格长度，保守的膨胀半径为1个栅格长度。
在一实施例中，在当次导航任务结束之后，地图构建系统可以将稳定型障碍物地图层的障碍物数据更新到静态地图层。更新后的静态地图层如图12所示。
实施例四
本实施例描述一个示例性的协商型障碍物的避障策略。如图13所示，本实施例协商避让方法的流程包括步骤401至步骤407。
在步骤401中，移动机器人导航过程中，由于前方新的障碍物的出现，导致原来规划的行进路径受阻。
在步骤402中，导航系统在主栅格地图上，提取该障碍物的类型标识，确定该障碍物是协商型障碍物，通知移动机器人语音交互业务模块。
在步骤403中，语音交互业务模块发起语音交互流程，请求前方的协商型障碍物避让，并等待响应。
在步骤404中，判断在设定时间内，协商型障碍物是否发出肯定的语音响应并采取避让措施，离开行进路径；基于协商型障碍物发出肯定的语音响应并采取避让措施，离开行进路径的判断结果，执行步骤405，基于协商型障碍物没有发出肯定的语音响应并采取避让措施，没有离开行进路径的判断结果，转入步骤406。
可以在得到协商型障碍物的肯定响应，或者确定该协商型障碍物已避让这两个条件之一满足时，即执行步骤405。
在步骤405中，导航系统不改变行进路径，指示机器人按照行进路径正常行进至目标，结束。
在步骤406中，导航系统以当前移动机器人的位置为起点，以目标点为终 点，重新规划行进路径，如重新规划行进路径成功，执行步骤407，如果失败，则表明已经没有可达目标点的路径，可以强制终止导航任务。
在步骤407中，导航系统指示移动机器人按照新的行进路径达到目标点，结束。
实施例五
本实施例的栅格地图构建系统封装有多种接口，可对外提供地图服务，本实施例将其称为栅格地图服务平台。本实施例描述移动机器人业务对栅格地图服务平台的调用流程，第三方移动机器人业务可以是建立在室内机器人自主导航基础上的其他应用场景。自主导航是基础性的功能，导航到目标点后做什么事会产生价值是第三方移动机器人业务解决的问题。
如图14所示，该流程包括以下步骤。
步骤一，第三方移动机器人业务调用栅格地图服务平台提供的静态地图构建接口，启动地图创建任务。
步骤二，栅格地图服务平台启动地图实时创建流程，并返回地图实时创建界面给第三方移动机器人业务。
步骤三，第三方移动机器人业务驱动移动机器人绕室内环境移动。
步骤四，第三方移动机器人业务将传感器扫描到的室内环境信息发送给栅格地图服务平台上的静态地图构建接口。
步骤五，栅格地图服务平台“更新地图”并通过静态地图构建接口发送给第三方移动机器人业务，进行地图界面的实时更新和呈现。
步骤六，重复步骤三至步骤五，直到室内环境地图构建完成，第三方移动机器人业务调用栅格地图服务平台的地图保存接口保存构建的地图。
步骤七，地图保存接口启动保存流程，将构建好的地图数据保存为外存中的地图文件。
步骤八，地图保存完成后，地图保存接口返回完成的状态给第三方移动机器人业务。
步骤九，第三方移动机器人执行自主导航，将上述保存的地图文件载入导航系统，并通过调用栅格地图服务平台提供的导航地图构建接口启动地图更新任务。
步骤十，导航地图构建接口执行流程，分别启动稳定型障碍物地图层处理 单元、协商型障碍物地图层处理单元、激变型障碍物地图层处理单元、风险型障碍物地图层处理单元以及地图层融合单元，设置为处理移动机器人导航过程中的各种类型障碍物对应的地图层构建和主地图生成。
步骤十一，导航地图构建接口将得到的用于实时导航的主栅格图返回给第三方移动机器人业务。
步骤十二，第三方移动机器人业务利用实时导航的主栅格地图执行导航和其他业务任务。
步骤十三，第三方移动机器人业务完成当次导航任务，调用栅格地图服务平台地图更新接口实现对静态层地图的更新。
步骤十四，地图更新接口执行流程，利用稳定型障碍物地图层的信息更新静态地图层。
步骤十五，地图更新完成后，地图更新接口将更新后的地图导出，替换原来的地图文件，并返回更新完成的状态给第三方移动机器人业务，整个流程结束。
本领域普通技术人员可以理解，上文中所公开方法中的全部或某些步骤、系统、装置中的功能模块/单元可以被实施为软件、固件、硬件及其适当的组合。在硬件实施方式中，在以上描述中提及的功能模块/单元之间的划分不一定对应于物理组件的划分；例如，一个物理组件可以具有多个功能，或者一个功能或步骤可以由若干物理组件合作执行。某些物理组件或所有物理组件可以被实施为由处理器，如中央处理器、数字信号处理器或微处理器执行的软件；可以被实施为硬件；或者可以被实施为集成电路，如专用集成电路。这样的软件可以分布在计算机可读介质上，计算机可读介质可以包括计算机存储介质(或非暂时性介质)和通信介质(或暂时性介质)。如本领域普通技术人员公知的，术语计算机存储介质包括在用于存储信息(诸如计算机可读指令、数据结构、程序模块或其他数据)的任何方法或技术中实施的易失性和非易失性、可移除和不可移除介质。计算机存储介质包括但不限于随机存储器(Random Access Memory，RAM)、只读存储器(Read-Only Memory，ROM)、带电可擦可编程只读存储器(Electrically Erasable Programmable Read-Only Memory，EEPROM)、闪存或其他存储器技术、只读光盘(Compact Disc Read-Only Memory，CD-ROM)、数字多功能盘(Digital Video Disc，DVD)或其他光盘存储、磁盒、磁带、磁盘存储 或其他磁存储装置、或者可以用于存储期望的信息并且可以被计算机访问的任何其他的介质。此外，本领域普通技术人员公知的是，通信介质通常包含计算机可读指令、数据结构、程序模块或者诸如载波或其他传输机制之类的调制数据信号中的其他数据，并且可包括任何信息递送介质。

Claims (32)

  1. 一种地图构建方法，包括：
    对检测到的障碍物进行识别，根据识别结果确定所述障碍物在根据障碍物特性分类得到的多种障碍物类型中所属的类型；
    构建地图并标记所述障碍物，记录所述障碍物所属的类型。
  2. 如权利要求1所述的方法，其中：
    所述障碍物特性包括障碍物的自主移动能力、交互能力、安全性和自主避险能力中的至少一种。
  3. 如权利要求2所述的方法，其中，所述根据障碍物特性分类得到的多种障碍物类型包括：稳定型、协商型、风险型和激变型中的至少二种，其中：
    稳定型障碍物包括至少一种不具备自主移动能力的障碍物；
    协商型障碍物包括至少一种具备自主移动能力和人机交互能力的障碍物；
    风险型障碍物包括至少一种作为被保护对象的障碍物；
    激变型障碍物包括至少一种具备自主避险能力的障碍物。
  4. 如权利要求1所述的方法，所述根据障碍物特性分类得到的多种障碍物类型包括：稳定型、协商型、风险型和激变型中的至少二种，其中：
    稳定型障碍物包括建筑构件和家具中的至少一种；
    协商型障碍物包括成人；
    风险型障碍物包括婴幼儿；
    激变型障碍物包括家庭宠物。
  5. 如权利要求1所述的方法，其中：
    所述地图为室内栅格地图；
    所述构建地图并标记所述障碍物，包括：为不同类型的障碍物构建不同的地图层，根据所述障碍物所属的类型在与所述障碍物所属的类型对应的地图层上标记所述障碍物。
  6. 如权利要求1所述的方法，其中：
    对检测到的障碍物进行识别，根据识别结果确定所述障碍物在根据障碍物特性分类得到的多种障碍物类型中所属的类型，包括：
    对视觉传感器采集的图像信息进行障碍物检测，对检测到的障碍物的图像进行图像识别，根据识别出的所述障碍物在预先设置的所述多种障碍物类型分别包含的障碍物的信息中进行查找，确定所述障碍物所属的类型。
  7. 如权利要求1所述的方法，其中：
    所述多种障碍物类型包括激变型，激变型障碍物包括家庭宠物；
    所述构建地图并标记所述障碍物，包括：在所述障碍物所属的类型为激变型时，构建激变型障碍物地图层并标记所述障碍物，所述激变型障碍物地图层的维护区域小于除所述激变型障碍物地图层之外的地图层的维护区域，所述除所述激变型障碍物地图层之外的地图层设置为标记所述多种障碍物类型中除所述激变型之外的类型的障碍物。
  8. 如权利要求1所述的方法，其中：
    所述多种障碍物类型包括风险型，风险型障碍物包括婴幼儿；
    构建地图标并记所述障碍物，在记录所述障碍物所属的类型之后，所述方法还包括：对所述障碍物的周边进行膨胀处理，在所述障碍物所属的类型为风险型时，在进行所述膨胀处理时使用第一膨胀半径，所述第一膨胀半径大于对所述多种障碍物类型中除所述风险型障碍物之外的类型的障碍物的周边进行所述膨胀处理时使用的膨胀半径。
  9. 如权利要求1-8中任一所述的方法，其中：
    所述地图是包括静态地图层的导航地图，所述静态地图层标记有导航进行之前环境中存在的障碍物；
    所述对检测到的障碍物进行识别，包括：在导航过程中，对在所述环境中检测到的新增的障碍物进行识别；
    所述构建地图并标记所述障碍物，记录所述障碍物所属的类型，包括：构建动态地图层并标记所述新增的障碍物，记录所述新增的障碍物所属的类型；
    构建动态地图并标记所述新增的障碍物之后，所述方法还包括：对障碍物的周边进行膨胀处理，融合所述静态地图层和所述动态地图层，得到更新后的导航主地图。
  10. 如权利要求9中任一所述的方法，其中：
    所述多种障碍物类型包括稳定型，稳定型障碍物包括建筑构件和家具中的至少一种；
    在导航结束之后，所述方法还包括：将在所述动态地图层标记的新增的稳定型障碍物添加到所述静态地图层。
  11. 如权利要求10中任一所述的方法，其中：
    所述构建动态地图层并标记所述新增的障碍物，包括：在所述新增的障碍物为稳定型障碍物时，构建稳定型障碍物地图层并标记所述新增的障碍物；在所述新增的障碍物为除所述稳定型之外的类型的障碍物时，构建除所述稳定型障碍物地图层之外的动态地图层标并记所述新增的障碍物；
    所述将新增的稳定型障碍物添加到所述静态地图层，包括：将在所述稳定型障碍物地图层上标记的所述新增的障碍物添加到所述静态地图层。
  12. 一种导航方法，包括：
    对行进路径上检测到的新增的障碍物进行识别，根据识别结果确定所述新增的障碍物在根据障碍物特性分类得到的多种障碍物类型中所属的类型；
    根据所述新增的障碍物所属的类型进行避障处理。
  13. 如权利要求12所述的方法，其中：
    所述多种障碍物类型包括协商型，协商型障碍物包括成人；
    根据所述新增的障碍物所属的类型进行避障处理，包括：在所述新增的障碍物为协商型障碍物时，与所述协商型障碍物进行交互，请求所述协商型障碍物避让。
  14. 如权利要求13所述的方法，所述与所述协商型障碍物进行交互，请求所述协商型障碍物避让之后，所述方法还包括：
    判断在设定的时间内，所述协商型障碍物是否执行以下至少之一的操作：离开所述行进路径和返回同意避让的响应；基于所述协商型障碍物执行了离开所述行进路径和返回同意避让的响应中至少一种的操作的判断结果，按照所述行进路径继续前进；基于所述协商型障碍物没有执行离开所述行进路径和返回同意避让的响应中至少一种的操作的判断结果，重新规划新的进行路径。
  15. 如权利要求12-14中任一所述的方法，其中：
    所述导航地图为移动机器人导航所使用的室内栅格地图。
  16. 一种地图构建系统，包括：
    障碍物识别单元，设置为对检测到的障碍物进行识别，得到识别结果；
    障碍物分类单元，设置为根据所述识别结果，确定所述障碍物在根据障碍物特性分类得到的多种障碍物类型中所属的类型；
    障碍物处理单元，设置为构建地图并标记所述障碍物，记录所述障碍物所属的类型。
  17. 如权利要求16所述的系统，其中：
    所述障碍物特性包括障碍物的自主移动能力、交互能力、安全性和自主避险能力中的至少一种。
  18. 如权利要求16所述的系统，所述根据障碍物特性分类得到的多种障碍物类型包括：稳定型、协商型、风险型和激变型中的至少二种，其中：
    稳定型障碍物包括建筑构件和家具中的至少一种；
    协商型障碍物包括成人；
    风险型障碍物包括婴幼儿；
    激变型障碍物包括家庭宠物。
  19. 如权利要求16所述的系统，其中：
    所述构建的地图为室内栅格地图；
    所述障碍物处理单元设置为通过以下操作构建地图并标记所述障碍物：为不同类型的障碍物构建不同的地图层，根据所述障碍物所属的类型在对应的地图层上标记所述障碍物。
  20. 如权利要求16所述的系统，所述系统还包括障碍物捕捉单元，设置为根据视觉传感器采集的图像信息进行障碍物检测；
    所述障碍物识别单元设置为通过以下操作对检测到的障碍物进行识别：对所述障碍物捕捉单元检测到的障碍物的图像进行图像识别；
    所述障碍物分类单元设置为通过以下操作确定所述障碍物在根据障碍物特性分类得到的多种障碍物类型中所属的类型：根据所述障碍物识别单元确定的障碍物，在预设的所述多种障碍物类型分别包含的障碍物的信息中进行查找，确定所述障碍物所属的类型。
  21. 如权利要求16-20中任一所述的系统，其中：
    所述多种障碍物类型包括激变型，激变型障碍物包括家庭宠物；
    所述障碍物处理单元设置为通过以下操作构建地图并标记所述障碍物：在所述障碍物所属的类型为激变型时，构建激变型障碍物地图层并标记所述障碍物，所述激变型障碍物地图层的维护区域小于除所述激变型障碍物地图层之外的地图层的维护区域，所述除所述激变型障碍物地图层之外的地图层设置为标记所述多种障碍物类型中除所述激变型之外的类型的障碍物。
  22. 如权利要求16-20中任一所述的系统，其中：
    所述多种障碍物类型包括风险型，风险型障碍物包括婴幼儿；
    所述系统还包括：障碍物膨胀单元，设置为对地图中标记的所述障碍物的周边进行膨胀处理，且在所述障碍物所属的类型为风险型时，在进行所述膨胀处理时使用第一膨胀半径，所述第一膨胀半径大于对所述多种障碍物类型中除所述风险型障碍物之外的类型的障碍物的周边进行所述膨胀处理时使用的膨胀半径。
  23. 如权利要求16-20中任一所述的系统，其中：
    所述地图为包括静态地图层的导航地图，所述静态地图层标记有导航进行之前环境中存在的障碍物；
    所述障碍物识别单元设置为通过以下操作对检测到的障碍物进行识别：在导航过程中，对在所述环境中检测到的新增的障碍物进行识别；
    所述障碍物处理单元设置为通过以下操作构建地图并标记所述障碍物：构建动态地图层并标记所述新增的障碍物；
    所述系统还包括：地图层融合单元，设置为对障碍物的周边进行膨胀处理，融合所述静态地图层和所述动态地图层，得到更新后的导航主地图。
  24. 如权利要求23所述的系统，所述多种障碍物类型包括稳定型，稳定型障碍物包括建筑构件和家具中的至少一种；
    所述系统还包括：障碍物更新单元，设置为在导航结束之后，将所述动态地图层标记的新增的稳定型障碍物添加到所述静态地图层。
  25. 如权利要求23所述的系统，所述地图构建系统还包括：
    导航地图构建接口，设置为接收对所述导航主地图进行更新的外部请求，启动所述障碍物识别单元、所述障碍物分类单元、所述障碍物处理单元和所述地图层融合单元以对所述导航主地图进行更新，并将更新后的所述导航主地图返回给移动机器人业务。
  26. 一种导航系统，包括导航模块和地图构建模块，其中：
    所述导航模块，设置为规划行进路径，通知所述地图构建模块启动导航地图更新任务；
    所述地图构建模块，设置为对所述行进路径上检测到的新增的障碍物进行识别，根据识别结果确定所述新增的障碍物在根据障碍物特性分类得到的多种障碍物类型中所属的类型，更新所述导航地图以标记所述新增的障碍物，并记 录所述新增的障碍物所属的类型，并将更新后的导航地图返回给所述导航模块；
    所述导航模块还设置为接收所述更新后的导航地图，根据所述新增的障碍物所属的类型进行避障处理。
  27. 如权利要求26所述的系统，其中，
    所述多种障碍物类型包括协商型，协商型障碍物包括成人；
    所述系统还包括人机交互模块；
    所述导航模块设置为通过以下操作来根据所述新增的障碍物所属的类型进行避障处理：在所述新增的障碍物为协商型障碍物时，通过所述人机交互模块与所述协商型障碍物进行交互，请求所述协商型障碍物避让。
  28. 如权利要求27所述的系统，所述导航模块还设置为在所述人机交互模块与所述协商型障碍物进行交互，请求所述协商型障碍物避让之后，
    判断在设定的时间内，所述协商型障碍物是否执行以下至少之一的操作：离开所述行进路径和返回同意避让的响应；基于所述协商型障碍物执行了离开所述行进路径和返回同意避让的响应中至少一种的操作的判断结果，保持所述行进路径不变；基于所述协商型障碍物没有执行离开所述行进路径和返回同意避让的响应中至少一种的操作的判断结果，重新规划新的进行路径。
  29. 一种地图构建装置，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，所述处理器执行所述程序时实现如权利要求1-11所述的地图构建方法。
  30. 一种导航装置，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，所述处理器执行所述程序时实现如权利要求13-15所述的导航方法。
  31. 一种计算机可读存储介质，存储有计算机可执行指令，所述计算机可执行指令设置为执行权利要求1-11任一项所述的地图构建方法。
  32. 一种计算机可读存储介质，存储有计算机可执行指令，所述计算机可执行指令设置为执行权利要求13-15任一项所述的导航方法。
PCT/CN2018/123164 2017-12-29 2018-12-24 一种地图构建、导航方法及装置、系统 Ceased WO2019128933A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP18893502.7A EP3734225A4 (en) 2017-12-29 2018-12-24 Map construction and navigation method, and device and system
US16/958,855 US11359930B2 (en) 2017-12-29 2018-12-24 Map construction and navigation method, and device and system

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201711499191.0 2017-12-29
CN201711499191.0A CN108344414A (zh) 2017-12-29 2017-12-29 一种地图构建、导航方法及装置、系统

Publications (1)

Publication Number Publication Date
WO2019128933A1 true WO2019128933A1 (zh) 2019-07-04

Family

ID=62961303

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/123164 Ceased WO2019128933A1 (zh) 2017-12-29 2018-12-24 一种地图构建、导航方法及装置、系统

Country Status (4)

Country Link
US (1) US11359930B2 (zh)
EP (1) EP3734225A4 (zh)
CN (1) CN108344414A (zh)
WO (1) WO2019128933A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111060116A (zh) * 2019-12-04 2020-04-24 江西洪都航空工业集团有限责任公司 一种基于视觉的草场自主建图系统
CN112150491A (zh) * 2020-09-30 2020-12-29 小狗电器互联网科技（北京）股份有限公司 图像检测方法、装置、电子设备和计算机可读介质

Families Citing this family (62)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108344414A (zh) * 2017-12-29 2018-07-31 中兴通讯股份有限公司 一种地图构建、导航方法及装置、系统
CN109238296B (zh) * 2018-08-16 2022-03-22 佛山科学技术学院 一种高精度室内导航方法及装置
CN110895409B (zh) * 2018-08-23 2020-11-27 珠海格力电器股份有限公司 避让障碍物的控制方法
CN110928282A (zh) * 2018-09-17 2020-03-27 格力电器（武汉）有限公司 清扫机器人的控制方法及装置
CN108827336A (zh) * 2018-09-26 2018-11-16 广东工业大学 一种基于无人驾驶的路径规划方法、装置及设备
CN110968083B (zh) * 2018-09-30 2023-02-10 科沃斯机器人股份有限公司 栅格地图的构建方法、避障的方法、设备及介质
CN109568093A (zh) * 2018-11-27 2019-04-05 湖北经济学院 一种步行安全综合管理系统及方法
CN109655073B (zh) * 2018-12-06 2021-07-23 宁波吉利汽车研究开发有限公司 一种在无或弱信号区域中的地图绘制方法、装置及车辆
CN111443695A (zh) * 2018-12-28 2020-07-24 珠海市一微半导体有限公司 扫地机器人控制方法、装置、存储介质及扫地机器人
CN111374607A (zh) * 2018-12-29 2020-07-07 尚科宁家(中国)科技有限公司 一种基于扫地机器人的目标识别方法及装置、设备、介质
CN109602343A (zh) * 2019-01-10 2019-04-12 珠海格力电器股份有限公司 基于微波雷达的扫地机器人的控制方法及装置、系统
CN111624989B (zh) * 2019-02-28 2023-05-26 科沃斯机器人股份有限公司 控制方法及自移动设备
CN110081893B (zh) * 2019-04-01 2020-09-25 东莞理工学院 一种基于策略重用和强化学习的导航路径规划方法
US11429110B1 (en) * 2019-05-24 2022-08-30 Amazon Technologies, Inc. System for obstacle avoidance by autonomous mobile device
CN110164288A (zh) * 2019-06-04 2019-08-23 浙江大学昆山创新中心 一种基于自建图的静态地图在线更新方法和装置
US20220343241A1 (en) * 2019-06-21 2022-10-27 Intel Corporation Technologies for enabling collective perception in vehicular networks
CN112256005B (zh) * 2019-07-02 2022-11-15 广州哨马智能装备科技有限公司 基于超声波传感器的避障系统和方法以及机器人
CN110522359B (zh) * 2019-09-03 2021-09-03 深圳飞科机器人有限公司 清洁机器人以及清洁机器人的控制方法
CN110658819B (zh) * 2019-09-30 2022-04-15 北京猎户星空科技有限公司 一种避障方法、装置、电子设备和存储介质
CN110764110B (zh) * 2019-11-12 2022-04-08 深圳创维数字技术有限公司 路径导航方法、装置及计算机可读存储介质
CN110736465B (zh) * 2019-11-15 2021-01-08 北京云迹科技有限公司 导航方法、装置、机器人及计算机可读存储介质
CN111338384B (zh) * 2019-12-17 2021-06-08 北京化工大学 一种仿蛇机器人自适应路径跟踪方法
CN111338336B (zh) * 2020-02-11 2021-07-13 腾讯科技（深圳）有限公司 一种自动驾驶方法及装置
CN111481105A (zh) * 2020-04-20 2020-08-04 北京石头世纪科技股份有限公司 一种自行走机器人避障方法、装置、机器人和存储介质
CN111631650B (zh) * 2020-06-05 2021-12-03 上海黑眸智能科技有限责任公司 基于障碍物高度检测的室内平面图生成方法、系统、终端以及扫地机器人
CN111829526B (zh) * 2020-07-23 2022-05-10 中国人民解放军国防科技大学 一种基于防撞半径的距离地图重构与跳点路径规划方法
CN111966097B (zh) * 2020-08-12 2024-06-14 深圳华芯信息技术股份有限公司 基于栅格地图区域化探索的建图方法、系统以及终端
CN112161624A (zh) * 2020-09-11 2021-01-01 上海高仙自动化科技发展有限公司 标记方法、标记装置、智能机器人及可读存储介质
CN112150490B (zh) * 2020-09-30 2024-02-02 北京小狗吸尘器集团股份有限公司 图像检测方法、装置、电子设备和计算机可读介质
CN112197763B (zh) * 2020-09-30 2024-02-27 灵动科技（北京）有限公司 地图构建方法、装置、设备和存储介质
CN112462768B (zh) * 2020-11-25 2024-03-29 深圳拓邦股份有限公司 移动机器人导航地图的创建方法、装置及移动机器人
CN112506204B (zh) * 2020-12-17 2022-12-30 深圳市普渡科技有限公司 机器人遇障处理方法、装置、设备和计算机可读存储介质
CN114691701B (zh) * 2020-12-31 2026-04-24 深圳引望智能技术有限公司 高精度地图更新方法及装置
US11960282B2 (en) * 2021-01-05 2024-04-16 Abb Schweiz Ag Systems and methods for servicing a data center using autonomous vehicle
CN112904841B (zh) * 2021-01-12 2023-11-03 北京布科思科技有限公司 非水平朝向的单线定位避障方法和装置、设备及存储介质
CN112882470A (zh) * 2021-01-14 2021-06-01 中广核工程有限公司 一种核电站试验机器人及试验方法
JP7521442B2 (ja) * 2021-02-01 2024-07-24 トヨタ自動車株式会社 自律移動システム、自律移動方法及び自律移動プログラム
KR20220121483A (ko) * 2021-02-25 2022-09-01 엘지전자 주식회사 지능적 맵 생성 방법 및 그 이동 로봇
CN113156949B (zh) * 2021-04-19 2024-09-10 深圳市云鼠科技开发有限公司 机器人航向角纠偏方法、装置、计算机设备及存储介质
CN113390417A (zh) * 2021-06-25 2021-09-14 深圳市优必选科技股份有限公司 机器人及其导航方法、装置和计算机可读存储介质
CN115687541A (zh) * 2021-07-21 2023-02-03 华为技术有限公司 一种地图、地图生成方法及装置
CN113485368B (zh) * 2021-08-09 2024-06-07 国电南瑞科技股份有限公司 一种架空输电线路巡线机器人导航、巡线方法及装置
CN113670292B (zh) * 2021-08-10 2023-10-20 追觅创新科技(苏州)有限公司 地图的绘制方法和装置、扫地机、存储介质、电子装置
CN113867349B (zh) * 2021-09-28 2024-04-09 浙江大华技术股份有限公司 一种机器人的避障方法、系统及智能机器人
CN114003036A (zh) * 2021-10-28 2022-02-01 广州赛特智能科技有限公司 机器人避障控制方法、装置、设备和介质
CN114296447B (zh) * 2021-12-07 2023-06-30 北京石头世纪科技股份有限公司 自行走设备的控制方法、装置、自行走设备和存储介质
CN114185353B (zh) * 2021-12-09 2024-03-22 吉林大学 一种自主作业装载机倒车避障与路径规划方法
CN114415659B (zh) * 2021-12-13 2024-05-28 烟台杰瑞石油服务集团股份有限公司 一种机器人安全避障方法、装置、机器人及存储介质
DE102021133614A1 (de) * 2021-12-17 2023-06-22 Still Gesellschaft Mit Beschränkter Haftung Verfahren zur Erzeugung einer Umgebungskarte für einen mobilen Logistik-Roboter und mobiler Logistik-Roboter
CN116473471A (zh) * 2022-01-14 2023-07-25 追觅创新科技(苏州)有限公司 防跌落控制方法、系统及自动清洁设备
CN114637287B (zh) * 2022-02-24 2025-01-14 深圳市优必选科技股份有限公司 路径生成方法、装置、机器人及计算机可读存储介质
CN115736733B (zh) * 2022-10-28 2025-10-28 广东盈峰智能环卫科技有限公司 室内机器人及其地图构建方法和装置
CN115755935A (zh) * 2022-12-16 2023-03-07 深圳拓邦股份有限公司 一种室内地图障碍物填充方法
CN115629612A (zh) * 2022-12-19 2023-01-20 科大讯飞股份有限公司 一种避障方法、装置、设备及存储介质
CN116798182A (zh) * 2022-12-23 2023-09-22 慧之安信息技术股份有限公司 基于动态视觉的儿童活动监测方法和系统
CN116380037A (zh) * 2023-02-24 2023-07-04 浙江安吉智电控股有限公司 局部代价地图更新方法、装置
CN116358528A (zh) * 2023-03-24 2023-06-30 深圳市正浩创新科技股份有限公司 地图更新方法、地图更新装置、自移动设备及存储介质
CN116466723B (zh) * 2023-04-26 2025-03-14 曲阜师范大学 一种消杀机器人避障方法、系统和设备
CN116862938B (zh) * 2023-07-13 2025-10-31 芜湖云从科技有限公司 矢量图形获取方法、计算机设备及计算机可读存储介质
US12588670B2 (en) * 2023-09-12 2026-03-31 Tobacco Research Institute of CAAS Automatic obstacle avoidance method and system of pesticide application robot and storage medium
WO2025067332A1 (zh) * 2023-09-28 2025-04-03 美智纵横科技有限责任公司 清洁设备的控制方法、装置及存储介质、控制器、设备
CN117216080A (zh) * 2023-09-28 2023-12-12 云鲸智能(深圳)有限公司 地图更新方法、清洁方法、清洁机器人及存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104808671A (zh) * 2015-05-19 2015-07-29 东南大学 一种家居环境下的机器人路径规划方法
CN105074602A (zh) * 2013-03-29 2015-11-18 株式会社日立产机系统 位置识别装置和具有该位置识别装置的移动机器人
CN105182979A (zh) * 2015-09-23 2015-12-23 上海物景智能科技有限公司 一种移动机器人障碍物检测及避让方法和系统
CN108344414A (zh) * 2017-12-29 2018-07-31 中兴通讯股份有限公司 一种地图构建、导航方法及装置、系统

Family Cites Families (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8989972B2 (en) * 2008-09-11 2015-03-24 Deere & Company Leader-follower fully-autonomous vehicle with operator on side
JP5891553B2 (ja) * 2011-03-01 2016-03-23 株式会社国際電気通信基礎技術研究所 ルートパースペクティブモデル構築方法およびロボット
CN103984037B (zh) * 2014-04-30 2017-07-28 深圳市墨克瑞光电子研究院 基于视觉的移动机器人障碍物检测方法和装置
CN104615138B (zh) * 2015-01-14 2017-09-08 上海物景智能科技有限公司 一种划分移动机器人室内区域动态覆盖方法及其装置
KR102521493B1 (ko) * 2015-10-27 2023-04-14 삼성전자주식회사 청소 로봇 및 그 제어방법
US9612123B1 (en) * 2015-11-04 2017-04-04 Zoox, Inc. Adaptive mapping to navigate autonomous vehicles responsive to physical environment changes
US10265859B2 (en) * 2016-02-09 2019-04-23 Cobalt Robotics Inc. Mobile robot with removable fabric panels
US10816654B2 (en) * 2016-04-22 2020-10-27 Huawei Technologies Co., Ltd. Systems and methods for radar-based localization
EP3453008A1 (en) * 2016-05-06 2019-03-13 PCMS Holdings, Inc. Method and system for collaborative sensing for updating dynamic map layers
CN106338996B (zh) * 2016-10-20 2019-05-31 上海物景智能科技有限公司 一种移动机器人安全控制的方法及系统
CN106530946A (zh) * 2016-11-30 2017-03-22 北京贝虎机器人技术有限公司 用于编辑室内地图的方法及装置
KR102012550B1 (ko) * 2017-02-20 2019-08-20 엘지전자 주식회사 돌발 장애물을 식별하는 방법 및 이를 구현하는 로봇
US10907974B2 (en) * 2017-04-17 2021-02-02 Cisco Technology, Inc. Real-time updates to maps for autonomous navigation
WO2018218146A1 (en) * 2017-05-26 2018-11-29 Google Llc Data layers for a vehicle map service
CN107256019B (zh) * 2017-06-23 2018-10-19 杭州九阳小家电有限公司 一种清洁机器人的路径规划方法
JP7062892B2 (ja) * 2017-07-13 2022-05-09 トヨタ自動車株式会社 ダイナミックマップ更新装置、ダイナミックマップ更新方法、ダイナミックマップ更新プログラム
US11562524B2 (en) 2017-10-31 2023-01-24 Hewlett-Packard Development Company, L.P. Mobile robots to generate occupancy maps
US10606269B2 (en) * 2017-12-19 2020-03-31 X Development Llc Semantic obstacle recognition for path planning

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105074602A (zh) * 2013-03-29 2015-11-18 株式会社日立产机系统 位置识别装置和具有该位置识别装置的移动机器人
CN104808671A (zh) * 2015-05-19 2015-07-29 东南大学 一种家居环境下的机器人路径规划方法
CN105182979A (zh) * 2015-09-23 2015-12-23 上海物景智能科技有限公司 一种移动机器人障碍物检测及避让方法和系统
CN108344414A (zh) * 2017-12-29 2018-07-31 中兴通讯股份有限公司 一种地图构建、导航方法及装置、系统

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3734225A4

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111060116A (zh) * 2019-12-04 2020-04-24 江西洪都航空工业集团有限责任公司 一种基于视觉的草场自主建图系统
CN112150491A (zh) * 2020-09-30 2020-12-29 小狗电器互联网科技（北京）股份有限公司 图像检测方法、装置、电子设备和计算机可读介质
CN112150491B (zh) * 2020-09-30 2023-08-18 北京小狗吸尘器集团股份有限公司 图像检测方法、装置、电子设备和计算机可读介质

Also Published As

Publication number Publication date
US20200340826A1 (en) 2020-10-29
EP3734225A1 (en) 2020-11-04
EP3734225A4 (en) 2022-01-26
US11359930B2 (en) 2022-06-14
CN108344414A (zh) 2018-07-31

Similar Documents

Publication Publication Date Title
WO2019128933A1 (zh) 一种地图构建、导航方法及装置、系统
US12001210B2 (en) Semantic obstacle recognition for path planning
KR102577785B1 (ko) 청소 로봇 및 그의 태스크 수행 방법
KR20240063820A (ko) 청소 로봇 및 그의 태스크 수행 방법
WO2021103987A1 (zh) 扫地机器人控制方法、扫地机器人及存储介质
CN109108973B (zh) 单向式路径调度方法及系统
CN105955258B (zh) 基于Kinect传感器信息融合的机器人全局栅格地图构建方法
CN112526993A (zh) 栅格地图更新方法、装置、机器人及存储介质
CN108062098B (zh) 智能机器人的地图构建方法和系统
CN111609852A (zh) 语义地图构建方法、扫地机器人及电子设备
CN110858098A (zh) 使用人-机器人交互的自驱动移动机器人
CN115639825A (zh) 机器人的避障方法和系统
Vardy et al. Cache consensus: rapid object sorting by a robotic swarm
CN113432610A (zh) 机器人通行规划方法、装置、机器人及存储介质
CN115326058A (zh) 机器人地图的生成方法、设备、存储介质和移动机器人
Ghani et al. Detecting negative obstacle using Kinect sensor
EP3795225B1 (en) Control device, control method, and program
KR20230134109A (ko) 청소 로봇 및 그의 태스크 수행 방법
CN110084825A (zh) 一种基于图像边缘信息导航的方法及系统
CN113420698B (zh) 一种基于机器人的环境识别方法及装置
CN116382308A (zh) 智能化移动机械自主寻径与避障方法、装置、设备及介质
US12504765B2 (en) Artificial intelligence (AI)-based system for autonomous navigation of robotic devices in dynamic human-centric environments and method thereof
CN111552286B (zh) 一种机器人及其移动控制方法和装置
KR20250066365A (ko) 라이다 기반 매핑 시 불필요 영역 필터링 시스템 및 그 방법
CN119596918A (zh) 用于自移动机器人的避障方法、系统及自移动机器人

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18893502

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2018893502

Country of ref document: EP

Effective date: 20200729

WWW Wipo information: withdrawn in national office

Ref document number: 2018893502

Country of ref document: EP