WO2023199846A1 - システム - Google Patents
システム Download PDFInfo
- Publication number
- WO2023199846A1 WO2023199846A1 PCT/JP2023/014300 JP2023014300W WO2023199846A1 WO 2023199846 A1 WO2023199846 A1 WO 2023199846A1 JP 2023014300 W JP2023014300 W JP 2023014300W WO 2023199846 A1 WO2023199846 A1 WO 2023199846A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- information
- layer
- learning
- agent
- agents
- 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
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/092—Reinforcement learning
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/098—Distributed learning, e.g. federated learning
Definitions
- the present invention relates to a system.
- Patent Document 1 describes multi-agent reinforcement learning. [Prior art documents] [Patent document] [Patent Document 1] Japanese Patent Application Publication No. 2004-227208
- a system may include a first layer agent of a first layer and a plurality of second layer agents of a second layer lower than the first layer, which are arranged in a network.
- Each of the plurality of second layer agents uses the collected information to perform learning in cooperation with other second layer agents, and transmits information generated using the learning results to the first layer agent. It's fine.
- the first layer agent may perform learning using information received from the plurality of second layer agents.
- the system may further include a plurality of third layer agents of a third layer lower than the second layer. Each of the plurality of third layer agents executes learning in cooperation with other third layer agents using the collected information, and the information generated using the learning results is transmitted to the plurality of second layer agents.
- Each of the plurality of third layer agents learns in cooperation with the other plurality of third layer agents using a third information collection unit that collects information and the information collected by the third information collection unit. and a third information transmitting unit that transmits information generated using the learning results by the third learning executing unit to at least one of the plurality of second layer agents. good.
- Each of the plurality of second layer agents includes a second information collection unit that collects information from the plurality of third layer agents, and uses the information collected by the second information collection unit to collect information from the other plurality of third layer agents.
- the first layer agent includes a first information collection section that collects information from the plurality of second layer agents, and a first learning execution section that executes learning using the information collected by the first information collection section. may have.
- the first layer agent may be placed in a first NW layer of a hierarchical network, and the plurality of second layer agents may be placed in a second NW layer lower than the first NW layer of the layered network.
- the plurality of third layer agents may be placed in a third network layer lower than the second network layer of the hierarchical network.
- the hierarchical network may be a cloud network, the first NW tier may be configured by cloud computing, the second NW tier may be configured by a plurality of fog computing, and the third NW tier may be configured by cloud computing. , may be configured by multiple edge computing systems.
- the hierarchical network may be a cloud network, the first NW tier may be configured by cloud computing, the second NW tier may be configured by a plurality of fog computing, and the third NW tier may be configured by cloud computing.
- the third information collection unit may collect information from a plurality of IoT (Internet of Things) devices by mobile communication, and the third information transmission unit may be configured by a plurality of edge computing units.
- At least one of the plurality of second layer agents selects information selected from the information collected by the third information collecting unit or information generated using the learning results.
- the second information collecting unit may collect the information transmitted by the third information transmitting unit, and the second information transmitting unit may use the learning result by the second learning execution unit. Then, information selected from the information collected by the second information collection unit or information generated using the learning result may be transmitted to the first layer agent.
- each of the plurality of third layer agents includes a third processing execution unit that executes a process on the information collected by the third information collection unit using learning results by the third learning execution unit. may have.
- the third processing execution unit may generate information to be sent to at least one of the plurality of second layer agents using the plurality of pieces of information collected by the third information collection unit.
- the third processing execution unit may generate transmission information including information selected from a plurality of pieces of information collected by the third information collection unit, using a learning result by the third learning execution unit.
- the third processing execution unit may generate summary information that summarizes a plurality of pieces of information collected by the third information collection unit, using a learning result by the third learning execution unit.
- the third information transmitting unit may transmit the information generated by the third processing execution unit to at least one of the plurality of second layer agents.
- each of the plurality of second layer agents includes a second processing execution unit that executes processing on the information collected by the second information collection unit using learning results by the second learning execution unit. may have.
- the second processing execution unit may generate information to be sent to the first layer agent using a plurality of pieces of information collected by the second information collection unit.
- the second processing execution unit may generate transmission information including information selected from a plurality of pieces of information collected by the second information collection unit, using a learning result by the second learning execution unit.
- the second processing execution unit may generate summary information that summarizes a plurality of pieces of information collected by the second information collection unit, using the learning result by the second learning execution unit.
- the second information transmitting unit may transmit the information generated by the second processing execution unit to the first layer agent.
- the first layer agent includes a first processing execution unit that executes processing on information collected by the first information collection unit using learning results by the first learning execution unit. good.
- the first processing execution unit may use the learning result by the first learning execution unit to execute processing aimed at stabilizing the entire network to which the system is applied.
- the first processing execution unit is configured to analyze at least one of the plurality of second tier agents or at least one of the plurality of third tier agents based on the result of analyzing the information collected by the first information collection unit. You may also generate instruction information for.
- the first processing execution unit instructs at least one of the plurality of third layer agents to send information to at least one of the plurality of second layer agents out of the information collected from the IoT device.
- Instruction information may be generated.
- the first processing execution unit generates instruction information that instructs at least one of the plurality of second-tier agents to process the information collected from at least one of the plurality of second-tier agents. good.
- the first processing execution unit may generate learning information to be transmitted to at least one of the plurality of third layer agents based on a result of analyzing the information collected by the first information collection unit.
- the first processing execution unit may generate learning information to be transmitted to at least one of the plurality of second layer agents based on a result of analyzing the information collected by the first information collection unit.
- the first processing execution unit may generate learning information including a reward set in accordance with a change in tendency of the information collected by the first information collection unit.
- the plurality of third layer agents perform analysis that requires more real-time performance, and the first layer agent: Analyzing trends over a fairly long period of time, the plurality of second layer agents may perform a corresponding analysis during that period.
- the plurality of third-tier agents collect image data and object detection data from a plurality of IoT devices placed in the area.
- third-tier agents in charge of geographically adjacent subareas execute learning to detect the occurrence of accidents while sharing information, and the third-tier agents use the learning results to , detects the occurrence of an accident, and a plurality of second-tier agents are assigned to a group of subareas, collect information from third-tier agents that collect information from IoT devices in the subareas of the group, and The second layer agents perform learning to predict the occurrence of accidents by cooperating with each other, the second layer agents predict the occurrence of accidents using the learning results, and the first layer agents Based on the prediction result by the second layer agent, learning for controlling the overall information may be performed.
- the first layer agent increases the amount of information collected and increases the types of information for subareas where accidents are expected to occur, and reduces the amount of information collected for other subareas, and increases the types of information collected. Control may be performed to reduce the number of types.
- the third layer agent may route the message. It's fine.
- the third layer agent may cooperate with other third layer agents to perform learning to control Topic, To (including Copy), and division of messages from publishers. Using the learning results, the third layer agent cooperates with other third layer agents to generate Topic, determine To, and send the message so that the message from the publisher reaches the appropriate subscriber. May control replication, message splitting, etc.
- the second layer agent monitors the routing by the third layer agent based on the information collected from the third layer agent, and performs learning so that it can perform processing to resolve problems that occur in the routing. good.
- the second layer agent buffers the message when there is a message with the same To with an unknown destination, and sends a part of the buffered message to the network when the amount of buffering exceeds a threshold. As a result, if the message comes back again, it will be buffered for a certain period of time until the message arrives, and when the message starts to arrive, it will send the message with the same To among the buffered messages. You may begin.
- the first layer agent may perform learning to detect problems in the network using information collected from the second layer agent. When the first layer agent detects a problem in the network using the learning results, it may notify the network operator of the detection results or output an instruction to change the configuration of the network.
- each of the multiple third-tier agents is installed in a vehicle located in each region divided into multiple regions.
- Information may be collected from IoT devices installed in traffic lights, IoT devices installed in traffic lights, etc., and IoT devices installed on roads, etc.
- the third layer agent collects vehicle location information, images captured by vehicle cameras and street cameras, vehicle detection information and human detection information detected by road sensors, vehicles detected by vehicle sensors, etc. At least one of the following information may be collected: distance between vehicles, weather information in the area, vehicle navigation information, and vehicle travel speed.
- the third layer agent is in charge of a certain area, collects information from the IoT device installed in the vehicle from the time a vehicle enters the area, and collects information from the IoT device installed in the vehicle when the vehicle leaves the area. In some cases, the third layer agent in charge of the area where the vehicle enters may take over the collection of information.
- the third layer lower agent shares information with other third layer agents to detect whether the distance between vehicles is shorter than a threshold or a danger such as a vehicle colliding with a person. good. When the third layer agent detects danger, it may transmit danger detection information to the target vehicle.
- the second tier agent may instruct the third tier agent what information to communicate.
- the second layer agent analyzes (learns) important information (many people, many cars, etc.) based on the information (day of the week, time, etc.) collected from the third layer agent in the past, and prioritizes collection. It may be possible to make it possible to determine information with a high degree of accuracy.
- the second layer agent may instruct the third layer agent to collect high priority information for each period.
- the second layer agent may be able to identify high priority information with high accuracy by learning in cooperation with other second layer agents. As a result of the learning, the second layer agent instructs the vehicle detection results to have a higher priority so that information can be collected on weekday mornings, and to increase the priority for the human detection results during the afternoon on holidays. It may be possible to provide instructions so that information can be collected.
- the first tier agent may perform a broader range of analysis.
- the first layer agent performs prediction several steps ahead (at a set time) based on traffic information (traffic volume, time of day, day of the week, event, weather) for the area in charge (city, prefecture, country), and Information may be provided to second-tier agents.
- traffic information traffic volume, time of day, day of the week, event, weather
- Information may be provided to second-tier agents.
- An example of a network 100 to which an autonomous decentralized system is applied is schematically shown.
- An example of a cloud network 300, which is an example of the network 100, is schematically shown.
- An example of an autonomous decentralized system 200 is schematically shown.
- An example of the functional configuration of the lower-level agent 230 is schematically shown.
- An example of the functional configuration of the intermediate agent 220 is schematically shown.
- An example of the functional configuration of the upper agent 210 is schematically shown.
- Another example of the autonomous decentralized system 200 is schematically shown.
- An example of the hardware configuration of a computer 1200 that implements an upper agent 210, an intermediate agent 220, or a lower agent 230 is schematically shown.
- An autonomous decentralized system which is one type of multi-agent system, is known.
- each of the multiple agents aims to independently learn autonomously and acquire cooperative behavior.
- a state in which each agent is exactly the same (structure and parameters) is called homogeneous, and a state in which each agent becomes different from homogeneous (heterogenious) is called functional differentiation.
- an autonomous distributed system there is a target and multiple hunters, and the multiple hunters learn how to capture the target. In other words, multiple hunter agents learn their own roles and cooperate to capture the target. There are examples like the catch-and-trace problem.
- an autonomous distributed system a group of agents with the same performance (homo) differentiates in function through learning, and each agent comes to have a role (hetero). For example, in the tracking problem, agents are divided into agents that chase straight and agents that go around. By using such technology, for example, it is possible to autonomously perform learning that adapts to the environment using the same algorithm, or to perform autonomous and distributed learning using multi-agents.
- an autonomous decentralized system is applied to the network.
- FIG. 1 schematically shows an example of a communication network 100 to which an autonomous decentralized system is applied.
- Communication network 100 is a hierarchical network.
- the communication network 100 is composed of an upper computing 110, multiple intermediate computing 120, and multiple lower computing 130.
- the upper computing 110 constitutes a first NW (NetWork) tier
- the plurality of intermediate computing 120 constitutes a second NW tier lower than the first NW tier
- the plurality of lower computing 130 constitutes a second NW tier lower than the second NW tier.
- a lower third NW layer is also configured.
- the communication network 100 to which the autonomous decentralized system is applied may be any communication network as long as it has a hierarchical structure.
- communication network 100 is a cloud network.
- FIG. 2 schematically shows an example of a cloud network 300, which is an example of the communication network 100.
- the cloud network 300 is configured by cloud computing 310, multiple fog computing units 320, and multiple edge computing units 330.
- Cloud computing 310 may be an example of upper level computing 110.
- Fog computing 320 may be an example of intermediate computing 120.
- Edge computing 330 may be an example of lower level computing 130.
- the first NW tier is configured by cloud computing 310
- the second NW tier is configured by multiple fog computing 320
- the third NW tier is configured by multiple edge computing 330.
- Each of the plurality of edge computing devices 330 communicates with one or more IoT devices 400 via a mobile communication system.
- Each of the plurality of IoT devices 400 may transmit information to at least one of the plurality of edge computing devices 330 via a wireless base station, a Wi-Fi (registered trademark) access point, or the like.
- the IoT device 400 may be any device as long as it can acquire and transmit some information.
- the IoT device 400 includes, for example, various sensors. Examples of information transmitted by the IoT device 400 include image data (still images, moving images), sound data, infrared data, position data, object detection data, distance data, weather data, temperature data, humidity data, etc. This is just an example, and any information may be used.
- the mobile communication system is, for example, a 5G (5th Generation) communication system.
- the mobile communication system may be an LTE (Long Term Evolution) communication system.
- the mobile communication system may be a 3G (3rd Generation) communication system.
- the mobile communication system may be a 6G (6th Generation) communication system or later.
- FIG. 3 schematically shows an example of an autonomous decentralized system 200.
- the autonomous decentralized system 200 is composed of a plurality of agents having a hierarchical structure.
- the autonomous decentralized system 200 includes a higher-level agent 210, multiple intermediate agents 220, and multiple lower-level agents 230.
- the autonomous decentralized system 200 is applied to the communication network 100.
- the upper level agent 210 is located at the upper level computing 110
- each of the multiple intermediate agents 220 is located at each of the multiple intermediate computing 120
- each of the multiple lower level agents 230 is located at each of the multiple lower level computing 110. 130 respectively.
- a plurality of intermediate agents 220 may be arranged for one intermediate computing 120.
- a plurality of lower-level agents 230 may be placed in one lower-level computing 130.
- FIG. 2 illustrates a case where the autonomous decentralized system 200 is configured with three layers, the present invention is not limited to this.
- the autonomous decentralized system 200 may have two hierarchies, or may have four or more hierarchies.
- the autonomous decentralized system 200 may include a higher level agent 210 and a plurality of lower level agents 230.
- the upper agent 210 may be an example of a first tier agent of the first tier
- the lower agent 230 may be an example of a plurality of second tier agents of the second tier.
- the upper level agent 210 is placed in the higher level computing 110
- each of the plurality of lower level agents 230 is placed in each of the plurality of lower level computing 130.
- the upper level agent 210 is placed in one of the plurality of intermediate computing units 120, and the plurality of lower level agents 230 are arranged in each of the plurality of lower level computing units 130.
- the communication network 100 may be configured with two layers without including the plurality of intermediate computing devices 120.
- the autonomous decentralized system 200 is applied to a cloud network 300, for example.
- the upper level agent 210 is placed in a cloud computing 310
- each of the plurality of intermediate agents 220 is placed in each of the plurality of fog computing 320
- each of the plurality of lower level agents 230 is placed in each of the plurality of edge computing 330 respectively.
- Multiple intermediate agents 220 may be deployed for one fog computing 320.
- a plurality of lower-level agents 230 may be placed in one edge computing 330.
- the upper agent 210 is placed in the cloud computing 310, and each of the plurality of lower agents 230 is placed in each of the plurality of edge computing 330. Further, for example, the upper level agent 210 is placed in one of the plurality of fog computing units 320, and the plurality of lower level agents 230 are placed in each of the plurality of edge computing units 330. Note that in this case, the cloud network 300 may be configured with two layers without including the plurality of fog computing devices 320.
- the lower agent 230 collects information.
- Lower agent 230 collects information sent by IoT device 400, for example.
- the lower-level agent 230 may collect information from the IoT device 400 through mobile communication.
- Lower level agent 230 may collect information sent by any device other than IoT device 400.
- the lower-level agent 230 may collect information from any device other than the IoT device 400 through mobile communication.
- the lower-level agent 230 uses the collected information to perform learning in cooperation with other lower-level agents 230.
- the lower agent 230 may perform learning according to pre-registered rewards.
- the lower agent 230 may perform learning according to knowledge registered in advance.
- the lower agent 230 may perform various processes using the learning results. For example, the lower-level agent 230 uses the learning results to select information to be transmitted to the intermediate agent 220 from among the plurality of pieces of collected information, and transmits the selected information to the intermediate agent 220.
- the lower agent 230 uses the learning results to generate summary information that summarizes a plurality of pieces of collected information and sends it to the intermediate agent 220.
- the lower agent 230 collects it from the IoT devices 400 and sends it to the intermediate agent 222 or The usefulness of the information sent to the higher-level agent 210 is registered as a reward.
- Each of the plurality of lower-level agents 230 autonomously advances learning while cooperating with other lower-level agents 230 to maximize the reward.
- the plurality of lower-level agents 230 collect information from one or more different IoT devices 400, and each selects information to be sent to the upper layer from the collected information or sends the collected information to the upper layer based on different criteria.
- a system can be constructed such that summary information is generated, and only information that is highly useful overall is transmitted to upper layers.
- the lower agent 230 may execute processing with relatively high real-time performance.
- the lower-level agent 230 uses the learning results to perform selection processing or summarize information collected from one or more IoT devices 400 at predetermined intervals such as 1 minute and 3 minutes. Performs information generation processing and sends the information to the upper layer. This makes it possible to adjust the information sent to the upper layer in real time.
- a plurality of lower-level agents 230 placed in an edge computing 330 that handles information that includes images during the day and infrared data at night
- many lower-level agents 230 handle images during the day and infrared data at night. Becomes in charge of.
- the lower-level agent 230 can dynamically adapt to the environment, for example, the IoT device 400 to be collected may be changed, or the type of information collected by the IoT device 400 to be collected may be changed. However, it is still applicable. In this way, the lower agent 230 may execute processing in real time, but may also transmit statistical information or summary information with a time delay. Lower agent 230 may be able to intentionally insert a delay.
- the intermediate agent 220 collects information.
- Intermediate agent 220 may collect information from subordinate agents 230.
- the intermediate agent 220 collects from the lower agent 230 the information that the lower agent 230 has collected from the IoT device 400.
- the intermediate agent 220 collects from the lower-level agent 230 information selected by the lower-level agent 230 from among the plurality of pieces of information that the lower-level agent 230 has collected from the IoT device 400 .
- the intermediate agent 220 collects from the lower agent 230 summary information that is a compilation of multiple pieces of information that the lower agent 230 has collected from the IoT device 400 .
- the intermediate agent 220 uses the collected information to perform learning in cooperation with other intermediate agents 220.
- the intermediate agent 220 may perform various processes using the learning results. For example, the intermediate agent 220 uses the learning results to select information to be transmitted to the intermediate agent 220 from among the plurality of pieces of information collected, and transmits the selected information to the upper agent 210. For example, the intermediate agent 220 uses the learning results to generate summary information that summarizes a plurality of pieces of collected information and sends it to the upper agent 210.
- the intermediate agent 220 may execute processing with lower real-time performance than the lower-level agent 230.
- the intermediate agent 220 performs processing using information collected from one or more lower-level agents 230 at predetermined intervals, such as one hour and one day.
- the intermediate agent 220 performs learning using information for a predetermined period.
- the intermediate agent 220 cooperates with other intermediate agents 220 to adjust the amount of information sent from the multiple lower-level agents 230 to the upper layer, and to adjust the amount of information transmitted from the multiple intermediate agents 220 to the higher-level agent 210. You may adjust the amount.
- the intermediate agent 220 may perform learning or processing that ensures robustness in the autonomous decentralized system 200.
- the intermediate agent 220 may play the role of Spinal Cord in the autonomous decentralized system 200.
- the upper level agent 210 collects information.
- Upper level agent 210 may collect information from intermediate agents 220.
- the upper agent 210 collects from the lower agent 230 the information collected from the intermediate agent 220 and from the lower agent 230 .
- the upper agent 210 collects from the intermediate agent 220 information selected by the intermediate agent 220 from among the plurality of pieces of information that the intermediate agent 220 has collected from the lower agent 230 .
- the upper agent 210 collects summary information from the intermediate agent 220, which is a compilation of multiple pieces of information that the intermediate agent 220 has collected from the lower agents 230.
- Upper level agents 210 may collect information from lower level agents 230.
- the upper agent 210 executes processing using the collected information.
- the upper agent 210 performs learning using the collected information, for example.
- the upper agent 210 executes various processes using the learning results, for example.
- the upper agent 210 may execute processing with lower real-time performance than the intermediate agent 220.
- the upper level agent 210 executes processing using the collected information every predetermined period such as one week, one month, and one year.
- the upper agent 210 performs learning using information for a predetermined period, for example.
- the upper level agent 210 may execute processing aimed at stabilizing the entire autonomous decentralized system 200. For example, the upper level agent 210 may perform factor analysis of the overall communication load of the communication network 100 and the cloud network 300 based on the collected information. The upper agent 210 identifies the location where the communication load on the network is increasing depending on the content of the collected information and the information collection status, and reduces the communication load by, for example, the IoT device 400, the lower agent 230, The amount of information sent by intermediate agent 220 may be adjusted and the propagation path may be adjusted.
- the autonomous decentralized system 200 by applying the autonomous decentralized system 200 according to this embodiment to a network, processing can be layered. According to the autonomous decentralized system 200, for example, while the low-level agents 230 execute highly real-time processing, only useful information is transmitted to the upper layer, and the intermediate agents 220 use the carefully selected information to perform further processing.
- the information is further narrowed down, and the upper level agent 210 performs information processing over a wider span using even more carefully selected information, and the entire network is stabilized. It can contribute to building an environment that can handle large amounts of information appropriately.
- FIG. 4 schematically shows an example of the functional configuration of the lower agent 230.
- the lower agent 230 includes a storage section 231 , a registration section 232 , an information collection section 233 , a learning execution section 234 , a processing execution section 235 , and an information transmission section 236 .
- the registration unit 232 performs various registrations.
- the registration unit 232 registers, for example, learning information used by the lower agent 230 for learning.
- the registration unit 232 stores the registered learning information in the storage unit 231.
- the learning information may include a reward.
- the learning information may include knowledge.
- the registration unit 232 registers learning information by accepting input from, for example, an operator of the autonomous decentralized system 200.
- the registration unit 232 registers learning information received from the upper agent 210, for example.
- the registration unit 232 registers learning information received from the intermediate agent 220, for example.
- the registration unit 232 may be an example of a third registration unit.
- the information collection unit 233 collects information.
- the information collection unit 233 may collect information transmitted by any device.
- the information collection unit 233 collects information transmitted by the IoT device 400, for example.
- the information collection unit 233 may collect information from the IoT device 400 through mobile communication.
- the information collection unit 233 stores the collected information in the storage unit 231.
- the information collection unit 233 may be an example of a third information collection unit.
- the learning execution unit 234 executes learning using the information collected by the information collection unit 233.
- the learning execution unit 234 may execute learning in cooperation with other lower-level agents 230 using the information collected by the information collection unit 233.
- the learning execution unit 234 stores the learning results in the storage unit 231.
- the learning result may include a model generated by learning by the learning execution unit 234.
- the learning result may include a neural network generated by learning by the learning execution unit 234.
- the learning execution unit 234 may be an example of a third learning execution unit.
- the learning execution unit 234 may use any learning method. For example, the learning execution unit 234 first executes pre-training by simulation, and updates the model, neural network, etc. based on the information collected by the information collection unit 233. Actual data collected by the information collection unit 233 in the past may be used for preliminary learning. Synthetic data may be used for pre-learning. If the environment does not allow trial and error in learning, it is effective to perform rule-based learning instead of reinforcement learning and learn only its parameters.
- the learning method may be ANN (Artificial Neural Network), DNN (Deep Neural Network), heuristics, or the like.
- the process execution unit 235 executes various processes.
- the processing execution unit 235 may use the learning results obtained by the learning execution unit 234 to execute processing on the information collected by the information collection unit 233.
- the processing execution unit 235 may generate information to be sent to the intermediate agent 220 using a plurality of pieces of information collected by the information collection unit 233.
- the processing execution unit 235 uses the learning results by the learning execution unit 234 to generate transmission information that includes information selected from the plurality of pieces of information collected by the information collection unit 233.
- the processing execution unit 235 uses the learning result by the learning execution unit 234 to generate summary information that summarizes a plurality of pieces of information collected by the information collection unit 233.
- the processing execution unit 235 may be an example of a third processing execution unit.
- the information transmitter 236 transmits information.
- the information transmitter 236 transmits the information collected by the information collector 233 to the intermediate agent 220.
- the information transmitting unit 236 transmits the information generated by the processing execution unit 235 to the intermediate agent 220.
- the information transmitter 236 may be an example of a third information transmitter.
- FIG. 5 schematically shows an example of the functional configuration of the intermediate agent 220.
- the intermediate agent 220 includes a storage section 221 , a registration section 222 , an information collection section 223 , a learning execution section 224 , a processing execution section 225 , and an information transmission section 226 .
- the registration unit 222 performs various registrations.
- the registration unit 222 registers learning information used for learning by the intermediate agent 220, for example.
- the registration unit 222 stores the registered learning information in the storage unit 221.
- the learning information may include a reward.
- the learning information may include knowledge.
- the registration unit 222 registers learning information by accepting input from, for example, an operator of the autonomous decentralized system 200.
- the registration unit 222 registers learning information received from the higher-level agent 210, for example.
- the registration unit 222 may be an example of a second registration unit.
- the information collection unit 223 collects information.
- the information collection unit 223 may collect information from the lower-level agents 230.
- the information collecting unit 223 collects, for example, information transmitted by the information transmitting unit 236 of the lower-level agent 230.
- the information collection unit 223 stores the collected information in the storage unit 221.
- the information collection unit 223 may be an example of a second information collection unit.
- the learning execution unit 224 executes learning using the information collected by the information collection unit 223.
- the learning execution unit 224 may execute learning in cooperation with other intermediate agents 220 using the information collected by the information collection unit 223.
- the learning execution unit 224 stores the learning results in the storage unit 221.
- the learning results may include a model generated by learning by the learning execution unit 224.
- the learning result may include a neural network generated by learning by the learning execution unit 224.
- the learning execution unit 224 may be an example of a second learning execution unit.
- the learning execution unit 224 may use any learning method. For example, the learning execution unit 224 first performs preliminary learning by simulation, and updates the model, neural network, etc. using information collected by the information collection unit 223. Actual data collected by the information collection unit 223 in the past may be used for preliminary learning. Synthetic data may be used for pre-learning. If the environment does not allow trial and error in learning, it is effective to perform rule-based learning instead of reinforcement learning and learn only its parameters.
- the learning method may be ANN, DNN, heuristics, or the like.
- the process execution unit 225 executes various processes.
- the processing execution unit 225 may use the learning results obtained by the learning execution unit 224 to execute processing on the information collected by the information collection unit 223.
- the processing execution unit 225 may generate information to be sent to the higher-level agent 210 using a plurality of pieces of information collected by the information collection unit 223.
- the processing execution unit 225 uses the learning results by the learning execution unit 224 to generate transmission information that includes information selected from the plurality of pieces of information collected by the information collection unit 223.
- the processing execution unit 225 uses the learning result by the learning execution unit 224 to generate summary information that summarizes a plurality of pieces of information collected by the information collection unit 223.
- the processing execution unit 225 may be an example of a second processing execution unit.
- the information transmitting unit 226 transmits information.
- the information transmitter 226 transmits the information collected by the information collector 223 to the higher-level agent 210.
- the information transmitting unit 226 transmits the information generated by the processing execution unit 225 to the higher-level agent 210.
- the information transmitter 226 may be an example of a second information transmitter.
- FIG. 6 schematically shows an example of the functional configuration of the upper agent 210.
- the upper agent 210 includes a storage section 211 , a registration section 212 , an information collection section 213 , a learning execution section 214 , a processing execution section 215 , and an information transmission section 216 .
- the registration unit 212 performs various registrations.
- the registration unit 212 registers learning information used for learning by the upper agent 210, for example.
- the registration unit 212 stores the registered learning information in the storage unit 211.
- the learning information may include a reward.
- the learning information may include knowledge.
- the registration unit 212 registers learning information by accepting input from, for example, an operator of the autonomous decentralized system 200.
- the registration unit 212 may be an example of a first registration unit.
- the information collection unit 213 collects information.
- the information collection unit 213 may collect information from the intermediate agent 220.
- the information collecting unit 213 collects information transmitted by the information transmitting unit 226 of the intermediate agent 220, for example.
- the information collection unit 213 stores the collected information in the storage unit 211.
- the information collection unit 213 may be an example of a first information collection unit.
- the learning execution unit 214 executes learning using the information collected by the information collection unit 213.
- the learning execution unit 214 may be an example of a first learning execution unit.
- the learning execution unit 214 may execute learning in cooperation with a plurality of intermediate agents 220 using the information collected by the information collection unit 213.
- the learning execution unit 214 may execute learning in cooperation with the plurality of lower-level agents 230 using the information collected by the information collection unit 213.
- the learning execution unit 214 may execute learning in cooperation with the plurality of intermediate agents 220 and the plurality of lower-level agents 230 using the information collected by the information collection unit 213.
- the learning execution unit 214 stores the learning results in the storage unit 211.
- the learning result may include a model generated by learning by the learning execution unit 214.
- the learning result may include a neural network generated by learning by the learning execution unit 214.
- the learning execution unit 214 may be an example of a first learning execution unit.
- the learning execution unit 214 may use any learning method. For example, the learning execution unit 214 first performs preliminary learning by simulation, and updates the model, neural network, etc. using information collected by the information collection unit 213. Actual data collected by the information collection unit 213 in the past may be used for preliminary learning. Synthetic data may be used for pre-learning. If the environment does not allow trial and error in learning, it is effective to perform rule-based learning instead of reinforcement learning and learn only its parameters.
- the learning method may be ANN, DNN, heuristics, or the like.
- the process execution unit 215 executes various processes.
- the processing execution unit 215 may be an example of a first processing execution unit.
- the processing execution unit 215 may use the learning results obtained by the learning execution unit 214 to execute processing on the information collected by the information collection unit 213.
- the processing execution unit 215 may use the learning results obtained by the learning execution unit 214 to execute processing aimed at stabilizing the entire network to which the autonomous decentralized system 200 is applied.
- the processing execution unit 215 may generate instruction information for the intermediate agent 220 or the lower-level agent 230 based on the result of analyzing the information collected by the information collection unit 213. For example, the processing execution unit 215 generates instruction information that instructs the lower-level agent 230 to send information to the intermediate agent 220 from among the information collected from the IoT device 400. For example, the processing execution unit 215 generates instruction information that instructs the intermediate agent 220 to process the information collected from the intermediate agent 220.
- the processing execution unit 215 may generate learning information to be sent to the lower agent 230 based on the results of analyzing the information collected by the information collection unit 213.
- the processing execution unit 215 may generate learning information to be sent to the intermediate agent 220 based on the result of analyzing the information collected by the information collection unit 213. For example, the processing execution unit 215 generates learning information including a reward set according to a change in the trend of the information collected by the information collection unit 213.
- the information transmitter 216 transmits information. For example, the information transmitting unit 216 transmits the instruction information generated by the processing execution unit 215 to the lower agent 230. For example, the information transmitting unit 216 transmits instruction information generated by the processing execution unit 215 to the intermediate agent 220.
- the information transmitting unit 216 transmits the learning information generated by the processing execution unit 215 to the lower agent 230.
- the information transmitting unit 216 transmits the learning information generated by the processing execution unit 215 to the intermediate agent 220.
- the lower-level agent 230 will perform analysis that requires more real-time performance, and the higher-level agent 210 will perform analysis that requires more real-time performance. Analyzing trends over long periods of time, intermediate agent 220 performs the corresponding analysis during that time.
- multiple lower-level agents 230 collect image data, object detection data, etc. from multiple IoT devices 400 placed in the area. do.
- the area is divided into a plurality of subareas, and each of the plurality of lower-level agents 230 collects information from the IoT device 400 placed in each of the plurality of subareas.
- a plurality of lower-level agents 230 in charge of geographically adjacent subareas execute learning for detecting the occurrence of an accident while sharing information.
- the lower agent 230 may use the learning results to detect the occurrence of an accident.
- a plurality of intermediate agents 220 are also assigned to a group of subareas to collect information from lower level agents 230 that collect information from IoT devices 400 in subareas of the group.
- the plurality of intermediate agents 220 perform learning to predict the occurrence of an accident by cooperating with each other.
- Intermediate agent 220 may use the learning results to predict the occurrence of an accident.
- the upper agent 210 performs learning for controlling overall information, for example, based on the prediction results of the plurality of intermediate agents 220.
- the upper level agent 210 may increase the amount of information collected or increase the types of information for subareas where accidents are expected to occur, and reduce the amount of information collected for other subareas. , perform control such as reducing the types of information.
- the lower-level agent 230 will route the message. implement.
- the lower-level agent 230 cooperates with other lower-level agents 230 to perform learning to control Topic, To (including Copy), and division of messages from the publisher. Using the learning results, the lower-level agent 230 cooperates with other lower-level agents 230 to generate a topic, determine To, duplicate the message, and send the message so that the message from the publisher reaches the appropriate subscriber. Control splitting, etc.
- the intermediate agent 220 monitors the routing by the lower agent 230 using information collected from the lower agent 230, and performs learning so that it can perform processing to resolve problems that occur in the routing. For example, the intermediate agent 220 buffers the message when there is a message with the same To whose destination is unknown, and sends a part of the buffered message to the network when the amount of buffering exceeds a threshold. Release. As a result, if it comes back again, it will be buffered for a certain period of time until it can be delivered. Then, when messages start to arrive, transmission of messages with the same To among the buffered messages is started.
- the upper agent 210 performs learning to detect problems in the network using information collected from the intermediate agent 220.
- the higher-level agent 210 detects a problem in the network using the learning results, it notifies the network operator of the detection result or outputs an instruction to change the network configuration.
- each of the plurality of lower-level agents 230 may be assigned to a vehicle located within each region divided into a plurality of regions.
- Information is collected from the IoT devices 400 installed, the IoT devices 400 installed in traffic lights, etc., the IoT devices 400 installed on roads, etc.
- the lower-level agent 230 may collect vehicle location information, images captured by a vehicle camera or street camera, vehicle detection information or human detection information detected by a road sensor, information detected by a vehicle sensor, etc.
- the system collects information such as the distance between vehicles, local weather information, vehicle navigation information, and vehicle speed.
- the lower agent 230 will be in charge of a certain area, and will collect information from the IoT device 400 installed in the vehicle from the time a certain vehicle enters the area, and When exiting the vehicle, information collection is handed over to the subordinate agent 230 in charge of the area into which the vehicle enters.
- the lower-level agent 230 shares information with other lower-level agents 230 to detect that the distance between vehicles is shorter than a threshold, or that the vehicle is in danger of colliding with a person or the like.
- the lower agent 230 detects danger, it transmits danger detection information to the target vehicle.
- the vehicle warns the driver or stops driving.
- the intermediate agent 220 instructs the lower-level agent 230 what information to communicate.
- the intermediate agent 220 analyzes (learns) important information (many people, many cars, etc.) based on information (day of the week, time, etc.) collected from the lower agent 230 in the past, and determines the collection priority. Make it possible to judge high-level information. Then, the intermediate agent 220 issues instructions to the lower agent 230 so that high priority information can be collected for each period.
- the intermediate agent 220 can identify high-priority information with high accuracy.
- the intermediate agent 220 instructs to prioritize vehicle detection results so that information can be collected in the morning on weekdays, and to prioritize human detection results in the afternoon on holidays. The system will now instruct users to collect information.
- the higher level agent 210 performs a broader range of analysis, for example.
- the upper level agent 210 performs prediction several steps ahead (at a set time) based on traffic information (traffic volume, time of day, day of the week, event, weather) for the area in charge (city, prefecture, country), and Provide information to 220.
- traffic information traffic volume, time of day, day of the week, event, weather
- area in charge city, prefecture, country
- FIG. 7 schematically shows another example of the autonomous decentralized system 200.
- the autonomous decentralized system 200 may further include a switch agent 250.
- Switch agent 250 may control the amount of information flowing through communication network 100. For example, when message routing is controlled by upper level agent 210, intermediate agent 220, and lower level agent 230, switch agent 250 performs communication traffic control.
- FIG. 8 schematically shows an example of the hardware configuration of a computer 1200 that implements the upper agent 210, intermediate agent 220, or lower agent 230.
- the program installed on the computer 1200 causes the computer 1200 to function as one or more "parts" of the apparatus according to the present embodiment, or causes the computer 1200 to perform operations associated with the apparatus according to the present embodiment or the one or more "parts" of the apparatus according to the present embodiment.
- Multiple units may be executed and/or the computer 1200 may execute a process or a stage of a process according to the present embodiments.
- Such programs may be executed by CPU 1212 to cause computer 1200 to perform certain operations associated with some or all of the blocks in the flowcharts and block diagrams described herein.
- a computer 1200 includes a CPU 1212, a RAM 1214, and a graphics controller 1216, which are interconnected by a host controller 1210.
- the computer 1200 also includes input/output units such as a communication interface 1222, a storage device 1224, a DVD drive 1226, and an IC card drive, which are connected to the host controller 1210 via an input/output controller 1220.
- DVD drive 1226 may be a DVD-ROM drive, a DVD-RAM drive, or the like.
- Storage device 1224 may be a hard disk drive, solid state drive, or the like.
- Computer 1200 also includes legacy input/output units, such as ROM 1230 and a keyboard, which are connected to input/output controller 1220 via input/output chips 1240.
- the CPU 1212 operates according to programs stored in the ROM 1230 and RAM 1214, thereby controlling each unit.
- Graphics controller 1216 obtains image data generated by CPU 1212, such as in a frame buffer provided in RAM 1214 or itself, and causes the image data to be displayed on display device 1218.
- the communication interface 1222 communicates with other electronic devices via the network.
- Storage device 1224 stores programs and data used by CPU 1212 within computer 1200.
- the DVD drive 1226 reads a program or data from a DVD-ROM 1227 or the like and provides it to the storage device 1224.
- the IC card drive reads programs and data from and/or writes programs and data to the IC card.
- ROM 1230 stores therein programs such as a boot program executed by computer 1200 upon activation and/or programs dependent on the computer 1200 hardware.
- I/O chip 1240 may also connect various I/O units to I/O controller 1220 via USB ports, parallel ports, serial ports, keyboard ports, mouse ports, etc.
- the program is provided by a computer readable storage medium such as a DVD-ROM 1227 or an IC card.
- the program is read from a computer-readable storage medium, installed in storage device 1224, RAM 1214, or ROM 1230, which are also examples of computer-readable storage media, and executed by CPU 1212.
- the information processing described in these programs is read by the computer 1200 and provides coordination between the programs and the various types of hardware resources described above.
- An apparatus or method may be configured to implement the operation or processing of information according to the use of computer 1200.
- the CPU 1212 executes a communication program loaded into the RAM 1214 and sends communication processing to the communication interface 1222 based on the processing written in the communication program. You may give orders.
- the communication interface 1222 reads transmission data stored in a transmission buffer area provided in a recording medium such as a RAM 1214, a storage device 1224, a DVD-ROM 1227, or an IC card under the control of the CPU 1212, and transmits the read transmission data. Data is transmitted to the network, or received data received from the network is written to a reception buffer area provided on the recording medium.
- the CPU 1212 causes the RAM 1214 to read all or a necessary part of a file or database stored in an external recording medium such as a storage device 1224, a DVD drive 1226 (DVD-ROM 1227), or an IC card. Various types of processing may be performed on the data. CPU 1212 may then write the processed data back to an external storage medium.
- an external recording medium such as a storage device 1224, a DVD drive 1226 (DVD-ROM 1227), or an IC card.
- Various types of processing may be performed on the data.
- CPU 1212 may then write the processed data back to an external storage medium.
- CPU 1212 performs various types of operations, information processing, conditional determination, conditional branching, unconditional branching, and information retrieval on data read from RAM 1214 as described elsewhere in this disclosure and specified by the program's instruction sequence. Various types of processing may be performed, including /substitutions, etc., and the results are written back to RAM 1214. Further, the CPU 1212 may search for information in a file in a recording medium, a database, or the like.
- the CPU 1212 selects the first entry from among the plurality of entries. Search for an entry whose attribute value matches the specified condition, read the attribute value of the second attribute stored in the entry, and then set the attribute value to the first attribute that satisfies the predetermined condition. An attribute value of the associated second attribute may be obtained.
- the programs or software modules described above may be stored in a computer-readable storage medium on or near computer 1200.
- a storage medium such as a hard disk or RAM provided in a server system connected to a dedicated communication network or the Internet can be used as a computer-readable storage medium, thereby allowing the program to be transferred to the computer 1200 via the network. provide.
- Blocks in the flowcharts and block diagrams of the present embodiments may represent a stage of a process in which an operation is performed or a "section" of a device responsible for performing the operation.
- Certain steps and units may be provided with dedicated circuitry, programmable circuitry provided with computer readable instructions stored on a computer readable storage medium, and/or provided with computer readable instructions stored on a computer readable storage medium. May be implemented by a processor.
- Dedicated circuitry may include digital and/or analog hardware circuits, and may include integrated circuits (ICs) and/or discrete circuits.
- Programmable circuits can perform AND, OR, EXCLUSIVE OR, NAND, NOR, and other logical operations, such as field programmable gate arrays (FPGAs), programmable logic arrays (PLAs), etc. , flip-flops, registers, and memory elements.
- FPGAs field programmable gate arrays
- PLAs programmable logic arrays
- flip-flops registers, and memory elements.
- a computer-readable storage medium may include any tangible device capable of storing instructions for execution by a suitable device such that a computer-readable storage medium with instructions stored therein may be illustrated in a flowchart or block diagram.
- a product will be provided that includes instructions that can be executed to create a means for performing specified operations.
- Examples of computer-readable storage media may include electronic storage media, magnetic storage media, optical storage media, electromagnetic storage media, semiconductor storage media, and the like. More specific examples of computer readable storage media include floppy disks, diskettes, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM or flash memory).
- EEPROM Electrically Erasable Programmable Read Only Memory
- SRAM Static Random Access Memory
- CD-ROM Compact Disc Read Only Memory
- DVD Digital Versatile Disk
- Blu-ray Disc Memory Stick
- integrated circuit cards and the like.
- Computer-readable instructions may include assembler instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state configuration data, or instructions such as Smalltalk®, JAVA®, C++, etc. any source code or object code written in any combination of one or more programming languages, including object-oriented programming languages such as may include.
- ISA instruction set architecture
- the computer-readable instructions are for producing means for a processor of a general purpose computer, special purpose computer, or other programmable data processing device, or programmable circuit to perform the operations specified in the flowchart or block diagrams.
- a general purpose computer, special purpose computer, or other programmable data processor locally or over a local area network (LAN), wide area network (WAN), such as the Internet, to execute the computer readable instructions. It may be provided in a processor or programmable circuit of the device. Examples of processors include computer processors, processing units, microprocessors, digital signal processors, controllers, microcontrollers, and the like.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Computer And Data Communications (AREA)
Abstract
Description
[先行技術文献]
[特許文献]
[特許文献1]特開2004-227208号公報
Claims (6)
- ネットワークに配置された、第1階層の第1階層エージェント及び前記第1階層よりも下位の第2階層の複数の第2階層エージェント
を備え、
前記複数の第2階層エージェントのそれぞれは、収集した情報を用いて、他の第2階層エージェントと協調して学習を実行し、学習結果を用いて生成した情報を前記第1階層エージェントに送信し、
前記第1階層エージェントは、前記複数の第2階層エージェントから受信した情報を用いて学習を実行する、
システム。 - 前記第2階層よりも下位の第3階層の複数の第3階層エージェント
を更に備え、
前記複数の第3階層エージェントのそれぞれは、収集した情報を用いて、他の第3階層エージェントと協調して学習を実行し、学習結果を用いて生成した情報を前記複数の第2階層エージェントの少なくともいずれかに送信する、請求項1に記載のシステム。 - 前記複数の第3階層エージェントのそれぞれは、
情報を収集する第3情報収集部と、
前記第3情報収集部によって収集された情報を用いて、他の複数の第3階層エージェントと協調して学習を実行する第3学習実行部と、
前記第3学習実行部による学習結果を用いて生成した情報を前記複数の第2階層エージェントの少なくともいずれかに送信する第3情報送信部と
を有し、
前記複数の第2階層エージェントのそれぞれは、
複数の前記第3階層エージェントから情報を収集する第2情報収集部と、
前記第2情報収集部によって収集された情報を用いて、他の複数の第2階層エージェントと協調して学習を実行する第2学習実行部と、
前記第2学習実行部による学習結果を用いて生成した情報を前記第1階層エージェントに送信する第2情報送信部と
を有し、
前記第1階層エージェントは、
複数の前記第2階層エージェントから情報を収集する第1情報収集部と、
前記第1情報収集部によって収集された情報を用いて学習を実行する第1学習実行部と
を有する、請求項2に記載のシステム。 - 前記第1階層エージェントは、階層型ネットワークの第1NW階層に配置され、
前記複数の第2階層エージェントは、前記階層型ネットワークの前記第1NW階層よりも下位の第2NW階層に配置され、
前記複数の第3階層エージェントは、前記階層型ネットワークの前記第2NW階層よりも下位の第3NW階層に配置される、請求項3に記載のシステム。 - 前記階層型ネットワークは、クラウドネットワークであり、
前記第1NW階層は、クラウドコンピューティングにより構成され、
前記第2NW階層は、複数のフォグコンピューティングにより構成され、
前記第3NW階層は、複数のエッジコンピューティングにより構成される、請求項4に記載のシステム。 - 前記階層型ネットワークは、クラウドネットワークであり、
前記第1NW階層は、クラウドコンピューティングにより構成され、
前記第2NW階層は、複数のフォグコンピューティングにより構成され、
前記第3NW階層は、複数のエッジコンピューティングにより構成され
前記第3情報収集部は、複数のIoTデバイスから移動体通信によって情報を収集し、
前記第3情報送信部は、前記第3学習実行部による学習結果を用いて、前記第3情報収集部が収集した情報から選択した情報、又は、前記学習結果を用いて生成した情報を前記複数の第2階層エージェントの少なくともいずれかに送信し、
前記第2情報収集部は、前記第3情報送信部によって送信された情報を収集し、
前記第2情報送信部は、前記第2学習実行部による学習結果を用いて、前記第2情報収集部が収集した情報から選択した情報、又は、前記学習結果を用いて生成した情報を前記第1階層エージェントに送信する、請求項4に記載のシステム。
Priority Applications (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202380033774.1A CN119096254A (zh) | 2022-04-14 | 2023-04-06 | 系统 |
| EP23788265.9A EP4502879B1 (en) | 2022-04-14 | 2023-04-06 | System |
| US18/911,251 US20250036921A1 (en) | 2022-04-14 | 2024-10-10 | System |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2022066766A JP7392027B2 (ja) | 2022-04-14 | 2022-04-14 | システム |
| JP2022-066766 | 2022-04-14 |
Related Child Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| US18/911,251 Continuation US20250036921A1 (en) | 2022-04-14 | 2024-10-10 | System |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2023199846A1 true WO2023199846A1 (ja) | 2023-10-19 |
Family
ID=88329680
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/JP2023/014300 Ceased WO2023199846A1 (ja) | 2022-04-14 | 2023-04-06 | システム |
Country Status (5)
| Country | Link |
|---|---|
| US (1) | US20250036921A1 (ja) |
| EP (1) | EP4502879B1 (ja) |
| JP (1) | JP7392027B2 (ja) |
| CN (1) | CN119096254A (ja) |
| WO (1) | WO2023199846A1 (ja) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2025176449A (ja) * | 2024-05-21 | 2025-12-04 | 三浦工業株式会社 | 環境情報収集システム |
| JP7672632B1 (ja) * | 2025-02-10 | 2025-05-08 | 久利寿 帝都 | ネットワーク化システムおよび情報処理方法 |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004227208A (ja) | 2003-01-22 | 2004-08-12 | Matsushita Electric Ind Co Ltd | ユーザ適応型行動決定装置および行動決定方法 |
| JP2021140825A (ja) * | 2014-12-22 | 2021-09-16 | インテル コーポレイション | ホリスティックグローバルなパフォーマンス及び電力管理 |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105610944B (zh) * | 2015-12-29 | 2019-03-05 | 北京物联远信息技术有限公司 | 一种面向物联网的雾计算架构 |
| CN111709542A (zh) * | 2020-06-12 | 2020-09-25 | 浪潮集团有限公司 | 一种基于雾计算环境的车辆预测诊断方法 |
| WO2022212079A1 (en) * | 2021-04-01 | 2022-10-06 | University Of South Florida | Deep reinforcement learning for adaptive network slicing in 5g for intelligent vehicular systems and smart cities |
| CN113465920B (zh) * | 2021-06-08 | 2022-04-22 | 西安交通大学 | 云、雾、边缘端协同的轴承状态监测与管理方法及系统 |
-
2022
- 2022-04-14 JP JP2022066766A patent/JP7392027B2/ja active Active
-
2023
- 2023-04-06 EP EP23788265.9A patent/EP4502879B1/en active Active
- 2023-04-06 WO PCT/JP2023/014300 patent/WO2023199846A1/ja not_active Ceased
- 2023-04-06 CN CN202380033774.1A patent/CN119096254A/zh active Pending
-
2024
- 2024-10-10 US US18/911,251 patent/US20250036921A1/en not_active Abandoned
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004227208A (ja) | 2003-01-22 | 2004-08-12 | Matsushita Electric Ind Co Ltd | ユーザ適応型行動決定装置および行動決定方法 |
| JP2021140825A (ja) * | 2014-12-22 | 2021-09-16 | インテル コーポレイション | ホリスティックグローバルなパフォーマンス及び電力管理 |
Non-Patent Citations (2)
| Title |
|---|
| ALEXANDER SASHA VEZHNEVETS; SIMON OSINDERO; TOM SCHAUL; NICOLAS HEESS; MAX JADERBERG; DAVID SILVER; KORAY KAVUKCUOGLU: "FeUdal Networks for Hierarchical Reinforcement Learning", ARXIV.ORG, 3 March 2017 (2017-03-03), XP080754067 * |
| See also references of EP4502879A4 |
Also Published As
| Publication number | Publication date |
|---|---|
| EP4502879A4 (en) | 2025-07-09 |
| EP4502879B1 (en) | 2026-02-04 |
| EP4502879A1 (en) | 2025-02-05 |
| CN119096254A (zh) | 2024-12-06 |
| JP2023157091A (ja) | 2023-10-26 |
| EP4502879C0 (en) | 2026-02-04 |
| JP7392027B2 (ja) | 2023-12-05 |
| US20250036921A1 (en) | 2025-01-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112400192B (zh) | 多模态深度交通信号控制的方法和系统 | |
| JP6738932B2 (ja) | シミュレーションプラットフォームに配置された、機械学習モデルを訓練するためのシステム及び方法 | |
| US11181913B2 (en) | Autonomous vehicle fleet model training and testing | |
| US11548518B2 (en) | Subjective route comfort modeling and prediction | |
| Alsrehin et al. | Intelligent transportation and control systems using data mining and machine learning techniques: A comprehensive study | |
| JP7542085B2 (ja) | 離脱予測とトリアージ支援のためのシステムと方法 | |
| JP6722280B2 (ja) | 自律走行車の交通予測における予測軌跡の評価フレームワーク | |
| US11398150B2 (en) | Navigation analysis for a multi-lane roadway | |
| JP7068260B2 (ja) | データ処理を優先順位付けするシステム及び方法 | |
| US20250036921A1 (en) | System | |
| EP3035314A1 (en) | A traffic data fusion system and the related method for providing a traffic state for a network of roads | |
| CN112629550B (zh) | 一种预测障碍物轨迹以及模型训练的方法及装置 | |
| Louati et al. | Multi-agent deep neural networks coupled with LQF-MWM algorithm for traffic control and emergency vehicles guidance | |
| JP2019192234A (ja) | 複数のキューを利用したオブジェクト追跡 | |
| Ikiriwatte et al. | Traffic density estimation and traffic control using convolutional neural network | |
| CN112783619A (zh) | 一种任务调度方法、装置和自动驾驶系统 | |
| Louati | A hybridization of deep learning techniques to predict and control traffic disturbances | |
| CN115158323A (zh) | 车辆的行驶路况预测方法和车辆 | |
| Mahmoud et al. | The future of digital twins for autonomous systems: analysis and opportunities | |
| US20240095151A1 (en) | Optimized test selection | |
| Tzortzoglou et al. | Teaching cars to drive | |
| US20220348232A1 (en) | Fleet-level av simulation system and method | |
| US12135632B2 (en) | Control loop for scaling agents in a cluster | |
| KR20220101535A (ko) | 딥러닝 기반 온-디바이스 실시간 교통제어 시스템 | |
| CN110807912A (zh) | 一种道路行车速度预测方法、终端设备及存储介质 |
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: 23788265 Country of ref document: EP Kind code of ref document: A1 |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 202380033774.1 Country of ref document: CN |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 2023788265 Country of ref document: EP |
|
| ENP | Entry into the national phase |
Ref document number: 2023788265 Country of ref document: EP Effective date: 20241101 |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 202417087480 Country of ref document: IN |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| WWG | Wipo information: grant in national office |
Ref document number: 2023788265 Country of ref document: EP |