WO2020010558A1 - 检测方法和检测装置 - Google Patents

检测方法和检测装置 Download PDF

Info

Publication number
WO2020010558A1
WO2020010558A1 PCT/CN2018/095339 CN2018095339W WO2020010558A1 WO 2020010558 A1 WO2020010558 A1 WO 2020010558A1 CN 2018095339 W CN2018095339 W CN 2018095339W WO 2020010558 A1 WO2020010558 A1 WO 2020010558A1
Authority
WO
WIPO (PCT)
Prior art keywords
positioning base
impulse response
line
base station
propagation
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/CN2018/095339
Other languages
English (en)
French (fr)
Inventor
于华俊
曾卓琦
王炜
刘史蒂文
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.)
Robert Bosch GmbH
Original Assignee
Robert Bosch GmbH
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 Robert Bosch GmbH filed Critical Robert Bosch GmbH
Priority to CN201880095467.5A priority Critical patent/CN112425220B/zh
Priority to PCT/CN2018/095339 priority patent/WO2020010558A1/zh
Priority to US17/257,965 priority patent/US11500055B2/en
Priority to EP18925748.8A priority patent/EP3823372A4/en
Publication of WO2020010558A1 publication Critical patent/WO2020010558A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
    • G01S5/02—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves
    • G01S5/0246—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves involving frequency difference of arrival or Doppler measurements
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
    • G01S5/02—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves
    • G01S5/0273—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves using multipath or indirect path propagation signals in position determination
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
    • G01S5/01—Determining conditions which influence positioning, e.g. radio environment, state of motion or energy consumption
    • G01S5/011—Identifying the radio environment
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
    • G01S5/02—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves
    • G01S5/0205—Details
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
    • G01S5/02—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves
    • G01S5/0205—Details
    • G01S5/0218—Multipath in signal reception
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
    • G01S5/02—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves
    • G01S5/0252—Radio frequency fingerprinting
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
    • G01S5/02—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves
    • G01S5/0278—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves involving statistical or probabilistic considerations
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W64/00—Locating users or terminals or network equipment for network management purposes, e.g. mobility management
    • H04W64/003—Locating users or terminals or network equipment for network management purposes, e.g. mobility management locating network equipment
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Definitions

  • the invention relates to the field of ultra-wideband (UWB) positioning, and in particular, to a detection method, a detection device, a computing device, and a machine-readable storage medium.
  • UWB ultra-wideband
  • UWB positioning is a technology that uses extremely narrow impulse response and bandwidth above 1GHz to locate objects indoors.
  • the UWB positioning system includes multiple positioning base stations and positioning labels placed on objects to be positioned.
  • the positioning tag sends a pulse signal that becomes a pulse response when it reaches the positioning base station after channel modulation.
  • the UWB positioning system uses the impulse response received by the positioning base station from the positioning tag to determine the positioning of the object.
  • the UWB positioning system can obtain accurate positioning of the object.
  • the signal transmission between the positioning base station and the positioning label is obstacle blocking Non-line-of-sight propagation
  • the positioning obtained by the UWB positioning system is inaccurate. Therefore, in UWB positioning, identification of non-line-of-sight propagation is very important.
  • the value of the impulse response characteristic calculated by using the impulse response received from the positioning tag received by the positioning base station has great differences, therefore, In order to be able to train and obtain a universal machine learning model in various situations, it is necessary to collect the impulse responses received by the positioning base station in various situations, and to calculate a large number of impulse response characteristics as training sample data to train the machine learning model. Therefore, the existing use of machine learning models to identify non-line-of-sight propagation is costly.
  • embodiments of the present invention provide a detection method, a detection device, a computing device, and a machine-readable storage medium, which can reduce the cost of identifying non-line-of-sight propagation using a classifier.
  • Embodiments of the present invention further provide a detection method, a detection device, a computing device, and a machine-readable storage medium, which can facilitate knowing a ranging error of a positioning base station in a hyper-bandwidth positioning system.
  • a detection method includes: when a positioning base station in an ultra-wideband positioning system receives an impulse response from a certain positioning tag this time, using the received impulse response to calculate a plurality of designated The value of the impulse response characteristic is used as the value of the multiple specified impulse response characteristics of the certain positioning base station this time; the value of the multiple specified impulse response characteristics of the certain positioning base station and the previous time are calculated The difference between the values of the plurality of specified impulse response characteristics of the certain positioning base station is used as the amount of change of the plurality of specified impulse response characteristics of the certain positioning base station this time; The amount of change in the plurality of specified impulse response characteristics of the certain positioning base station, using a trained classifier to determine whether the signal propagation that the certain positioning base station participates in this time is non-line-of-sight propagation, wherein the The classifier is trained to be able to classify the arbitrary positioning base station according to the amount of change in the plurality of specified impulse response characteristics of any positioning base station in the ultra-wideband positioning system.
  • a detection method includes: when a positioning base station in an ultra-wideband positioning system receives an impulse response from a positioning tag, using the received impulse response to calculate a plurality of specified impulse responses The value of the feature; based on the calculated value of the plurality of impulse response features, using a trained classifier to determine the level of the ranging error of the certain positioning base station, wherein the classifier is trained to be able to The values of the plurality of impulse response characteristics of arbitrary positioning base stations of the ultra-wideband positioning system classify the ranging errors of the arbitrary positioning base stations into corresponding levels.
  • a detection device includes: a first calculation module, configured to: when a positioning base station in an ultra-wideband positioning system receives an impulse response from a positioning tag this time, use the received Impulse response, calculating a plurality of specified impulse response characteristics as the value of the plurality of specified impulse response characteristics of a certain positioning base station this time; a second calculation module for calculating a certain positioning base station this time The difference between the value of the multiple specified impulse response characteristics of the and the value of the multiple specified impulse response characteristics of the certain positioning base station of the previous time is used as the multiple of the certain positioning base station this time.
  • An amount of change in the specified impulse response characteristics configured to use the trained classifier to determine the one based on at least the amount of change in the plurality of specified impulse response characteristics of a certain positioning base station this time Whether the signal propagation that a positioning base station participates in this time is non-line-of-sight, where the classifier is trained to be capable of locating base stations based on any positioning base station in the ultra-wideband positioning system.
  • the plurality of specified impulse response characteristics are used to classify the signal propagation in which the arbitrary positioning base station participates as line-of-sight propagation or non-line-of-sight propagation.
  • a detection device includes a calculation module for calculating an impulse response from a certain positioning tag when a certain positioning base station in an ultra-wideband positioning system uses the received impulse response to calculate Values of multiple specified impulse response characteristics; a determining module, configured to determine a level of a ranging error of the certain positioning base station by using a trained classifier based on the calculated values of the multiple impulse response characteristics, where The classifier is trained to be able to classify the ranging error of the arbitrary positioning base station into a corresponding level based on the values of the plurality of impulse response characteristics of the arbitrary positioning base station.
  • a computing device includes: a processor; and a memory that stores executable instructions that, when executed, cause the processor to perform the foregoing method.
  • a machine-readable storage medium has executable instructions thereon, and when the executable instructions are executed, the machine is caused to perform the foregoing method.
  • the solution of the embodiment of the present invention uses a classifier trained to be able to classify the signal propagation in which the positioning base station participates as line-of-sight propagation or non-line-of-sight propagation based on the amount of change in the impulse response characteristics of the positioning base station in the ultra-wideband positioning system. To determine whether the signal propagation involved by the positioning base station in the UWB positioning system is non-line-of-sight. Because the impulse response characteristics calculated by using the impulse responses received from the positioning tags received by the positioning base stations collected at various occasions at different times are very different from each other, it is only necessary to collect the positioning base stations at different times in different typical occasions.
  • the solution of the embodiment of the present invention can reduce the cost of identifying non-line-of-sight propagation using a classifier.
  • the solution of the embodiment of the present invention uses a classifier trained to be able to classify the ranging error of the positioning base station into a corresponding level according to the value of the impulse response characteristics of the positioning base station in the ultra-wideband positioning system to determine the ultra-wideband positioning.
  • the level of the ranging error of the positioning base station in the system can easily know the magnitude of the ranging error of the positioning base station. Therefore, compared with the prior art, the solution of the embodiment of the present invention can easily obtain the ranging of the positioning base station. The magnitude of the error.
  • FIG. 1 shows an overall flowchart of a method for model training according to a first embodiment of the present invention
  • FIG. 2 shows an overall flowchart of a detection method according to a first embodiment of the present invention
  • FIG. 3 shows an overall flowchart of a method for model training according to a second embodiment of the present invention
  • FIG. 4 shows an overall flowchart of a detection method according to a second embodiment of the present invention
  • FIG. 5A shows a flowchart of a detection method according to an embodiment of the present invention
  • 5B shows a flowchart of a detection method according to another embodiment of the present invention.
  • FIG. 6A shows a schematic diagram of a detection device according to an embodiment of the present invention.
  • FIG. 6B shows a schematic diagram of a detection device according to another embodiment of the present invention.
  • FIG. 7 shows a schematic diagram of a computing device according to an embodiment of the invention.
  • the inventor has found through a large number of experiments and studies that in UWB positioning systems, there are four transformation methods for the signal propagation between the positioning base station and the positioning tag over time, namely: from line-of-sight propagation to line-of-sight propagation, and from line-of-sight Propagation transforms into non-line-of-sight propagation, from non-line-of-sight propagation to line-of-sight propagation, and from non-line-of-sight propagation to non-line-of-sight propagation, and in the same signal propagation transformation mode, the data collected on various occasions are used.
  • the changes in the impulse response characteristics calculated by the impulse response received by the positioning base station from the positioning tag at different times are very small.
  • the embodiment of the present invention uses the variation of the impulse response characteristics as training sample data to train a machine learning model for identifying non-line-of-sight propagation, and only needs to collect positioning base stations at different times to receive from Locate the impulse response of the tag, and use the collected impulse responses to calculate the change in impulse response characteristics.
  • a machine learning model for identifying non-line-of-sight propagation a general machine learning model can be obtained in various situations. Can greatly reduce the cost of identifying non-line-of-sight propagation using machine learning models.
  • FIG. 1 shows an overall flowchart of a method for model training according to a first embodiment of the present invention.
  • the method 100 shown in FIG. 1 may be implemented by any computing device having computing capabilities.
  • the computing device may be, but is not limited to, a desktop computer, a notebook computer, a tablet computer, a server, or a smart phone.
  • Each impulse response pair includes a first impulse response and a second impulse response received by a positioning base station in a UWB positioning system located on a certain occasion on a plurality of occasions, wherein the first impulse response is the certain positioning
  • the base station receives from a positioning tag under line-of-sight or non-line-of-sight propagation at a certain time, and the second impulse response is that the certain positioning base station is in the video at another time after the certain time.
  • the multiple occasions may include an airport, a train station, a parking lot, a shopping mall, and the like.
  • the acquired impulse response pairs are divided into a first signal group, a second signal group, a third signal group, and a fourth signal group.
  • each signal group includes several impulse responses.
  • the first signal group corresponds to the case where the signal propagation is changed from line-of-sight propagation to line-of-sight propagation, that is, any of the impulse responses included in the first signal group to the first impulse response g1i-1 in G1i is receiving the first pulse
  • the signal propagation between the positioning base station responding to g1i-1 and the positioning tag transmitting the impulse signal corresponding to the first impulse response g1i-1 (that is, the signal propagation participating in the positioning base station receiving the first impulse response g1i-1) is Is received in the case of range propagation
  • the second impulse response g1i-2 in this impulse response to G1i is the positioning base station that receives the second impulse response g1i-2 and sends the second impulse response g1i-2
  • the signal propagation between the positioning tags of the corresponding pulse signals is received in the case of line-of-sight propagation.
  • the plurality of impulse response pairs included in the first signal group are received by the plurality of positioning base stations F1D during signal propagation in which the plurality of positioning base stations F1D participate in the transformation from line-of-sight propagation to line-of-sight propagation.
  • the second signal group corresponds to the case where the signal propagation is changed from line-of-sight propagation to non-line-of-sight propagation, that is, any impulse response included in the second signal group responds to the first impulse response g2i-1 in G2i.
  • the signal propagation between the positioning base station with the impulse response g2i-1 and the positioning tag that transmits the impulse signal corresponding to the first impulse response g2i-1 is received in the case of line-of-sight propagation, and any one of the impulse responses to G2i
  • the second impulse response g2i-2 is a case where the signal propagation between the positioning base station receiving the second impulse response g2i-2 and the positioning tag transmitting the impulse signal corresponding to the second impulse response g2i-2 is non-line-of-sight propagation.
  • the plurality of impulse response pairs included in the second signal group are received by the plurality of positioning base stations F2D during a signal propagation in which the plurality of positioning base stations F2D participates changes from line-of-sight propagation to non-line-of-sight propagation.
  • the third signal group corresponds to the case where the signal propagation is changed from non-line-of-sight propagation to line-of-sight propagation, that is, any impulse response included in the third signal group responds to the first impulse response g3i-1 in G3i.
  • the signal propagation between the positioning base station with the impulse response g3i-1 and the positioning tag transmitting the impulse signal corresponding to the first impulse response g3i-1 is received when the line-of-sight propagation is non-line-of-sight, and any of the impulse responses to G3i
  • the second impulse response g3i-2 is the case where the signal propagation between the positioning base station receiving the second impulse response g3i-2 and the positioning tag transmitting the impulse signal corresponding to the second impulse response g3i-2 is line-of-sight.
  • the plurality of impulse response pairs included in the third signal group are received by the plurality of positioning base stations F3D during a signal propagation in which the plurality of positioning base stations F3D participates changes from line-of-sight propagation to non-line-of-sight propagation.
  • the fourth signal group corresponds to a case where the signal propagation is changed from non-line-of-sight propagation to non-line-of-sight propagation, that is, any of the impulse responses included in the fourth signal group responds to the first impulse response g4i-1 in G4i.
  • the signal propagation between a positioning base station with an impulse response g4i-1 and a positioning tag transmitting an impulse signal corresponding to the first impulse response g4i-1 is received in the case of non-line-of-sight propagation, and either of these impulse response pairs
  • the second impulse response g4i-2 in G4i is non-line-of-sight propagation between the positioning base station that receives the second impulse response g4i-2 and the positioning tag that sends the impulse signal corresponding to the second impulse response g4i-2 The case was received.
  • the plurality of impulse response pairs included in the fourth signal group are received by the plurality of positioning base stations F4D during a signal propagation in which the plurality of positioning base stations F4D participates changes from line-of-sight propagation to non-line-of-sight propagation.
  • any eigenvalue vector pair ck of each signal group Cj includes a first eigenvalue vector and a second eigenvalue vector, and the first eigenvalue vector in the any eigenvalue vector pair ck is included by using the signal group Cj
  • the value of a plurality of specified impulse response characteristics PPT1 calculated by the first impulse response in one of the impulse response pairs is formed, and the second eigenvalue vector in any of the eigenvalue vector pairs ck is included by using the signal group Cj
  • the value of the plurality of designated impulse response characteristics PPT1 calculated by the second impulse response in one of the impulse response pairs is calculated.
  • the plurality of specified impulse response characteristics PPT1 can be selected from the following impulse response characteristics according to the actual situation: the distance between the positioning base station and the positioning label, the received signal energy, the maximum amplitude, the maximum amplitude rise time, and the standard Poor, power difference between the first path and the strongest path, power ratio of the first path and the strongest path, signal-to-noise ratio (SNR), form factor, delay from the peak of the received pulse to the start time, average excess delay, mean square delay Spread, kurtosis, crest factor, peak-to-average power ratio, and skewness.
  • SNR signal-to-noise ratio
  • a plurality of feature change vectors for each of the first signal group, the second signal group, the third signal group, and the fourth signal group are calculated.
  • Any feature change vector of each signal group Cj is a difference between a first feature value vector and a second feature value vector included in one of the feature value vector pairs of the signal group Cj.
  • each feature change vector of the first signal group represents a change amount of the plurality of designated impulse response features PPT1 of one of the plurality of positioning base stations F1D
  • each feature change vector of the second signal group represents the multiple One of the positioning base stations F2D
  • each feature change vector of the third signal group represents the plurality of designated pulses of one of the plurality of positioning base stations F3D
  • each feature change vector of the fourth signal group represents a change amount of the plurality of designated impulse response characteristics PPT1 of one of the plurality of positioning base stations F4D.
  • the feature change vector of the first signal group is used as the negative training sample and the feature change vector of the second signal group is used as the positive training sample to train as a classifier for the plurality of impulse responses according to the positioning base station.
  • the variation of the characteristic PPT1 is used to classify the signal propagation in which the positioning base station participates as a line-of-sight or non-line-of-sight propagation first machine learning model M1
  • the feature change vector of the third signal group is used as the negative training sample and the fourth signal
  • the set of feature change vectors is used as a positive training sample to train as a classifier for classifying the signal propagation involved in the positioning base station as line-of-sight propagation or non-line-of-sight based on the amount of change in the multiple impulse response features PPT1 of the positioning base station.
  • Propagated second machine learning model M2 Propagated second machine learning model M2.
  • the first machine learning model M1 is suitable for the case where the signal propagation involved in positioning the base station is line-of-sight propagation before transformation
  • the second machine learning model M2 is suitable for the signal propagation involved in positioning the base station is non-line-of-sight propagation before transformation.
  • the first machine learning model M1 and the second machine learning model M2 may be a decision tree, a neural network, a support vector machine, and the like.
  • FIG. 2 shows an overall flowchart of the detection method according to the first embodiment of the present invention.
  • the method 200 shown in FIG. 2 may be implemented by any computing device having computing capabilities.
  • the computing device may be, but is not limited to, a desktop computer, a notebook computer, a tablet computer, a server, or a smart phone.
  • the received impulse response PUL is used to calculate the multiple
  • the value of the designated response signal characteristic PPT1 is used as the value of the multiple designated impulse response characteristics PPT1 of the positioning base station CP this time.
  • the difference between the values of the multiple specified impulse response characteristics PPT1 of the positioning base station CP and the values of the multiple specified impulse response characteristics PPT1 of the previous positioning base station CP is calculated as the current positioning base station The amount of change in the plurality of designated impulse response characteristics PPT1 of the CP.
  • a corresponding machine learning model is selected from the trained first machine learning model M1 and the second machine learning model M2 according to whether the signal propagation previously participated by the positioning base station CP is line-of-sight or non-line-of-sight.
  • the first learning model M1 is selected when the signal propagation previously participated by the positioning base station CP was line-of-sight propagation
  • the second learning model M2 was selected when the signal propagation previously participated by the positioning base station CP was non-line-of-sight propagation.
  • the change amount of the plurality of specified impulse response characteristics PPT1 of the positioning base station CP is input into the selected machine learning model to determine the signal propagation of the positioning base station CP this time (i.e., the positioning base station CP and the positioning Whether signal propagation between tags BQ) is non-line-of-sight propagation.
  • Time-of-arrival (TDOA) positioning is a commonly used wireless positioning technique that uses the time difference between signals received by multiple base stations from the object to be measured to calculate the position of the object to be measured.
  • TDOA positioning technology is used in the UWB positioning system to calculate the position of a positioning tag
  • multiple positioning base stations receive impulse responses from the same positioning tag, sometimes based on some of the multiple positioning base stations
  • the positions of the positioning tags calculated by the impulse responses received by the positioning base station are accurate, but the positions of the positioning tags calculated based on the impulse responses received by the other positioning base stations of the plurality of positioning base stations are not accurate.
  • the ranging errors of several positioning base stations receiving impulse responses from the same positioning tag are very different from each other, then use TDOA positioning technology
  • the positions of the positioning tags calculated based on the impulse responses received by these positioning base stations are inaccurate, because the ranging errors of several positioning base stations receiving impulse responses from the same positioning tag are not significantly different from each other
  • the ranging errors of these positioning base stations can basically cancel each other out, so that the calculated positioning The position of the label is basically accurate.
  • the ranging error of the positioning base station refers to the estimated distance between the positioning base station and the positioning tag and the true distance between the positioning base station and the positioning tag calculated by using the impulse response received from the positioning base station. Difference.
  • the ranging error of the positioning base station can be known in advance, when using the TDOA positioning technology to calculate the position of the positioning tag, the impulse responses received by those positioning base stations with substantially the same ranging error can be used to calculate the position of the positioning tag.
  • the true position of the positioning tag is calculated.
  • the ranging error of the positioning base station is usually not fixed, and the prior art does not provide a solution for determining the magnitude of the ranging error of the positioning base station.
  • the inventors have found through a large number of studies that if the signal transmission line between the positioning base station and the positioning label is more severely blocked, the ranging error of the positioning base station is greater, and if the signal transmission line between the positioning base station and the positioning label is The occlusion severity changes, so the value of the impulse response characteristic calculated based on the impulse response received by the positioning base station from the positioning tag also changes accordingly.
  • the embodiment of the present invention divides the ranging error of the positioning base station into multiple different ranging error ranges and assigns a different level to each ranging error range.
  • the training is used as a classifier for positioning based on positioning.
  • the base station's impulse response characteristic value is used to classify the ranging error of the positioning base station into a corresponding level of machine learning model, and then use the trained machine learning model to determine the level of ranging error of the positioning base station, thereby facilitating the knowledge of the positioning base station.
  • Ranging error magnitude is used to classify the ranging error of the positioning base station into a corresponding level of machine learning model, and then use the trained machine learning model to determine the level of ranging error of the positioning base station, thereby facilitating the knowledge of the positioning base station.
  • FIG. 3 shows an overall flowchart of a method for model training according to a second embodiment of the present invention.
  • the method 300 shown in FIG. 3 may be implemented by any computing device having computing power.
  • the computing device may be, but is not limited to, a desktop computer, a notebook computer, a tablet computer, a server, or a smart phone.
  • a plurality of impulse responses PPS received by a plurality of positioning base stations PBS and respective real distances for the plurality of positioning base stations PBS are obtained.
  • the plurality of positioning base stations PBS are positioning base stations in one or more UWB positioning systems. Each impulse response in the plurality of impulse responses PPS is received by one of the plurality of positioning base stations PBS from a positioning tag.
  • the true distance for any positioning base station in the plurality of positioning base stations PBS indicates that when any positioning base station receives the impulse response PPSi received by the any positioning base station in the plurality of impulse response PPSs.
  • the true distance between the base station and the positioning tag that sends the impulse signal corresponding to the impulse response PPSi For example, but not limited to, the actual distance between any positioning base station and a positioning tag that transmits a pulse signal corresponding to the impulse response PPSi can be captured by a photographing device installed on the occasion where the positioning base station is located. The image is obtained after image processing.
  • an estimated distance is calculated for each of the plurality of positioning base stations PBS.
  • the estimated distance for any one of the plurality of positioning base stations PBS indicates that it is received at any one of the positioning base stations calculated using the impulse response PPSi received by the any one of the plurality of impulse response PPS.
  • the distance between any positioning base station and the positioning tag that sends a pulse signal corresponding to the impulse response PPSi may be calculated as an estimate for the positioning base station distance.
  • a ranging error of each of the plurality of positioning base stations PBS is calculated.
  • the ranging error of any positioning base station in the plurality of positioning base stations PBS represents an absolute value of a difference between an estimated distance and a real distance for the any positioning base station.
  • a mapping relationship YG between a plurality of levels L of the ranging error and a plurality of different ranging error ranges R is set.
  • each of the plurality of levels L corresponds to one of the plurality of ranging error ranges R.
  • the plurality of ranging error ranges R may include three ranging error ranges of 0 to 20 cm, 20 to 40 cm, and 40 to 60 cm.
  • a value of a plurality of designated impulse response characteristics PPT2 is calculated as the The values of the plurality of designated impulse response characteristics PPT2, so as to obtain the values of the plurality of designated impulse response characteristics PPT2 of the plurality of positioning base stations PBS.
  • the multiple specified impulse response characteristics PPT2 can be selected from the following impulse response characteristics according to the actual situation: the distance between the positioning base station and the positioning tag, the received signal energy, the maximum amplitude, the maximum amplitude rise time, the standard deviation, the first path and Power difference of the strongest path, power ratio of the first path to the strongest path, signal-to-noise ratio (SNR), form factor, peak delay of the received pulse to start time, average excess delay, mean square delay spread, kurtosis, crest Factor, peak-to-average power ratio, and skewness.
  • SNR signal-to-noise ratio
  • the values of the plurality of designated impulse response features PPT2 and the levels of the ranging errors of the plurality of positioning base stations PBS are used as training sample data to train and obtain a machine learning model M3 as a classifier.
  • the machine learning model M3 is trained to be able to classify the ranging error of the arbitrary positioning base station into a corresponding level according to the values of the plurality of impulse response characteristics PPT2 of the arbitrary positioning base station.
  • FIG. 4 shows an overall flowchart of a detection method according to a second embodiment of the present invention.
  • the method 400 shown in FIG. 4 may be implemented by any computing device having computing capabilities.
  • the computing device may be, but is not limited to, a desktop computer, a notebook computer, a tablet computer, a server, or a smart phone.
  • a positioning base station CP in the UWB positioning system uses the received impulse response PUL to calculate the plurality of designated responses.
  • the value of the signal characteristic PPT2 is used as the value of the plurality of designated response signal characteristics PPT2 of the positioning base station CP.
  • the values of the plurality of designated response signal characteristics PPT2 of the positioning base station CP are input into the trained machine learning model M3 to determine the level of the ranging error of the positioning base station CP.
  • the determined level of the ranging error of the positioning base station CP can be displayed for use by the user.
  • the method 400 can be used to determine the level of the ranging error of the multiple positioning base stations, and then the measurement in the multiple positioning base stations can be used.
  • the impulse responses received by those positioning base stations with the same level of distance error determine the location of the positioning label through TDOA positioning technology, which is compared with the use of pulse signals received by those positioning base stations with different levels of ranging error for positioning using TDOA positioning technology. Can effectively improve positioning accuracy.
  • the machine learning for classifying the signal propagation in which the positioning base station participates as line-of-sight propagation or non-line-of-sight propagation is based on the amount of change in the impulse response characteristics of the positioning base station
  • the model includes two models, namely: a first machine learning model M1 and a first machine learning model M2, however, the present invention is not limited thereto.
  • the machine learning model used to classify the signal propagation in which the positioning base station participates as line-of-sight propagation or non-line-of-sight propagation according to the variation of the impulse response characteristics of the positioning base station may also be a single machine learning model.
  • the feature change vectors of the first signal group and the third signal group can be used as negative training samples, and the feature change vectors of the second signal group and the fourth signal group are used as positive training samples to obtain The single machine learning model, and the method 200 does not include block 210.
  • the method 100 is only an example method of training a machine learning model for classifying a signal propagation in which a positioning base station participates into line-of-sight propagation or non-line-of-sight propagation according to a change amount of an impulse response characteristic of the positioning base station, Any other suitable method may also be used to train a machine learning model for classifying the signal propagation involved in the positioning base station into line-of-sight propagation or non-line-of-sight propagation based on the amount of change in the impulse response characteristics of the positioning base station.
  • the method 300 is only an example method of training a machine learning model for classifying a ranging error of a positioning base station into a corresponding level according to a value of an impulse response characteristic of the positioning base station, and any other suitable The method is used to train a machine learning model for classifying the ranging error of the positioning base station into corresponding levels according to the values of the impulse response characteristics of the positioning base station.
  • the classifier used is a machine learning model (ie, the first machine learning model M1, the first machine learning model M2, and the machine learning module M3 )
  • the present invention is not limited to this. In some other embodiments of the present invention, the classifier used may also be any other suitable classifier other than a machine learning model.
  • FIG. 5A shows a flowchart of a detection method according to an embodiment of the present invention.
  • the detection method 500 shown in FIG. 5A may be implemented by any computing device having computing capabilities.
  • the computing device may be, but is not limited to, a desktop computer, a notebook computer, a tablet computer, a server, or a smart phone.
  • the detection method 500 may include, in block 502, when a certain positioning base station in an ultra-wideband positioning system receives an impulse response from a certain positioning tag this time, using the received impulse response to calculate The values of the multiple specified impulse response characteristics are used as the values of the multiple specified impulse response characteristics of a certain positioning base station this time.
  • the detection method 500 may further include, in block 504, calculating values of the plurality of specified impulse response characteristics of the certain positioning base station this time and the plurality of specified impulse response characteristics of the certain positioning base station the previous time. The difference between the values is used as the change amount of the multiple specified impulse response characteristics of a certain positioning base station this time.
  • the detection method 500 may further include, at block 506, using a trained classifier to determine the location of a certain positioning base station based on at least changes in the plurality of specified impulse response characteristics of the certain positioning base station this time. Whether the second-participant signal propagation is non-line-of-sight propagation, wherein the classifier is trained to be capable of converting the arbitrary according to the amount of change in the plurality of specified impulse response characteristics of any positioning base station in the ultra-wideband positioning system The signal propagation that the positioning base station participates in is classified as line-of-sight propagation or non-line-of-sight propagation.
  • the classifier includes a first classifier and a second classifier, wherein the positive training samples and the negative training samples used to train the first classifier are signal propagations participating in the first plurality of positioning base stations, respectively The amount of change in the plurality of designated impulse response characteristics of the first plurality of positioning base stations and the signal propagation participating in the second plurality of positioning base stations during the conversion from line-of-sight propagation to non-line-of-sight propagation. The amount of change in the plurality of specified impulse response characteristics of the second plurality of positioning base stations during propagation, and the positive training samples and the negative training samples used to train the second classifier are in the third plurality of positioning, respectively The signal propagation involved in the base station is converted from non-line-of-sight propagation to non-line-of-sight propagation during the non-line-of-sight propagation.
  • Non-line-of-sight propagation is converted into a change amount of the plurality of specified impulse response characteristics of the fourth plurality of positioning base stations during line-of-sight propagation, and wherein the certain positioning is determined Whether the signal propagation that the base station participates in this time is non-line-of-sight propagation includes: according to whether the previous signal propagation that a certain base station participated in is line-of-sight or non-line-of-sight, from the first classifier and the second class A corresponding classifier is selected among the classifiers, wherein the first classifier is selected when the signal propagation previously participated by the certain positioning base station is line-of-sight propagation, otherwise the second classifier is selected; and based on The amount of change in the plurality of specified impulse response characteristics of a certain positioning base station this time uses the selected classifier to determine whether the signal propagation that the certain positioning base station participates in this time is non-line-of-sight propagation.
  • the classifier is a single classifier, wherein the positive training samples used to train the classifier include signals transmitted during the conversion of line-of-sight propagation to non-line-of-sight propagation involved in the fifth plurality of positioning base stations.
  • the amount of change in the plurality of designated impulse response characteristics of the base station, and the negative training samples used to train the classifier include the signal propagation involved in the seventh plurality of positioning base stations being converted from line-of-sight propagation to line-of-sight propagation
  • the amount of change in the plurality of designated impulse response characteristics of the seventh plurality of positioning base stations and the signal propagation involved in the eighth plurality of positioning base stations are converted from non-line-of-sight propagation to line-of-sight propagation during
  • FIG. 5B shows a flowchart of a detection method according to another embodiment of the present invention.
  • the detection method 560 shown in FIG. 5B can be implemented by any computing device having computing capabilities.
  • the computing device may be, but is not limited to, a desktop computer, a notebook computer, a tablet computer, a server, or a smart phone.
  • the detection method 560 may include, in block 562, when a certain positioning base station in the UWB positioning system receives an impulse response from a certain positioning tag, using the received impulse response to calculate multiple Specify the value of the impulse response characteristic.
  • the detection method 560 may further include, in block 564, based on the calculated values of the plurality of impulse response characteristics, using a trained classifier to determine a level of the ranging error of the certain positioning base station, where The classifier is trained to be able to classify the ranging errors of the arbitrary positioning base stations into corresponding levels based on the values of the plurality of impulse response characteristics of arbitrary positioning base stations in the ultra-wideband positioning system.
  • FIG. 6A shows a schematic diagram of a detection device according to an embodiment of the present invention.
  • the detection device 600 shown in FIG. 6A may be implemented by using software, hardware, or a combination of software and hardware.
  • the detection device 600 shown in FIG. 6A can be installed in any computing device having computing capabilities, for example.
  • the detection device 600 may include a first calculation module 602, a second calculation module 604, and a determination module 606.
  • the first calculation module 602 is configured to calculate a value of multiple specified impulse response characteristics by using a received impulse response when a certain positioning base station in an ultra-wideband positioning system receives an impulse response from a certain positioning tag this time. Values of the multiple specified impulse response characteristics of a certain positioning base station this time.
  • the second calculation module 604 is configured to calculate a value between the values of the multiple specified impulse response characteristics of the certain positioning base station this time and the values of the multiple specified impulse response characteristics of the certain positioning base station the previous time. The difference is used as the change amount of the plurality of specified impulse response characteristics of a certain positioning base station this time.
  • the determining module 606 is configured to determine whether the signal propagation involved in the current positioning of the certain positioning base station is right or wrong based on at least the change amount of the plurality of specified impulse response characteristics of the certain positioning base station this time.
  • Line-of-sight propagation wherein the classifier is trained to be able to classify the signal propagation that the arbitrary positioning base station participates in according to the change amount of the plurality of specified impulse response characteristics of any positioning base station in the ultra-wideband positioning system For line-of-sight or non-line-of-sight.
  • the classifier includes a first classifier and a second classifier, wherein the positive training samples and the negative training samples used to train the first classifier are signal propagations participating in the first plurality of positioning base stations, respectively.
  • the amount of change in the plurality of specified impulse response characteristics of the second plurality of positioning base stations during propagation, and the positive training samples and the negative training samples used to train the second classifier are in the third plurality of positioning, respectively
  • the signal propagation involved in the base station is converted from non-line-of-sight propagation to non-line-of-sight propagation during the non-line-of-sight propagation during the change in the plurality of specified impulse response characteristics of the plurality of positioning base stations and the signal propagation participating in the fourth plurality of positioning base stations from Non-line-
  • the classifier is a single classifier, wherein the positive training samples used to train the single classifier include a period during which signal propagation involved in a fifth plurality of positioning base stations transitions from line-of-sight propagation to non-line-of-sight propagation A change amount of the plurality of designated impulse response characteristics of the fifth plurality of positioning base stations and a signal propagation involved in the sixth plurality of positioning base stations is changed from non-line-of-sight propagation to non-line-of-sight propagation. The amount of change in the plurality of specified impulse response characteristics of the positioning base station, and the negative training samples used to train the classifier include signals transmitted during the seventh line of sight of the positioning base stations from line-of-sight propagation to line-of-sight propagation The change amount of the plurality of specified impulse response characteristics of the seventh plurality of positioning base stations and the signal propagation involved in the eighth plurality of positioning base stations are converted from non-line-of-sight propagation to line-of-sight propagation during the eighth plurality
  • FIG. 6B shows a schematic diagram of a detection device according to another embodiment of the present invention.
  • the detection device 660 shown in FIG. 6B may be implemented by using software, hardware, or a combination of software and hardware.
  • the detection device 660 shown in FIG. 6B may be installed in any computing device having computing capabilities, for example.
  • the detection device 660 may include a calculation module 662 and a determination module 664.
  • the calculation module 662 is configured to calculate a value of a plurality of specified impulse response characteristics by using a received impulse response when a certain positioning base station in an ultra-wideband positioning system receives an impulse response from a certain positioning tag.
  • the determining module 664 is configured to determine a level of a ranging error of the certain positioning base station by using a trained classifier based on the calculated values of the multiple impulse response characteristics, where the classifier is trained to be capable of The ranging errors of the arbitrary positioning base stations are classified into corresponding levels based on the values of the multiple impulse response characteristics of the arbitrary positioning base stations.
  • FIG. 7 shows a schematic diagram of a computing device according to an embodiment of the invention.
  • the computing device 700 may include a processor 702 and a memory 704 coupled with the processor 702.
  • the memory 704 stores executable instructions that, when executed, cause the processor 702 to execute any of the foregoing methods.
  • An embodiment of the present invention further provides a machine-readable storage medium on which executable instructions are stored. When the executable instructions are executed, the machine executes any of the foregoing methods.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Probability & Statistics with Applications (AREA)
  • Position Fixing By Use Of Radio Waves (AREA)

Abstract

一种检测方法（500）和检测装置（600），其中，检测方法（500）包括：当超宽带定位系统中的某一定位基站本次接收到脉冲响应时，利用所接收的脉冲响应，计算多个指定脉冲响应特征的值，作为本次某一定位基站的多个指定脉冲响应特征的值（502）；计算本次某一定位基站的多个指定脉冲响应特征的值和前一次某一定位基站的多个指定脉冲响应特征的值之间的差值，作为本次某一定位基站的多个指定脉冲响应特征的变化量（504）；至少基于本次某一定位基站的多个指定脉冲响应特征的变化量，利用已训练的分类器来确定某一定位基站本次参与的信号传播是否是非视距传播（506）。利用检测方法（500）和检测装置（600），能够减少利用分类器识别非视距传播的成本。

Description

检测方法和检测装置 技术领域
本发明涉及超宽带(UWB)定位领域，尤其涉及检测方法、检测装置、计算设备和机器可读存储介质。
背景技术
UWB定位是一种利用极窄的脉冲响应和1GHz以上带宽在室内对物体进行定位的技术。UWB定位系统包括多个定位基站和放置在要定位的对象上的定位标签。定位标签发送脉冲信号，该脉冲信号经过信道调制到达定位基站时变成脉冲响应。UWB定位系统利用定位基站接收的来自定位标签的脉冲响应来确定对象的定位。
当定位基站与定位标签之间的信号传播是没有障碍物阻挡的视距传播时，UWB定位系统可以获得对象的准确定位，然而，如果定位基站与定位标签之间的信号传播是有障碍物阻挡的非视距传播，那么UWB定位系统获得的定位是不准确的。因此，在UWB定位中，对非视距传播的识别是非常重要的。
目前已经出现利用机器学习模型作为分类器来识别UWB定位中的非视距传播的技术，其中，机器学习模型被训练为能够根据利用定位基站所接收的脉冲响应而计算的脉冲响应特征的值来将定位基站与定位标签之间的信号传播分类为视距传播或非视距传播。
然而，在不同的场合(例如，机场、停车场、火车站，学校等)中，利用定位基站接收的来自定位标签接收的脉冲响应而计算的脉冲响应特征的值具有很大的差异，因此，为了能训练得到在各种场合下通用的机器学习模型，需要收集在各种场合中定位基站接收的脉冲响应，以计算得到大量的脉冲响应特征的值作为训练样本数据来训练机器学习模型。从而，现有的利用机器学习模型识别非视距传播的成本较高。
发明内容
考虑到现有技术的以上问题，本发明的实施例提供检测方法、检测装置、计算设备和机器可读存储介质，其能够减少利用分类器识别非视距传播的成本。
本发明的实施例还提供检测方法、检测装置、计算设备和机器可读存储介质，其能够便于获知超带宽定位系统中的定位基站的测距误差大小。
按照本发明的实施例的一种检测方法，包括：当超宽带定位系统中的某一定位基站本次接收到来自某一定位标签的脉冲响应时，利用所接收的脉冲响应，计算多个指定脉冲响应特征的值，作为本次所述某一定位基站的所述多个指定脉冲响应特征的值；计算本次所述某一定位基站的所述多个指定脉冲响应特征的值和前一次所述某一定位基站的所述多个指定脉冲响应特征的值之间的差值，作为本次所述某一定位基站的所述多个指定脉冲响应特征的变化量；以及，至少基于本次所述某一定位基站的所述多个指定脉冲响应特征的变化量，利用已训练的分类器来确定所述某一定位基站本次参与的信号传播是否是非视距传播，其中，所述分类器被训练为能够根据所述超宽带定位系统中的任意定位基站的所述多个指定脉冲响应特征的变化量来将所述任意定位基站参与的信号传播分类为视距传播或非视距传播。
按照本发明的实施例的一种检测方法，包括：当超宽带定位系统中的某一定位基站接收到来自某一定位标签的脉冲响应时，利用所接收的脉冲响应，计算多个指定脉冲响应特征的值；基于所计算的所述多个脉冲响应特征的值，利用已训练的分类器来确定所述某一定位基站的测距误差的级别，其中，所述分类器被训练为能够基于所述超宽带定位系统的任意定位基站的所述多个脉冲响应特征的值将所述任意定位基站的测距误差分类为相应的级别。
按照本发明的实施例的一种检测装置，包括：第一计算模块，用于当超宽带定位系统中的某一定位基站本次接收到来自某一定位标签的脉冲响应时，利用所接收的脉冲响应，计算多个指定脉冲响应特 征的值，作为本次所述某一定位基站的所述多个指定脉冲响应特征的值；第二计算模块，用于计算本次所述某一定位基站的所述多个指定脉冲响应特征的值和前一次所述某一定位基站的所述多个指定脉冲响应特征的值之间的差值，作为本次所述某一定位基站的所述多个指定脉冲响应特征的变化量；以及，确定模块，用于至少基于本次所述某一定位基站的所述多个指定脉冲响应特征的变化量，利用已训练的分类器来确定所述某一定位基站本次参与的信号传播是否是非视距传播，其中，所述分类器被训练为能够根据所述超宽带定位系统中的任意定位基站的所述多个指定脉冲响应特征的变化量来将所述任意定位基站参与的信号传播分类为视距传播或非视距传播。
按照本发明的实施例的一种检测装置，包括：计算模块，用于当超宽带定位系统中的某一定位基站接收到来自某一定位标签的脉冲响应时，利用所接收的脉冲响应，计算多个指定脉冲响应特征的值；确定模块，用于基于所计算的所述多个脉冲响应特征的值，利用已训练的分类器来确定所述某一定位基站的测距误差的级别，其中，所述分类器被训练为能够基于所述任意定位基站的所述多个脉冲响应特征的值将所述任意定位基站的测距误差分类为相应的级别。
按照本发明的实施例的一种计算设备，包括：处理器；以及，存储器，其存储有可执行指令，所述可执行指令当被执行时使得所述处理器执行前述方法。
按照本发明的实施例的一种机器可读存储介质，其上具有可执行指令，当所述可执行指令被执行时，使得机器执行前述方法。
本发明的实施例的方案利用被训练为能够根据超宽带定位系统中的定位基站的脉冲响应特征的变化量来将该定位基站参与的信号传播分类为视距传播或非视距传播的分类器来确定超宽带定位系统中的定位基站参与的信号传播是否是非视距传播。由于利用在各种场合下收集的定位基站在不同时间接收的来自定位标签的脉冲响应计算得到的脉冲响应特征的变化相互之间差异很小，因而只需在一些典型场合中收集定位基站在不同时间接收的来自定位标签的脉冲响应，并利用所收集的脉冲响应计算得到脉冲响应特征的变化量作为训练 样本数据来训练用于识别非视距传播的分类器，就能得到在各种场合下通用的分类器。因此，与现有技术相比，本发明的实施例的方案能够减少利用分类器识别非视距传播的成本。
此外，本发明的实施例的方案利用被训练为能够根据超宽带定位系统中的定位基站的脉冲响应特征的值来将定位基站的测距误差分类为相应的级别的分类器来确定超宽带定位系统中的定位基站的测距误差的级别，很方便地就能够获知定位基站的测距误差大小，因此，与现有技术相比，本发明的实施例的方案能够便于获知定位基站的测距误差大小。
附图说明
本发明的其它特征、特点、益处和优点通过以下结合附图的详细描述将变得更加显而易见。其中：
图1示出了按照本发明的第一实施例的用于模型训练的方法的总体流程图；
图2示出了按照本发明的第一实施例的检测方法的总体流程图；
图3示出了按照本发明的第二实施例的用于模型训练的方法的总体流程图；
图4示出了按照本发明的第二实施例的检测方法的总体流程图；
图5A示出了按照本发明的一个实施例的检测方法的流程图；
图5B示出了按照本发明的另一个实施例的检测方法的流程图；
图6A示出了按照本发明的一个实施例的检测装置的示意图；
图6B示出了按照本发明的另一个实施例的检测装置的示意图；以及
图7示出了按照本发明的一个实施例的计算设备的示意图。
具体实施方式
下面将参考附图详细描述本发明的各个实施例。
一、非视距传播的识别
发明人经过大量的试验和研究发现，在UWB定位系统中，定位基站与定位标签之间的信号传播随时间变化存在四种变换方式，即：从视距传播变换为视距传播、从视距传播变换为非视距传播、从非视距传播变换为视距传播和从非视距传播变换为非视距传播，并且，在相同的信号传播变换方式下，利用在各种场合下收集的定位基站在不同时间接收的来自定位标签的脉冲响应计算得到的脉冲响应特征的变化相互之间差异很小。
基于以上发现，本发明的实施例利用脉冲响应特征的变化量作为训练样本数据来训练用于识别非视距传播的机器学习模型，只需在一些典型场合中收集定位基站在不同时间接收的来自定位标签的脉冲响应，并利用所收集的脉冲响应计算得到脉冲响应特征的变化量作为用于识别非视距传播的机器学习模型，就能得到在各种场合下通用的机器学习模型，这将能大大减少利用机器学习模型识别非视距传播的成本。
图1示出了按照本发明的第一实施例的用于模型训练的方法的总体流程图。图1所示的方法100可以由具有计算能力的任何计算设备来实现。该计算设备可以是但不局限于台式计算机、笔记本电脑、平板电脑、服务器或智能手机等。
如图1所示，在方框102，获取多个脉冲响应对。每一个脉冲响应对包括由位于多个场合中的某个场合的UWB定位系统中的某个定位基站接收的第一脉冲响应和第二脉冲响应，其中，该第一脉冲响应是该某个定位基站在某个时间在视距传播或非视距传播的情况下从某个定位标签中接收的，而该第二脉冲响应是该某个定位基站在该某个时间之后的另一个时间在视距传播或非视距传播的情况下从某个定位标签中接收的。例如但不局限于，该多个场合可以包括机场、火车站、停车场、大型购物中心等。
在方框106，将所获取的脉冲响应对分成第一信号组、第二信号组、第三信号组和第四信号组。其中，每一个信号组包括若干脉冲响 应对。
第一信号组对应于信号传播从视距传播变换为视距传播的情形，即，第一信号组所包括的任一脉冲响应对G1i中的第一脉冲响应g1i-1是在接收第一脉冲响应g1i-1的定位基站与发送跟第一脉冲响应g1i-1对应的脉冲信号的定位标签之间的信号传播(即，接收第一脉冲响应g1i-1的定位基站参与的信号传播)是视距传播的情况下被接收的，以及，该任一脉冲响应对G1i中的第二脉冲响应g1i-2是在接收第二脉冲响应g1i-2的定位基站与发送跟第二脉冲响应g1i-2对应的脉冲信号的定位标签之间的信号传播是视距传播的情况下被接收的。实际上，第一信号组所包括的多个脉冲响应对是在多个定位基站F1D参与的信号传播从视距传播变换为视距传播期间该多个定位基站F1D接收的。
第二信号组对应于信号传播从视距传播变换为非视距传播的情形，即，第二信号组所包括的任一脉冲响应对G2i中的第一脉冲响应g2i-1是在接收第一脉冲响应g2i-1的定位基站与发送跟第一脉冲响应g2i-1对应的脉冲信号的定位标签之间的信号传播是视距传播的情况下被接收的，以及，该任一脉冲响应对G2i中的第二脉冲响应g2i-2是在接收第二脉冲响应g2i-2的定位基站与发送跟第二脉冲响应g2i-2对应的脉冲信号的定位标签之间的信号传播是非视距传播的情况下被接收的。实际上，第二信号组所包括的多个脉冲响应对是在多个定位基站F2D参与的信号传播从视距传播变换为非视距传播期间该多个定位基站F2D接收的。
第三信号组对应于信号传播从非视距传播变换为视距传播的情形，即，第三信号组所包括的任一脉冲响应对G3i中的第一脉冲响应g3i-1是在接收第一脉冲响应g3i-1的定位基站与发送跟第一脉冲响应g3i-1对应的脉冲信号的定位标签之间的信号传播是非视距传播的情况下被接收的，以及，该任一脉冲响应对G3i中的第二脉冲响应g3i-2是在接收第二脉冲响应g3i-2的定位基站与发送跟第二脉冲响应g3i-2对应的脉冲信号的定位标签之间的信号传播是视距传播的情况下被接收的。实际上，第三信号组所包括的多个脉冲响应对是在多个定位 基站F3D参与的信号传播从视距传播变换为非视距传播期间该多个定位基站F3D接收的。
第四信号组对应于信号传播从非视距传播变换为非视距传播的情形，即，第四信号组所包括的任一脉冲响应对G4i中的第一脉冲响应g4i-1是在接收第一脉冲响应g4i-1的定位基站与发送跟第一脉冲响应g4i-1对应的脉冲信号的定位标签之间的信号传播是非视距传播的情况下被接收的，以及，该任一脉冲响应对G4i中的第二脉冲响应g4i-2是在接收第二脉冲响应g4i-2的定位基站与发送跟第二脉冲响应g4i-2对应的脉冲信号的定位标签之间的信号传播是非视距传播的情况下被接收的。实际上，第四信号组所包括的多个脉冲响应对是在多个定位基站F4D参与的信号传播从视距传播变换为非视距传播期间该多个定位基站F4D接收的。
在方框110，计算第一信号组、第二信号组、第三信号组和第四信号组各自的多个特征值向量对。其中，每一个信号组Cj的任一特征值向量对ck包括第一特征值向量和第二特征值向量，该任一特征值向量对ck中的第一特征值向量由利用信号组Cj所包括的其中一个脉冲响应对中的第一脉冲响应计算的多个指定脉冲响应特征PPT1的值构成，以及，该任一特征值向量对ck中的第二特征值向量由利用信号组Cj所包括的该其中一个脉冲响应对中的第二脉冲响应计算的该多个指定脉冲响应特征PPT1的值构成。
利用脉冲响应来计算脉冲响应特征的值在UWB定位领域是公知技术，在此省略对其的详细描述。例如但不局限于，该多个指定脉冲响应特征PPT1可以根据实际情况从以下脉冲响应特征中选取：定位基站与定位标签之间的距离，接收信号能量，最大幅度，最大幅度的上升时间，标准差，第一路径和最强路径的功率差，第一路径和最强路径的功率比，信噪比(SNR)、波形因数，接收脉冲峰值至开始时间延迟，平均超额延迟，均方时延扩展，峰度，波峰因数，峰值与平均功率比，以及，偏斜度。
在方框114，计算第一信号组、第二信号组、第三信号组和第四信号组各自的多个特征变化向量。每一个信号组Cj的任一特征变化 向量是信号组Cj的其中一个特征值向量对所包括的第一特征值向量和第二特征值向量之间的差值。这里，第一信号组的每一个特征变化向量表示该多个定位基站F1D的其中一个定位基站的该多个指定脉冲响应特征PPT1的变化量，第二信号组的每一个特征变化向量表示该多个定位基站F2D的其中一个定位基站的该多个指定脉冲响应特征PPT1的变化量，第三信号组的每一个特征变化向量表示该多个定位基站F3D的其中一个定位基站的该多个指定脉冲响应特征PPT1的变化量，以及，第四信号组的每一个特征变化向量表示该多个定位基站F4D的其中一个定位基站的该多个指定脉冲响应特征PPT1的变化量。
在方框118，使用第一信号组的特征变化向量作为负训练样本和第二信号组的特征变化向量作为正训练样本，来训练得到作为分类器的用于根据定位基站的该多个脉冲响应特征PPT1的变化量来将定位基站参与的信号传播分类为视距传播或非视距传播的第一机器学习模型M1，以及，使用第三信号组的特征变化向量作为负训练样本和第四信号组的特征变化向量作为正训练样本，来训练得到作为分类器的用于根据定位基站的该多个脉冲响应特征PPT1的变化量来将定位基站参与的信号传播分类为视距传播或非视距传播的第二机器学习模型M2。
其中，第一机器学习模型M1适用于定位基站参与的信号传播在变换前为视距传播的情形，而第二机器学习模型M2适用于定位基站参与的信号传播在变换前为非视距传播的情形。例如但不局限于，第一机器学习模型M1和第二机器学习模型M2可以是决策树、神经网络、支持向量机等。
图2示出了按照本发明的第一实施例的检测方法的总体流程图。图2所示的方法200可以由具有计算能力的任何计算设备来实现。该计算设备可以是但不局限于台式计算机、笔记本电脑、平板电脑、服务器或智能手机等。
如图2所示，在方框202，当UWB定位系统中的某个定位基站 CP本次接收到来自某个定位标签BQ的脉冲响应PUL时，利用所接收的脉冲响应PUL，计算该多个指定响应信号特征PPT1的值，作为本次定位基站CP的该多个指定脉冲响应特征PPT1的值。
在方框206，计算本次定位基站CP的该多个指定脉冲响应特征PPT1的值和前一次定位基站CP的该多个指定脉冲响应特征PPT1的值之间的差值，作为本次定位基站CP的该多个指定脉冲响应特征PPT1的变化量。
在方框210，根据定位基站CP前一次参与的信号传播是视距传播还是非视距传播，从已训练的第一机器学习模型M1和第二机器学习模型M2中选择相应的机器学习模型。这里，当定位基站CP前一次参与的信号传播是视距传播时第一学习模型M1被选择，以及，当定位基站CP前一次参与的信号传播是非视距传播时第二学习模型M2被选择。
在方框214，将本次定位基站CP的该多个指定脉冲响应特征PPT1的变化量输入所选择的机器学习模型，以确定定位基站CP本次参与的信号传播(即，定位基站CP与定位标签BQ之间的信号传播)是否是非视距传播。
二、定位基站的测距误差的级别的确定
到达时间差(TDOA)定位是一种常用的无线定位技术，其利用多个基站从待测对象接收到的信号的时间差来计算该待测对象的位置。当在UWB定位系统中应用TDOA定位技术来计算定位标签的位置时，在多个定位基站都接收到来自同一定位标签的脉冲响应的情况下，有时会出现基于该多个定位基站中的某些定位基站所接收的脉冲响应而计算的定位标签的位置是准确，但基于该多个定位基站中的另外一些定位基站所接收的脉冲响应而计算的定位标签的位置却是不准确。
发明人经过大量分析发现，在UWB定位系统中，如果接收到来自同一定位标签的脉冲响应的几个定位基站的测距误差相互之间相差不大，那么使用TDOA定位技术基于这几个定位基站所接收的脉 冲响应而计算的定位标签的位置基本上是准确，然而，如果接收到来自同一定位标签的脉冲响应的几个定位基站的测距误差相互之间相差很大，那么使用TDOA定位技术基于这几个定位基站所接收的脉冲响应而计算的定位标签的位置是不准确，这是因为在接收到来自同一定位标签的脉冲响应的几个定位基站的测距误差相互之间相差不大的情况下，当使用TDOA定位技术基于这几个定位基站所接收的脉冲响应来计算定位标签的位置时，这几个定位基站的测距误差基本上能相互抵消掉，从而使得所计算的定位标签的位置基本上是准确的。这里，定位基站的测距误差是指利用该定位基站接收的来自定位标签的脉冲响应而计算的该定位基站与该定位标签之间的估算距离与该定位基站与该定位标签之间的真实距离之差。
因此，如果能够事先知道定位基站的测距误差，那么在使用TDOA定位技术计算定位标签的位置时，选用测距误差基本上相同的那些定位基站接收的脉冲响应来计算定位标签的位置，就能计算得到定位标签的真实位置。然而，定位基站的测距误差通常不是固定不变的，现有技术没有提供便于确定定位基站的测距误差大小的方案。
发明人经过大量研究发现，如果定位基站与定位标签之间的信号传播线路被遮挡得越严重，那么定位基站的测距误差越大，并且，如果定位基站与定位标签之间的信号传播线路的被遮挡严重程度发生变化，那么基于该定位基站接收的来自该定位标签的脉冲响应而计算的脉冲响应特征的值也相应地发生变化。
基于以上发现，本发明的实施例将定位基站的测距误差划分成多个不同的测距误差范围并给每一个测距误差范围赋予一个不同的级别，训练得到作为分类器的用于根据定位基站的脉冲响应特征的值来将定位基站的测距误差分类为相应的级别的机器学习模型，然后利用所训练的机器学习模型来确定定位基站的测距误差的级别，从而便于获知定位基站的测距误差大小。
图3示出了按照本发明的第二实施例的用于模型训练的方法的总体流程图。图3所示的方法300可以由具有计算能力的任何计算设 备来实现。该计算设备可以是但不局限于台式计算机、笔记本电脑、平板电脑、服务器或智能手机等。
如图3所示，在方框302，获取多个定位基站PBS接收的多个脉冲响应PPS和针对该多个定位基站PBS各自的真实距离。
其中，该多个定位基站PBS是一个或多个UWB定位系统中的定位基站。该多个脉冲响应PPS中的每一个脉冲响应是该多个定位基站PBS的其中一个定位基站从某个定位标签接收的。针对该多个定位基站PBS中的任一定位基站的真实距离表示当该任一定位基站接收到该多个脉冲响应PPS中的该任一定位基站所接收的那个脉冲响应PPSi时该任一定位基站与发送跟脉冲响应PPSi对应的脉冲信号的那个定位标签之间的真正距离。例如但不局限于，该任一定位基站与发送跟脉冲响应PPSi对应的脉冲信号的定位标签之间的真实距离可以通过对安装在该任一定位基站所处的场合中的照相设备所拍摄的图像进行图像处理后获得。
在方框306，计算针对该多个定位基站PBS各自的估算距离。
其中，针对该多个定位基站PBS中的任一定位基站的估算距离表示利用该多个脉冲响应PPS中的该任一定位基站所接收的那个脉冲响应PPSi计算的在该任一定位基站接收到脉冲响应PPSi时该任一定位基站与发送跟脉冲响应PPSi对应的脉冲信号的那个定位标签之间的距离。例如但不局限于，可以计算该任一定位基站接收到脉冲响应PPSi的时刻和脉冲响应PPSi中携带的其被发送的时刻之间的时间差与光速的乘积，作为针对该任一定位基站的估算距离。
在方框310，计算该多个定位基站PBS各自的测距误差。该多个定位基站PBS中的任一定位基站的测距误差表示针对该任一定位基站的估算距离与真实距离之间的差值的绝对值。
在方框314，设置测距误差的多个级别L与多个不同的测距误差范围R的映射关系YG。在该映射关系YG中，该多个级别L中的每一个级别对应于该多个测距误差范围R的其中一个测距误差范围。例如但不局限于，该多个测距误差范围R可以包括0～20厘米、20～40厘米和40～60厘米三个测距误差范围。
在方框318，根据该多个定位基站PBS各自的测距误差所落入的测距误差范围和映射关系YG，确定该多个定位基站PBS各自的测距误差的级别。
在方框322，利用该多个脉冲响应PPS中的该多个定位基站PBS的任一定位基站所接收的那个脉冲响应，计算多个指定脉冲响应特征PPT2的值，作为该任一定位基站的该多个指定脉冲响应特征PPT2的值，从而得到该多个定位基站PBS各自的该多个指定脉冲响应特征PPT2的值。
该多个指定脉冲响应特征PPT2可以根据实际情况从以下脉冲响应特征中选取：定位基站与定位标签之间的距离，接收信号能量，最大幅度，最大幅度的上升时间，标准差，第一路径和最强路径的功率差，第一路径和最强路径的功率比，信噪比(SNR)、波形因数，接收脉冲峰值至开始时间延迟，平均超额延迟，均方时延扩展，峰度，波峰因数，峰值与平均功率比，以及，偏斜度。
在方框326，使用该多个定位基站PBS各自的该多个指定脉冲响应特征PPT2的值和测距误差的级别作为训练样本数据，训练得到作为分类器的机器学习模型M3。机器学习模型M3被训练为能够根据任意定位基站的该多个脉冲响应特征PPT2的值将该任意定位基站的测距误差分类为相应的级别。
图4示出了按照本发明的第二实施例的检测方法的总体流程图。图4所示的方法400可以由具有计算能力的任何计算设备来实现。该计算设备可以是但不局限于台式计算机、笔记本电脑、平板电脑、服务器或智能手机等。
如图4所示，在方框402，当UWB定位系统中的某个定位基站CP接收到来自某个定位标签BQ的脉冲响应PUL时，利用所接收的脉冲响应PUL，计算该多个指定响应信号特征PPT2的值，作为定位基站CP的该多个指定响应信号特征PPT2的值。
在方框406，将定位基站CP的该多个指定响应信号特征PPT2的值输入已训练的机器学习模型M3，以确定定位基站CP的测距误 差的级别。
所确定的定位基站CP的测距误差的级别可以显示出来供用户使用。
当UWB定位系统中的多个定位基站接收到来自同一定位标签的脉冲响应时，可以利用方法400确定该多个定位基站各自的测距误差的级别，然后可以利用该多个定位基站中的测距误差的级别相同的那些定位基站接收的脉冲响应通过TDOA定位技术来确定定位标签的位置，这与利用测距误差的级别不同的那些定位基站接收的脉冲信号通过TDOA定位技术进行定位相比，能够有效地提高定位准确性。
其它变型
本领域技术人员将理解，虽然在上面的第一实施例中，用于根据定位基站的脉冲响应特征的变化量来将定位基站参与的信号传播分类为视距传播或非视距传播的机器学习模型包括两个模型，即：第一机器学习模型M1和第一机器学习模型M2，然而，本发明并不局限于此。在本发明的其它一些实施例中，用于根据定位基站的脉冲响应特征的变化量来将定位基站参与的信号传播分类为视距传播或非视距传播的机器学习模型也可以是单个机器学习模型。在单个机器学习模型的情况下，可以使用第一信号组和第三信号组的特征变化向量作为负训练样本以及第二信号组和第四信号组的特征变化向量作为正训练样本，来训练得到该单个机器学习模型，并且，方法200不包括方框210。
本领域技术人员将理解，方法100仅是训练用于根据定位基站的脉冲响应特征的变化量来将定位基站参与的信号传播分类为视距传播或非视距传播的机器学习模型的示例方法，还可以采用其它任何合适的方法来训练用于根据定位基站的脉冲响应特征的变化量来将定位基站参与的信号传播分类为视距传播或非视距传播的机器学习模型。
本领域技术人员将理解，方法300仅是训练用于根据定位基站的脉冲响应特征的值将定位基站的测距误差分类为相应的级别的机器 学习模型的示例方法，还可以采用其它任何合适的方法来训练用于根据定位基站的脉冲响应特征的值将定位基站的测距误差分类为相应的级别的机器学习模型。
本领域技术人员将理解，虽然在上面的第一和第二实施例中，所使用的分类器是机器学习模型(即，第一机器学习模型M1、第一机器学习模型M2和机器学习模块M3)，然而，本发明并不局限于此。在本发明的其它一些实施例中，所使用的分类器也可以是除了机器学习模型之外的任何合适的其它类型的分类器。
图5A示出了按照本发明的一个实施例的检测方法的流程图。图5A所示的检测方法500可以由具有计算能力的任何计算设备来实现。该计算设备可以是但不局限于台式计算机、笔记本电脑、平板电脑、服务器或智能手机等。
如图5A所示，检测方法500可以包括，在方框502，当超宽带定位系统中的某一定位基站本次接收到来自某一定位标签的脉冲响应时，利用所接收的脉冲响应，计算多个指定脉冲响应特征的值，作为本次所述某一定位基站的所述多个指定脉冲响应特征的值。
检测方法500还可以包括，在方框504，计算本次所述某一定位基站的所述多个指定脉冲响应特征的值和前一次所述某一定位基站的所述多个指定脉冲响应特征的值之间的差值，作为本次所述某一定位基站的所述多个指定脉冲响应特征的变化量。
检测方法500还可以包括，在方框506，至少基于本次所述某一定位基站的所述多个指定脉冲响应特征的变化量，利用已训练的分类器来确定所述某一定位基站本次参与的信号传播是否是非视距传播，其中，所述分类器被训练为能够根据所述超宽带定位系统中的任意定位基站的所述多个指定脉冲响应特征的变化量来将所述任意定位基站参与的信号传播分类为视距传播或非视距传播。
在一个方面，所述分类器包括第一分类器和第二分类器，其中，训练所述第一分类器使用的正训练样本和负训练样本分别是在第一多个定位基站参与的信号传播从视距传播转换为非视距传播期间所 述第一多个定位基站的所述多个指定脉冲响应特征的变化量和在第二多个定位基站参与的信号传播从视距传播转换为视距传播期间所述第二多个定位基站的所述多个指定脉冲响应特征的变化量，以及，训练所述第二分类器使用的正训练样本和负训练样本分别是在第三多个定位基站参与的信号传播从非视距传播转换为非视距传播期间所述第三多个定位基站的所述多个指定脉冲响应特征的变化量和在第四多个定位基站参与的信号传播从非视距传播转换为视距传播期间所述第四多个定位基站的所述多个指定脉冲响应特征的变化量，以及，其中，确定所述某一定位基站本次参与的信号传播是否是非视距传播包括：根据所述某一定位基站前一次参与的信号传播是视距传播还是非视距传播，从所述第一分类器和所述第二分类器中选择相应的分类器，其中，当所述某一定位基站前一次参与的信号传播是视距传播时所述第一分类器被选择，否则所述第二分类器被选择；以及，基于本次所述某一定位基站的所述多个指定脉冲响应特征的变化量，利用所选择的分类器来确定所述某一定位基站本次参与的信号传播是否是非视距传播。
在另一个方面，所述分类器是单个分类器，其中，训练所述分类器使用的正训练样本包括在第五多个定位基站参与的信号传播从视距传播转换为非视距传播期间所述第五多个定位基站的所述多个指定脉冲响应特征的变化量和在第六多个定位基站参与的信号传播从非视距传播转换为非视距传播期间所述第六多个定位基站的所述多个指定脉冲响应特征的变化量，以及，训练所述分类器使用的负训练样本包括在第七多个定位基站参与的信号传播从视距传播转换为视距传播期间所述第七多个定位基站的所述多个指定脉冲响应特征的变化量和在第八多个定位基站参与的信号传播从非视距传播转换为视距传播期间所述第八多个定位基站的所述多个指定脉冲响应特征的变化量。
图5B示出了按照本发明的另一个实施例的检测方法的流程图。图5B所示的检测方法560可以由具有计算能力的任何计算设备来实 现。该计算设备可以是但不局限于台式计算机、笔记本电脑、平板电脑、服务器或智能手机等。
如图5B所示，检测方法560可以包括，在方框562，当超宽带定位系统中的某一定位基站接收到来自某一定位标签的脉冲响应时，利用所接收的脉冲响应，计算多个指定脉冲响应特征的值。
检测方法560还可以包括，在方框564，基于所计算的所述多个脉冲响应特征的值，利用已训练的分类器来确定所述某一定位基站的测距误差的级别，其中，所述分类器被训练为能够基于所述超宽带定位系统中的任意定位基站的所述多个脉冲响应特征的值将所述任意定位基站的测距误差分类为相应的级别。
图6A示出了按照本发明的一个实施例的检测装置的示意图。图6A所示的检测装置600可以利用软件、硬件或软硬件结合的方式来执行。图6A所示的检测装置600例如可以安装在具有计算能力的任何计算设备中。
如图6A所示，检测装置600可以包括第一计算模块602、第二计算模块604和确定模块606。第一计算模块602用于当超宽带定位系统中的某一定位基站本次接收到来自某一定位标签的脉冲响应时，利用所接收的脉冲响应，计算多个指定脉冲响应特征的值，作为本次所述某一定位基站的所述多个指定脉冲响应特征的值。第二计算模块604用于计算本次所述某一定位基站的所述多个指定脉冲响应特征的值和前一次所述某一定位基站的所述多个指定脉冲响应特征的值之间的差值，作为本次所述某一定位基站的所述多个指定脉冲响应特征的变化量。确定模块606用于至少基于本次所述某一定位基站的所述多个指定脉冲响应特征的变化量，利用已训练的分类器来确定所述某一定位基站本次参与的信号传播是否是非视距传播，其中，所述分类器被训练为能够根据所述超宽带定位系统中的任意定位基站的所述多个指定脉冲响应特征的变化量来将所述任意定位基站参与的信号传播分类为视距传播或非视距传播。
在一个方面，所述分类器包括第一分类器和第二分类器，其中， 训练所述第一分类器使用的正训练样本和负训练样本分别是在第一多个定位基站参与的信号传播从视距传播转换为非视距传播期间所述第一多个定位基站的所述多个指定脉冲响应特征的变化量和在第二多个定位基站参与的信号传播从视距传播转换为视距传播期间所述第二多个定位基站的所述多个指定脉冲响应特征的变化量，以及，训练所述第二分类器使用的正训练样本和负训练样本分别是在第三多个定位基站参与的信号传播从非视距传播转换为非视距传播期间所述第三多个定位基站的所述多个指定脉冲响应特征的变化量和在第四多个定位基站参与的信号传播从非视距传播转换为视距传播期间所述第四多个定位基站的所述多个指定脉冲响应特征的变化量，以及，所述确定模块包括：选择模块，用于根据所述某一定位基站前一次参与的信号传播是视距传播还是非视距传播，从所述第一分类器和所述第二分类器中选择相应的分类器，其中，当所述某一定位基站前一次参与的信号传播是视距传播时所述第一分类器被选择，否则所述第二分类器被选择；以及，分类模块，用于基于本次所述某一定位基站的所述多个指定脉冲响应特征的变化量，利用所选择的分类器来确定所述某一定位基站本次参与的信号传播是否是非视距传播。
在另一个方面，所述分类器是单个分类器，其中，训练所述单个分类器使用的正训练样本包括在第五多个定位基站参与的信号传播从视距传播转换为非视距传播期间所述第五多个定位基站的所述多个指定脉冲响应特征的变化量和在第六多个定位基站参与的信号传播从非视距传播转换为非视距传播期间所述第六多个定位基站的所述多个指定脉冲响应特征的变化量，以及，训练所述分类器使用的负训练样本包括在第七多个定位基站参与的信号传播从视距传播转换为视距传播期间所述第七多个定位基站的所述多个指定脉冲响应特征的变化量和在第八多个定位基站参与的信号传播从非视距传播转换为视距传播期间所述第八多个定位基站的所述多个指定脉冲响应特征的变化量。
图6B示出了按照本发明的另一个实施例的检测装置的示意图。 图6B所示的检测装置660可以利用软件、硬件或软硬件结合的方式来执行。图6B所示的检测装置660例如可以安装在具有计算能力的任何计算设备中。
如图6B所示，检测装置660可以包括计算模块662和确定模块664。计算模块662用于当超宽带定位系统中的某一定位基站接收到来自某一定位标签的脉冲响应时，利用所接收的脉冲响应，计算多个指定脉冲响应特征的值。确定模块664用于基于所计算的所述多个脉冲响应特征的值，利用已训练的分类器来确定所述某一定位基站的测距误差的级别，其中，所述分类器被训练为能够基于所述任意定位基站的所述多个脉冲响应特征的值将所述任意定位基站的测距误差分类为相应的级别。
图7示出了按照本发明的一个实施例的计算设备的示意图。如图7所示，计算设备700可以包括处理器702和与处理器702耦合的存储器704。其中，存储器704存储有可执行指令，所述可执行指令当被执行时使得处理器702执行前述的任意方法。
本发明实施例还提供一种机器可读存储介质，其上存储可执行指令，所述可执行指令当被执行时使得机器执行前述的任意方法。
本领域技术人员应当理解，上面所公开的各个实施例可以在不偏离发明实质的情况下做出各种变形、修改和改变，这些变形、修改和改变都应当落入在本发明的保护范围之内。因此，本发明的保护范围由所附的权利要求书来限定。

Claims (10)

  1. 一种检测方法，包括：
    当超宽带定位系统中的某一定位基站本次接收到来自某一定位标签的脉冲响应时，利用所接收的脉冲响应，计算多个指定脉冲响应特征的值，作为本次所述某一定位基站的所述多个指定脉冲响应特征的值；
    计算本次所述某一定位基站的所述多个指定脉冲响应特征的值和前一次所述某一定位基站的所述多个指定脉冲响应特征的值之间的差值，作为本次所述某一定位基站的所述多个指定脉冲响应特征的变化量；以及
    至少基于本次所述某一定位基站的所述多个指定脉冲响应特征的变化量，利用已训练的分类器来确定所述某一定位基站本次参与的信号传播是否是非视距传播，
    其中，所述分类器被训练为能够根据所述超宽带定位系统中的任意定位基站的所述多个指定脉冲响应特征的变化量来将所述任意定位基站参与的信号传播分类为视距传播或非视距传播。
  2. 如权利要求1所述的检测方法，其中，
    所述分类器包括第一分类器和第二分类器，其中，训练所述第一分类器使用的正训练样本和负训练样本分别是在第一多个定位基站参与的信号传播从视距传播转换为非视距传播期间所述第一多个定位基站的所述多个指定脉冲响应特征的变化量和在第二多个定位基站参与的信号传播从视距传播转换为视距传播期间所述第二多个定位基站的所述多个指定脉冲响应特征的变化量，以及，训练所述第二分类器使用的正训练样本和负训练样本分别是在第三多个定位基站参与的信号传播从非视距传播转换为非视距传播期间所述第三多个定位基站的所述多个指定脉冲响应特征的变化量和在第四多个定位基站参与的信号传播从非视距传播转换为视距传播期间所述第四多个定位基站的所述多个指定脉冲响应特征的变化量，以及
    其中，确定所述某一定位基站本次参与的信号传播是否是非视距传播包括：
    根据所述某一定位基站前一次参与的信号传播是视距传播还是非视距传播，从所述第一分类器和所述第二分类器中选择相应的分类器，其中，当所述某一定位基站前一次参与的信号传播是视距传播时所述第一分类器被选择，否则所述第二分类器被选择；以及
    基于本次所述某一定位基站的所述多个指定脉冲响应特征的变化量，利用所选择的分类器来确定所述某一定位基站本次参与的信号传播是否是非视距传播。
  3. 如权利要求1所述的检测方法，其中，
    所述分类器是单个分类器，
    其中，训练所述单个分类器使用的正训练样本包括在第五多个定位基站参与的信号传播从视距传播转换为非视距传播期间所述第五多个定位基站的所述多个指定脉冲响应特征的变化量和在第六多个定位基站参与的信号传播从非视距传播转换为非视距传播期间所述第六多个定位基站的所述多个指定脉冲响应特征的变化量，以及，训练所述分类器使用的负训练样本包括在第七多个定位基站参与的信号传播从视距传播转换为视距传播期间所述第七多个定位基站的所述多个指定脉冲响应特征的变化量和在第八多个定位基站参与的信号传播从非视距传播转换为视距传播期间所述第八多个定位基站的所述多个指定脉冲响应特征的变化量。
  4. 一种检测方法，包括：
    当超宽带定位系统中的某一定位基站接收到来自某一定位标签的脉冲响应时，利用所接收的脉冲响应，计算多个指定脉冲响应特征的值；
    基于所计算的所述多个脉冲响应特征的值，利用已训练的分类器来确定所述某一定位基站的测距误差的级别，
    其中，所述分类器被训练为能够基于所述超宽带定位系统的任意定位基站的所述多个脉冲响应特征的值将所述任意定位基站的测距 误差分类为相应的级别。
  5. 一种检测装置，包括：
    第一计算模块，用于当超宽带定位系统中的某一定位基站本次接收到来自某一定位标签的脉冲响应时，利用所接收的脉冲响应，计算多个指定脉冲响应特征的值，作为本次所述某一定位基站的所述多个指定脉冲响应特征的值；
    第二计算模块，用于计算本次所述某一定位基站的所述多个指定脉冲响应特征的值和前一次所述某一定位基站的所述多个指定脉冲响应特征的值之间的差值，作为本次所述某一定位基站的所述多个指定脉冲响应特征的变化量；以及
    确定模块，用于至少基于本次所述某一定位基站的所述多个指定脉冲响应特征的变化量，利用已训练的分类器来确定所述某一定位基站本次参与的信号传播是否是非视距传播，
    其中，所述分类器被训练为能够根据所述超宽带定位系统中的任意定位基站的所述多个指定脉冲响应特征的变化量来将所述任意定位基站参与的信号传播分类为视距传播或非视距传播。
  6. 如权利要求5所述的检测装置，其中，
    所述分类器包括第一分类器和第二分类器，其中，训练所述第一分类器使用的正训练样本和负训练样本分别是在第一多个定位基站参与的信号传播从视距传播转换为非视距传播期间所述第一多个定位基站的所述多个指定脉冲响应特征的变化量和在第二多个定位基站参与的信号传播从视距传播转换为视距传播期间所述第二多个定位基站的所述多个指定脉冲响应特征的变化量，以及，训练所述第二分类器使用的正训练样本和负训练样本分别是在第三多个定位基站参与的信号传播从非视距传播转换为非视距传播期间所述第三多个定位基站的所述多个指定脉冲响应特征的变化量和在第四多个定位基站参与的信号传播从非视距传播转换为视距传播期间所述第四多个定位基站的所述多个指定脉冲响应特征的变化量，以及
    其中，所述确定模块包括：
    选择模块，用于根据所述某一定位基站前一次参与的信号传播是视距传播还是非视距传播，从所述第一分类器和所述第二分类器中选择相应的分类器，其中，当所述某一定位基站前一次参与的信号传播是视距传播时所述第一分类器被选择，否则所述第二分类器被选择；以及
    分类模块，用于基于本次所述某一定位基站的所述多个指定脉冲响应特征的变化量，利用所选择的分类器来确定所述某一定位基站本次参与的信号传播是否是非视距传播。
  7. 如权利要求5所述的检测装置，其中，
    所述分类器是单个分类器，
    其中，训练所述单个分类器使用的正训练样本包括在第五多个定位基站参与的信号传播从视距传播转换为非视距传播期间所述第五多个定位基站的所述多个指定脉冲响应特征的变化量和在第六多个定位基站参与的信号传播从非视距传播转换为非视距传播期间所述第六多个定位基站的所述多个指定脉冲响应特征的变化量，以及，训练所述分类器使用的负训练样本包括在第七多个定位基站参与的信号传播从视距传播转换为视距传播期间所述第七多个定位基站的所述多个指定脉冲响应特征的变化量和在第八多个定位基站参与的信号传播从非视距传播转换为视距传播期间所述第八多个定位基站的所述多个指定脉冲响应特征的变化量。
  8. 一种检测装置，包括：
    计算模块，用于当超宽带定位系统中的某一定位基站接收到来自某一定位标签的脉冲响应时，利用所接收的脉冲响应，计算多个指定脉冲响应特征的值；
    确定模块，用于基于所计算的所述多个脉冲响应特征的值，利用已训练的分类器来确定所述某一定位基站的测距误差的级别，
    其中，所述分类器被训练为能够基于所述任意定位基站的所述多 个脉冲响应特征的值将所述任意定位基站的测距误差分类为相应的级别。
  9. 一种计算设备，包括：
    处理器；以及
    存储器，其存储有可执行指令，所述可执行指令当被执行时使得所述处理器执行权利要求1-4中的任意一个所述的方法。
  10. 一种机器可读存储介质，其上具有可执行指令，当所述可执行指令被执行时，使得机器执行权利要求1-4中的任意一个所述的方法。
PCT/CN2018/095339 2018-07-12 2018-07-12 检测方法和检测装置 Ceased WO2020010558A1 (zh)

Priority Applications (4)

Application Number Priority Date Filing Date Title
CN201880095467.5A CN112425220B (zh) 2018-07-12 2018-07-12 检测方法和检测装置
PCT/CN2018/095339 WO2020010558A1 (zh) 2018-07-12 2018-07-12 检测方法和检测装置
US17/257,965 US11500055B2 (en) 2018-07-12 2018-07-12 Detection method and detection apparatus
EP18925748.8A EP3823372A4 (en) 2018-07-12 2018-07-12 DETECTION METHOD AND DETECTION DEVICE

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2018/095339 WO2020010558A1 (zh) 2018-07-12 2018-07-12 检测方法和检测装置

Publications (1)

Publication Number Publication Date
WO2020010558A1 true WO2020010558A1 (zh) 2020-01-16

Family

ID=69143016

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/095339 Ceased WO2020010558A1 (zh) 2018-07-12 2018-07-12 检测方法和检测装置

Country Status (4)

Country Link
US (1) US11500055B2 (zh)
EP (1) EP3823372A4 (zh)
CN (1) CN112425220B (zh)
WO (1) WO2020010558A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114374985A (zh) * 2020-10-14 2022-04-19 Oppo广东移动通信有限公司 超带宽信号的接收信号强度修正方法与装置、电子设备
EP4204841A2 (en) * 2020-08-31 2023-07-05 QUALCOMM Incorporated Unsupervised learning for simultaneous localization and mapping in deep neural networks using channel state information

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11551006B2 (en) * 2019-09-09 2023-01-10 International Business Machines Corporation Removal of personality signatures
US12082147B2 (en) 2020-09-18 2024-09-03 Samsung Electronics Co., Ltd. Line of sight (LoS)/non-line of sight (NLoS) point identification in wireless communication networks using artificial intelligence
CN113660639B (zh) * 2021-08-18 2023-07-14 电子科技大学 高速磁浮列车地面基站通信探测一体化系统及方法
CN115184859B (zh) * 2022-06-27 2024-06-21 南京森林警察学院 一种构建非视距传播场景下测距和测角误差消除方法
CN115278876B (zh) * 2022-09-19 2022-12-09 四川创智联恒科技有限公司 一种在5g网络和uwb共同定位的方法
CN117056818B (zh) * 2023-10-08 2024-02-02 广东工业大学 一种基于cir第一路径的超宽带nlos信号识别方法

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103399297A (zh) * 2013-08-16 2013-11-20 哈尔滨工业大学深圳研究生院 一种基于机器学习的超宽带非视距鉴别方法
CN103684524A (zh) * 2013-12-13 2014-03-26 北京邮电大学 一种基于uwb通信信号的目标识别方法
US20160048710A1 (en) * 2013-07-12 2016-02-18 Lawrence Livermore National Security, Llc Long-range uwb remote powering capability at fcc regulated limit using multiple antennas
CN105911539A (zh) * 2016-04-19 2016-08-31 北京邮电大学 一种到达时间差的误差补偿值确定方法及装置
CN107037402A (zh) * 2017-05-24 2017-08-11 南京图贝斯韦智能技术有限公司 一种基于uwb测距的室内非视距环境下的定位方法
CN107205226A (zh) * 2017-04-18 2017-09-26 上海交通大学 基于信道分类的室内定位跟踪方法及系统

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2008249396A (ja) * 2007-03-29 2008-10-16 Brother Ind Ltd 移動局の静止判定を利用した位置検出システム
US9002286B2 (en) * 2009-03-30 2015-04-07 Massachusetts Institute Of Technology Method and system for identification and mitigation of errors in non-line-of-sight distance estimation
US9229102B1 (en) * 2009-12-18 2016-01-05 L-3 Communications Security And Detection Systems, Inc. Detection of movable objects
US9720069B2 (en) * 2012-10-10 2017-08-01 Lg Electronics Inc. Apparatus and method for measuring location of user equipment located indoors in wireless network
KR20160107007A (ko) * 2015-03-03 2016-09-13 삼성전자주식회사 혈압 측정 장치 및 방법
CN105445699B (zh) * 2015-11-13 2018-04-13 哈尔滨工业大学深圳研究生院 一种非视距误差消除的测距方法及系统
CN107734479A (zh) * 2017-09-11 2018-02-23 广东广业开元科技有限公司 一种基于无线传感技术的消防人员定位方法、系统和装置
CN112567834B (zh) * 2018-08-14 2023-11-24 罗伯特·博世有限公司 用于检测信号传播类型的方法和装置

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160048710A1 (en) * 2013-07-12 2016-02-18 Lawrence Livermore National Security, Llc Long-range uwb remote powering capability at fcc regulated limit using multiple antennas
CN103399297A (zh) * 2013-08-16 2013-11-20 哈尔滨工业大学深圳研究生院 一种基于机器学习的超宽带非视距鉴别方法
CN103684524A (zh) * 2013-12-13 2014-03-26 北京邮电大学 一种基于uwb通信信号的目标识别方法
CN105911539A (zh) * 2016-04-19 2016-08-31 北京邮电大学 一种到达时间差的误差补偿值确定方法及装置
CN107205226A (zh) * 2017-04-18 2017-09-26 上海交通大学 基于信道分类的室内定位跟踪方法及系统
CN107037402A (zh) * 2017-05-24 2017-08-11 南京图贝斯韦智能技术有限公司 一种基于uwb测距的室内非视距环境下的定位方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3823372A4 *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP4204841A2 (en) * 2020-08-31 2023-07-05 QUALCOMM Incorporated Unsupervised learning for simultaneous localization and mapping in deep neural networks using channel state information
CN114374985A (zh) * 2020-10-14 2022-04-19 Oppo广东移动通信有限公司 超带宽信号的接收信号强度修正方法与装置、电子设备
CN114374985B (zh) * 2020-10-14 2023-09-19 Oppo广东移动通信有限公司 超带宽信号的接收信号强度修正方法与装置、电子设备

Also Published As

Publication number Publication date
EP3823372A4 (en) 2023-02-01
EP3823372A1 (en) 2021-05-19
US11500055B2 (en) 2022-11-15
CN112425220A (zh) 2021-02-26
CN112425220B (zh) 2023-05-30
US20210223351A1 (en) 2021-07-22

Similar Documents

Publication Publication Date Title
CN112425220B (zh) 检测方法和检测装置
CN107659893B (zh) 一种误差补偿方法、装置、电子设备及可读存储介质
CN106793087B (zh) 一种基于aoa和pdoa的阵列天线室内定位方法
CN102928813A (zh) 基于rssi质心加权算法的无源rfid标签定位方法
CN106093852A (zh) 一种提高WiFi指纹定位精度与效率的方法
CN108828501B (zh) 在室内声场环境中对移动声源进行实时跟踪定位的方法
US7050652B2 (en) Methods and arrangements to enhance correlation
JP2011214920A (ja) 位置推定装置、位置推定方法及びプログラム
CN112929826A (zh) 一种基于uwb三角定位与指纹信息的室内混合定位方法
CN112927565B (zh) 提升机坪综合航迹监视数据精度的方法、装置及系统
CN107205226B (zh) 基于信道分类的室内定位跟踪方法及系统
CN115767427B (zh) 一种基于众包机制的多径辅助蓝牙aoa室内定位方法
WO2023281506A2 (en) System and method for determining object location
CN108989988B (zh) 基于机器学习的室内定位方法
CN114051209A (zh) 一种基于智能反射面和场景几何模型的指纹定位方法
CN116761252A (zh) 一种uwb室内定位系统基站自动标定的方法
JP2014106725A (ja) 点群解析装置、点群解析方法及び点群解析プログラム
WO2017054673A1 (zh) 基于无线网络的定位方法和定位装置
US11791911B2 (en) Method and apparatus for detecting signal propagation type
US10012729B2 (en) Tracking subjects using ranging sensors
Qu et al. OODT: LOS signal identification for acoustic indoor localization from stream perspective
CN111836188A (zh) 一种基于Wi-Fi RSS的在线协同定位及系统
US10061012B2 (en) Locating and ranging using coherent array reconciliation tomography
CN108055504B (zh) 一种智能室内巡逻装置
US20250294516A1 (en) System and method using 5g positioning with non-line of sight information from sensors

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: 18925748

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE