Method for processing perception data

Abstract

This invention provides a method for processing sensing data. The method includes: mapping sensing data to multiple grid regions to obtain grid-level data corresponding to the multiple grid regions; determining the association relationship between multiple sensing targets corresponding to the sensing data based on the grid-level data; and, if the association relationship determines that two sensing targets are the same sensing target, integrating the sensing data corresponding to the two sensing targets. This invention solves the problems of high computational load and low processing efficiency in existing methods for matching and associating sensing targets from multiple sites, thereby improving the accuracy of sensing data.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of communications, and more specifically, to a method for processing sensing data. Background Technology

[0002] Currently, single-station sensing technology is relatively mature, capable of distinguishing different sensing targets with high accuracy. However, the sensing range of a single station is limited. When a sensing target moves continuously over a large area, multiple stations within the region need to collaborate using sensing technology to jointly sense the target and achieve more complete tracking of its trajectory.

[0003] In collaborative sensing technologies, different stations often cannot determine which targets they identify are the same and which are different, requiring matching and association methods. However, existing methods for matching and associating sensing targets from multiple stations suffer from high computational cost and low processing efficiency. Summary of the Invention

[0004] This invention provides a method for processing sensing data, which at least solves the problems of large computational load and low processing efficiency in existing methods for matching and associating sensing targets from multiple sites in related technologies.

[0005] According to an embodiment of the present invention, a method for processing sensing data is provided, comprising: mapping sensing data to multiple grid regions to obtain grid-level data corresponding to the multiple grid regions; determining the association relationship between multiple sensing targets corresponding to the sensing data based on the grid-level data; and, if it is determined based on the association relationship that there are two sensing targets that are the same sensing target, integrating the sensing data corresponding to the two sensing targets.

[0006] According to yet another embodiment of the present invention, a computer-readable storage medium is also provided, wherein a computer program is stored therein, wherein the computer program is configured to perform the steps in any of the above method embodiments when executed.

[0007] According to yet another embodiment of the present invention, an electronic device is also provided, including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the steps in any of the above method embodiments.

[0008] According to yet another embodiment of the present invention, a computer program product is also provided, comprising a computer program that, when executed by a processor, implements the steps in any of the above method embodiments.

[0009] Through the above embodiments of the present invention, since all sensing data is mapped to multiple grid regions, data processing is facilitated on a grid region-by-grid basis, reducing the amount of data computation and improving data processing efficiency. Then, the correlation between the data within each individual grid region is determined, and the sensing data is integrated based on this correlation to obtain more accurate sensing data. Therefore, this solves the problems of high computational load and low processing efficiency in existing methods for matching and associating sensing targets across multiple sites, thereby improving the accuracy of sensing data. Attached Figure Description

[0010] Figure 1 This is a hardware structure block diagram of the network device used in the embodiments of the method of the present invention;

[0011] Figure 2 This is a schematic diagram of a network architecture for communication-aware networking according to an embodiment of the present invention;

[0012] Figure 3 This is a flowchart of a method for processing perceived data according to an embodiment of the present invention;

[0013] Figure 4 This is a schematic diagram of a station sensing a target according to a scenario embodiment of the present invention;

[0014] Figure 5 This is a flowchart of target matching according to an embodiment of the present invention;

[0015] Figure 6 This is a flowchart of target matching according to an embodiment of the present invention. Detailed Implementation

[0016] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings and examples.

[0017] It should be noted that the terms "first," "second," etc., in the specification, claims, and drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0018] The methods and embodiments provided in this application can be executed in a network device or a similar computing device. Taking running on a network device as an example, Figure 1 This is a hardware structure block diagram of the network device used in the embodiments of the method of the present invention. For example... Figure 1 As shown, a network device may include one or more ( Figure 1Only one is shown in the diagram. A processor 102 (which may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.) and a memory 104 for storing data are also shown. The network device may further include a transmission device 106 for communication functions and an input / output device 108. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the network device described above. For example, the network device may also include components that are more complex than those described above. Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0019] The memory 104 can be used to store computer programs, such as application software programs and modules, like the computer program corresponding to the data processing method in this embodiment of the invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, thus implementing the above-described method. The memory 104 may include high-speed random access memory and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to network devices via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0020] The transmission device 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by a communication provider for the network device. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 may be a Radio Frequency (RF) module used for wireless communication with the Internet.

[0021] In the 5G wireless air interface, the areas requiring research in sensor fusion include: sensing reference signal design, sensing resource allocation, multi-antenna sensing beam management, single-site sensing technology, cooperative sensing technology, sensing measurement design, and measurement feedback, to meet the diverse needs of sensing scenarios. In particular, when the sensed target moves continuously over a large area, multiple sites within the region need to jointly sense the target. However, different sites often cannot determine which targets they identify are the same and which are different, necessitating a method for matching and associating them.

[0022] Existing matching and association methods can determine whether the sensing targets perceived by multiple stations are the same by directly transmitting information between stations, or by centrally processing information through a centralized sensing function (SF) to determine whether the sensing targets perceived by multiple stations are the same.

[0023] For centralized SF processing, the processing object is the sensing data of all its subordinate sensing devices, resulting in a large data volume. Therefore, associating sensing targets is particularly important. Conventional sensing target association methods are based on initial judgment and screening of sensing data according to time difference and location difference, and then calculate the distance between each pair of sensing targets for further detailed judgment to determine whether the sensing targets can be associated. However, this method involves a large number of data combinations, a large amount of computation, low processing efficiency, and a certain impact on performance. In addition, if there are conflicts in the sensing results, a secondary verification and re-judgment are required, further increasing the workload.

[0024] To address the issues of high computational load and low processing efficiency in conventional target association methods, this invention proposes a spatial grid-based inter-station target association method in multi-site network scenarios. In this method, a Sensing Function (SF) controls multiple subordinate sensing devices to sense targets, collect sensing measurement data, perform centralized processing, and finally output the sensing results. Specifically, the method includes: first, dividing the sensing area into spatial grids and mapping all sensing data to corresponding grid areas (referred to as grids); then, performing association judgments on the data between pairs of stations within a single grid area. For example, if factors such as time and distance meet the conditions, the sensing targets between the corresponding two stations are considered to be matched; otherwise, they are not matched (matching means identical); finally, comprehensively judging the results within each grid area, outputting the final target association result, and optimizing and merging some sensing data based on this result to achieve more complete and accurate trajectory tracking of targets over a larger area.

[0025] By mapping data to raster regions, data is broken down into smaller parts. Pairwise correlation judgments are performed only on data within each raster region, significantly reducing the number of data combinations and computational load, thus improving overall data processing efficiency. Based on the combined judgment of results from each raster region, if conflicts arise, raster information allows for faster and more accurate problem identification and correction. Furthermore, the raster-based approach can also improve efficiency to some extent for optimizing and integrating multi-station sensing data of the same target in overlapping coverage areas.

[0026] The inter-station target association method of this invention can be run in Figure 2 In the network architecture of the communication-aware networking shown, such as Figure 2As shown, the network architecture includes: sensing network element (SF), sensing device (gNB) and sensing target (UE), where inter-site refers to the communication between sensing devices.

[0027] The Sensing Element (SF) is used to send control messages to its subordinate sensing devices to control the activation or deactivation of their sensing functions. It also collects and centrally processes the sensing data uploaded by each device, ultimately integrating and outputting the final sensing results. If the SF detects an anomaly in the data processing of a sensing device, it can also send a message to that device instructing it to adjust its sensing configuration.

[0028] Sensing devices may include multiple ones, such as Figure 2 As shown, it includes gNB1 to gNB n n is an integer greater than 1, used to activate the sensing function based on a certain mechanism, to sense the sensing targets within its coverage area, and to upload sensing data to SF; it is also used to adjust the relevant configuration of the sensing function based on the message indication of SF.

[0029] The sensing device in this embodiment of the invention can distinguish different sensing targets within its own area.

[0030] Perceived targets also include multiple ones, such as Figure 2 As shown, including UE1 to UE2 m m is an integer greater than 1, used to be sensed by the sensing device.

[0031] Figure 3 This is a flowchart of a method for processing sensing data according to an embodiment of the present invention, that is, a flowchart of an inter-station target association method, as shown below. Figure 3 As shown, the process includes the following steps:

[0032] Step S302: Map the sensing data to multiple grid regions to obtain grid-level data corresponding to multiple grid regions;

[0033] In some embodiments, the sensing data is sensing data received by the sensing network element from multiple sensing devices, and the sensing data includes at least one of the following: sensing device identifier, sensing target identifier, the time point of sensing data collection, and the location coordinate information of the sensing target.

[0034] In some exemplary embodiments, the collected single-site (single gNB) sensing data includes at least: sensing device identifier, sensing target identifier, sensing data collection time point, and sensing target location coordinate information, i.e., location coordinates (longitude, latitude, altitude), such as {gNBID, UEID, time, loc}.

[0035] After raster mapping based on the location coordinate information in any single-station sensing data, information similar to that shown in Table 1 is obtained:

[0036] Table 1

[0037]

[0038] Before step S302 in this embodiment, the method further includes: dividing the sensing area into spatial grids to obtain multiple grid areas; wherein the sensing area is determined according to the control area corresponding to the sensing network element; or, determined according to the position boundary corresponding to the sensing data of all sensing devices collected by the sensing network element.

[0039] Since the target being sensed may be moving in the air, such as a drone, the spatial grid division of the sensing area is a three-dimensional spatial grid division.

[0040] In some embodiments, spatial grid division of the sensing region includes one of the following: dividing the sensing region into spatial grids according to a preset first distance interval; or dividing the sensing region into spatial grids according to latitude / longitude differences and height differences; or dividing the sensing region into multiple sub-sensing regions and spatially dividing the multiple sub-sensing regions into spatial grids according to a preset second distance interval; or spatial grid division of the sensing region through machine learning. The first distance interval or the second distance interval can be a distance interval with equal length / width / height values, or a distance interval with unequal length / width / height values.

[0041] In some embodiments, the sensing area is spatially gridded, and the method further includes: determining the data volume and data density of new grid-level data based on the mapping of new sensing data in multiple grid areas; and re-dividing the sensing area into spatial grids in response to the case where the data volume of new grid-level data exceeds a first preset threshold and / or the data density of new grid-level data exceeds a second preset threshold.

[0042] In some exemplary embodiments, the spatial grid size is a*b*c, which can be defined in multiple ways and can be selected according to specific circumstances, for example:

[0043] Divide the space into equal sections of 20m x 20m x 20m, or;

[0044] Divide the sensing area into 0.1 degrees * 0.1 degrees * 20 meters based on the latitude and longitude difference and altitude difference, or;

[0045] Divide the area into zones based on distance, for example: 20m*20m*20m below 500 meters altitude, 20m*20m*50m above 500 meters altitude, or;

[0046] Based on the perceived data, a grid is adaptively divided using machine learning. The higher the density of the perceived data, the smaller the granularity of the grid division. The machine learning methods used can be, but are not limited to, K-Means clustering and DBSCAN clustering.

[0047] In some embodiments, the spatial grid division of the sensing region is further performed according to different grid division methods and precision requirements, including:

[0048] 1) Fixed division.

[0049] a. No Update: Once the grid is initially divided according to a certain rule, it is used continuously without re-dividing or adjusting the grid boundaries. For example, after dividing the grid into 20m*20m*20m sections based on equal Euclidean distances, the single grid area is always kept at 20m*20m*20m.

[0050] b. Condition Update: After initial raster division according to a certain rule, while collecting new sensing data for raster mapping, the amount and density of sensing data at the raster level are statistically evaluated. When a certain threshold is exceeded, raster re-division is triggered. Re-division can target only the problematic raster / region, or all raster regions.

[0051] 2) Adaptive partitioning, such as K-Means clustering.

[0052] a. Each update: Raster division and data mapping are always based on the latest collected perception dataset.

[0053] b. Condition Update: After initial raster division based on sensing data, when collecting new sensing data, data mapping is first performed based on the existing raster, while the amount and density of sensing data at the raster level are statistically evaluated. When a certain threshold is exceeded, raster re-division is triggered. Re-division can target only the problematic raster / region, or all raster regions.

[0054] Step S304: Determine the correlation between multiple sensing targets corresponding to the sensing data based on the grid-level data.

[0055] In step S304 of this embodiment, for any grid area, based on the grid-level data in the grid area, perception target matching is performed on each pair of perception targets corresponding to the grid area. A pair of perception targets includes two perception targets, and the two perception targets correspond to different perception devices. Perception target matching is used to determine whether the two perception targets in each pair of perception targets are the same perception target. Based on the result of perception target matching for each grid area, the association relationship between multiple perception targets is determined. The perception target matching result is one of the following: exact matching result, no matching result, or fuzzy matching result.

[0056] In some embodiments, performing perception target matching for each pair of perception targets corresponding to the grid area includes: stopping perception target matching in response to the grid-level data of the grid area containing only the perception data of one perception device; and performing perception target matching for the pair of perception targets corresponding to any two perception devices based on the perception data of any two perception devices in the grid area when the grid-level data of the grid area contains the perception data of multiple perception devices.

[0057] In this embodiment, since the target within a single station (i.e., containing only one sensing device) is clear, no further processing is required.

[0058] In some exemplary embodiments, for any grid containing a sensing device, the grid region outputs a result of "0: no match".

[0059] In some embodiments, the scenario where the output result indicates no match may be due to the fact that the grid location is covered by only one station, or that the grid is covered by multiple stations, but only one of the stations activates the sensing function when the target passes through the grid.

[0060] In some embodiments, when each of the multiple sensing devices senses only one sensing target, sensing target matching is performed on the sensing target pairs corresponding to any two sensing devices based on the sensing data of any two sensing devices in the grid area. This includes: determining the sensing time difference and sensing distance difference for each pair of sensing target pairs corresponding to the sensing device that senses only one sensing target, based on the sensing data of any two sensing devices; for any pair of sensing target pairs corresponding to the sensing device that senses only one sensing target, if the sensing time difference corresponding to the pair of sensing target pairs is less than or equal to a third preset threshold and the sensing distance difference is less than or equal to a fourth preset threshold, determining the sensing target matching result as a precise matching result; wherein, the precise matching result is used to indicate that the two sensing targets in the pair of sensing target pairs are the same sensing target; if the sensing time difference corresponding to the pair of sensing target pairs is greater than the third preset threshold and / or the sensing distance difference is greater than the fourth preset threshold, determining the sensing target matching result as a non-matching result; wherein, the non-matching result is used to indicate that the two sensing targets in the pair of sensing target pairs are different sensing targets.

[0061] In some exemplary embodiments, there are multiple stations, each corresponding to a sensing target. The sensing targets are related as stations. Based on the sensing data of any two sensing devices, a mapping and matching judgment is performed on the two sensing targets corresponding to those two sensing devices. Here, an inter-station relationship means that two sensing targets correspond to two different sensing devices. If the mapping conditions are met, such as the time difference and distance difference between the two sensing targets being less than a specified threshold, the result for the sensing target pair corresponding to those two sensing targets is "1: Exact Match". Otherwise, the result is "3: No Match".

[0062] In some embodiments, the scenario where all sensing targets are inter-station relationships may correspond to the following scenarios: when the same sensing target passes through the grid, it is simultaneously sensed by multiple stations. In this case, these multiple sets of sensing data are usually very close in time and space, satisfying the mapping condition and can be matched; or different targets pass through the grid at different times and are sensed by different single stations respectively. In this case, the mapping condition is not satisfied and they are not matched.

[0063] In some embodiments, when at least one of the two sensing devices can sense multiple sensing targets, sensing target matching is performed on the sensing target pairs corresponding to any two sensing devices based on the sensing data of any two sensing devices in the grid area. This includes: based on the sensing data of any two sensing devices, for any sensing target corresponding to any two sensing devices, determining the sensing time difference and sensing distance difference of all sensing target pairs corresponding to that sensing target; if there is only one set of sensing target pairs among all sensing target pairs whose sensing time difference is less than or equal to a third preset threshold and whose sensing distance difference is less than or equal to a fourth preset threshold, then the sensing target matching result is determined to be a precise matching result; wherein, the precise matching result is used... If the two perceived targets in a given pair are identified as the same perceived target, and if multiple pairs of perceived target pairs have a perception time difference less than or equal to a third preset threshold and a perception distance difference less than or equal to a fourth preset threshold, the result of the perceived target matching is determined to be a fuzzy matching result. The fuzzy matching result indicates that it is impossible to determine if the two perceived targets in the multiple pairs of perceived target pairs are the same perceived target. If the perception time difference of all pairs of perceived target pairs is greater than the third preset threshold, and / or all perception distance differences are greater than the fourth preset threshold, the result of the perceived target matching is determined to be a non-matching result. The non-matching result indicates that all perceived targets in the multiple pairs of perceived target pairs are different perceived targets.

[0064] In some exemplary embodiments, there are multiple stations, and at least one station has multiple target data. Therefore, pairwise judgments are performed on all inter-station relationships between perceived targets (i.e., perceived target pairs). Since one station may have multiple perceived targets, a single perceived target will have multiple processing results. For example, if station 1 has two perceived targets and station 2 has one perceived target, then the perceived target of station 2 needs to be judged separately with the two perceived targets of station 1. The possible judgment results include:

[0065] 1) If the mapping condition is met and the result is unique, then the result of the perceived target pair is "1: exact match".

[0066] 2) If the mapping condition is met, but not unique, then the result of the corresponding perceived target pair is "2: Fuzzy Match".

[0067] 3) If the mapping condition is not met, the result of the perception target pair is "3: mismatch".

[0068] In various embodiments of the present invention, two sensing targets with inter-station relationships are defined as a sensing target pair. Except for a grid area containing only one station, each grid area contains one or more sensing target pairs. In the above embodiments, the sensing targets in the grid area are mapped and matched, that is, the sensing target pairs in the grid area are mapped and matched.

[0069] In some embodiments, for a pair of perceived targets whose output result is a fuzzy match, the multi-grid comprehensive judgment method can be used to further determine whether the pair of perceived targets matches the mapping.

[0070] In some embodiments, for a pair of perceived targets whose output results are exactly matched, further judgment can be made through multi-grid integrated judgment to verify whether the perceived target pair matches the mapping.

[0071] In some embodiments, determining the association between multiple sensing targets corresponding to sensing data based on the sensing target matching results of each grid region includes: determining the matching degree value of each sensing target pair based on the sensing target matching results of each sensing target pair in different grid regions and a preset weight value, wherein the preset weight value corresponds to the sensing target matching results in different grid regions; performing uniqueness verification on each sensing target pair based on the matching degree value, and determining the two sensing targets corresponding to the sensing target pair that passes the uniqueness verification as the same sensing target.

[0072] In some exemplary embodiments, the mapping matching results of all perceived target pairs within a single grid are comprehensively judged, and the final perceived target association result is output.

[0073] 1. For any pair of sensed targets, calculate the matching degree of the pair of sensed targets, and combine it with the judgment threshold to output the initial mapping result.

[0074] For example: matching degree = ratio1 * weight1 + ratio2 * weight2 + ratio3 * weight3,

[0075] If the matching degree ≥ Thresholdl, it can be mapped.

[0076] If the matching degree ≤ Threshold2, it cannot be mapped.

[0077] If Threshold2 < matching degree < Threshold1, further judgment is required.

[0078] Among them, ratioX is the ratio of the number of grids with result = X, and the value range is [0, 1]; weightX is the weight value of result = X, and there are multiple choices for the value range, for example, [0, 1] or [-1, 1].

[0079] In this embodiment, if the weight of a certain result is 0, it means that this value is not currently concerned in the rule, and the number of grids with this value is not included in the denominator when calculating the ratios of other values.

[0080] For example: If weight1 - 3 are all not 0, then ratio1 = (the number of grids with result = 1) / (the number of grids with result = 1 + the number of grids with result = 2 + the number of grids with result = 3). If weight2 = 0, then ratio1 = (the number of grids with result = 1) / (the number of grids with result = 1 + the number of grids with result = 3).

[0081] By adjusting the configuration of weight and Threshold to adjust the judgment conditions, the tendency and severity of the mapping rule can be controlled.

[0082] Illustrate with an example:

[0083] If weigh1 = 1, weigh2 = 0, weigh3 = -1, Threshold1 = 1, Threshold2 = -1, the initial mapping result shown in Table 2 can be obtained:

[0084] Table 2

[0085] 1: Exact Match 2: Fuzzy matching 3: Mismatch Match Mapping results √(Exists) 1 (Not paying attention) × (Does not exist) 1 Mappable × - √ -1 Unmapping √0.9 - √0.1 0.8 Further assessment is needed. × √ × 0 Further assessment is needed.

[0086] In some embodiments, after obtaining the initial mapping result, uniqueness verification can also be performed on the initial mapping result.

[0087] In some exemplary embodiments, if station 1 corresponds to multiple sensing targets (including sensing target 1) and station 2 corresponds to multiple sensing targets (including sensing target 2), then the uniqueness condition may include:

[0088] 1) There is a mapping between the perception target 1 of station 1 and the perception target 2 of station 2.

[0089] 2) If multiple sensing targets are reported by station 1, then other sensing targets are not mappable to sensing target 2 (it may be explicitly unmappable, or a fuzzy match that cannot be determined at the moment).

[0090] 3) If multiple sensing targets are reported by station 2, then the other sensing targets are not mappable to sensing target 1 (it may be explicitly unmappable, or a fuzzy match that cannot be determined at the moment).

[0091] If the uniqueness condition described above is met, the mapping relationship of the perceived target pair is output as the final result. Otherwise, a mapping conflict is considered to exist, and a secondary matching judgment is required for further identification.

[0092] For example, if perception target 1 is not only mappable to perception target 2, but also mappable to other perception targets x in station 2, then further mapping matching judgment needs to be performed on perception target 1, i.e., secondary matching judgment.

[0093] In some embodiments, a secondary matching judgment is performed on the sensing target pairs that do not meet the uniqueness condition, including: if the sensing data of the sensing target contains velocity estimation information, then the set of sensing target pairs with the closest velocity is output as the unique mappable relationship, and the other sets are modified to be unmappable; or, the trajectory directions of each sensing target within the station are compared, and the set of sensing target pairs with the closest direction or the smallest angle is output as the unique mappable relationship, and the other sets are modified to be unmappable.

[0094] If, after a series of matching checks, there are still more than one unmatched sensing target, then pair them up between stations, and then check each pair of sensing target pairs to see if they can be mapped.

[0095] 1) Report whether the two stations of the target are adjacent. If so, continue to 2); otherwise, exit.

[0096] 2) Does the raster containing the perceived data of the target exist in adjacent or similar raster cells? If so, continue to 3); otherwise, exit.

[0097] Because the judgment is based on spatial grids, it is possible that the data of the same sensing target being sensed by different stations are mapped to two adjacent grids, which is why it is marked as no match.

[0098] To determine whether grid cells are adjacent or close, we can rely on their grid index numbers, i.e., the difference between their x, y, or z values ​​must be within a certain range. For example, if x and y are equal, and the difference in z is 1, then they are adjacent grid cells.

[0099] 3) If the perceived data of the target satisfies the mapping conditions, such as the time difference and distance difference being less than the specified threshold, then continue to 4); otherwise, exit.

[0100] 4) Determine the estimated velocity of the perceived target or the angle between its trajectory and the target. If the conditions are met, it is considered that it can be mapped; otherwise, it cannot be mapped.

[0101] Step S306: If it is determined that there are two sensing targets that are the same sensing target based on the correlation relationship, the sensing data corresponding to the two sensing targets are integrated and processed.

[0102] In some embodiments, when it is determined that there are two sensing targets that are the same sensing target based on the association relationship, the sensing data is integrated, including: for any pair of sensing targets, when the two sensing targets corresponding to the pair of sensing targets are the same sensing target, the sensing data in the overlapping grid area corresponding to the pair of sensing targets is integrated; wherein, the overlapping grid area is the grid area that both sensing targets in the pair of sensing targets have reached; and the integrated sensing data is used as the final movement trajectory of the sensing target.

[0103] Specifically, for any pair of sensing targets, the overlapping area corresponding to the pair of sensing targets is the data of the two sensing targets that contain the pair of sensing targets. When the two sensing targets are the same sensing target, the data of the two sensing targets in the overlapping area are integrated, that is, the data of the two sensing targets in the overlapping area are merged.

[0104] For example, if sensing target 1 and sensing target 2 are the same sensing target, then the trajectories of sensing target 1 and sensing target 2 are the same trajectory of sensing target 1. The data of sensing target 1 and sensing target 2 are present in each grid cell of the overlapping area, and the data in the grid cells can be merged.

[0105] Through the matching and determination in the previous embodiments, a unique mapping relationship can be obtained between the sensing targets at different stations. Based on these mapping relationships, similar trajectory points of the same sensing target can be integrated and optimized, and the trajectories of sensing targets can also be integrated and presented as a whole.

[0106] 1. Partial integration and optimization of sensing data

[0107] In overlapping coverage areas of stations, multiple stations perceive the same target and output perception results. For data from nearby points, data can be filtered or merged based on certain principles to obtain the integrated trajectory of the perceived target. For example, the arithmetic mean of multiple sets of perception data belonging to the same grid for the same perceived target can be taken, and this result can be used to replace and update the original sets of perception data within the grid.

[0108] 2. Integration and presentation of perceived target trajectories

[0109] The final tracking trajectory of the target is displayed on the interface, which is the result of integrating and optimizing all the sensing data from multiple stations for the same target, after partial integration of sensing data.

[0110] In some embodiments, after determining the association between multiple sensing targets corresponding to sensing data, the method further includes: if it is determined that there is a sensing target pair with an abnormal sensing target matching based on the association between the multiple sensing targets, sending a control message to the corresponding sensing device, wherein the control message is used to instruct the sensing device to adjust the sensing configuration information.

[0111] In some exemplary embodiments, the overall sensing target matching results are evaluated. If an anomaly is found in the sensing target matching at a certain site, the sensing function configuration of that site may have a problem and needs to be optimized and adjusted. At this time, SF sends a control message to the problematic site, containing information instructing it to adjust its sensing configuration.

[0112] Site-based target matching anomaly identification: Within the sensing area, the site is not located at a boundary and is not an isolated area with adjacent sites, but the proportion of unmatched targets under the site is very high, or even all targets are unmatched. Unmatched targets refer to targets passing through overlapping coverage areas between sites without a match, not simply targets appearing only in the same raster cell along their path. This is because it's possible that a target doesn't traverse multiple sites but only moves within the scope of a single site, in which case no match is normal.

[0113] Through the above steps, mapping all sensing data to multiple grid regions facilitates data processing on a grid-region basis, reducing computational load and improving processing efficiency. Then, the relationships between data within each individual grid region are determined, and sensing data is integrated based on these relationships to obtain more accurate sensing data. Therefore, this method solves the problems of high computational load and low processing efficiency in existing methods for matching and associating sensing targets across multiple sites, thus improving the accuracy of sensing data.

[0114] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0115] It should be noted that the above modules can be implemented by software or hardware. For the latter, they can be implemented in the following ways, but are not limited to: all the above modules are located in the same processor; or, the above modules are located in different processors in any combination.

[0116] Figure 4 This is a schematic diagram illustrating a station sensing a target according to an embodiment of the present invention, such as... Figure 4 As shown, the two stations overlap and cover each other, sensing two sensing targets.

[0117] Figure 5 This is a flowchart of target matching according to an embodiment of the present invention, such as... Figure 5 The process shown includes the following steps:

[0118] Step S11: Divide the sensing area controlled by SF into spatial grids and map the sensing data of the two stations into the divided grid areas.

[0119] The spatial grid division of the sensing area controlled by SF includes:

[0120] 1. The raster index is defined as (x, y, z). Sensing area boundaries: longitude range {Xmin, Xmax}, latitude range {Ymin, Ymax}, altitude range {Zmin, Zmax}. Raster size: 20m * 20m * 20m.

[0121] 2. Determining z: Based on Zmin, a height layer is divided every 20m, and z is the number of the height layer, starting from 0. For example, in the location information loc of sensing data 1, if Zmin <= height < (Zmin + 20), then the z corresponding to sensing data 1 is 0, that is, the index of its grid is (x, y, 0).

[0122] 3. Determining y: Calculate the latitude difference deltaY across each grid cell based on a grid size of 20m. Using Ymin as a base, divide the area into latitude layers every deltaY, with y representing the layer number, starting from 0. For example, in the location information loc of sensing data 1, if Ymin <= latitude < (Ymin + deltaY), then the y value corresponding to sensing data 1 is 0, meaning its grid index is (x, 0, z).

[0123] 4. Determining x: The method is the same as for y.

[0124] Where deltaY = (20*180) / (π*R), deltaX = (20*180) / (π*R*cosY), and R is the Earth's radius of 6,371,000 m. The calculation of the longitude difference is related to the latitude circle. For simplification, Y takes a fixed value, Ymin or Ymax, or an intermediate value Ymid. High precision needs to be maintained when calculating the longitude and latitude difference.

[0125] Step S12: Perform target matching for any pair of perceived targets within any grid area.

[0126] For ease of distinction, the sensing targets of site 1 are denoted as UE11 and UE12, and the sensing targets of site 2 are denoted as UE21 and UE22.

[0127] At this point, for the overlapping coverage area, there are multiple sets of raster data. It is necessary to determine four combinations of different sites: UE11 and UE21, UE11 and UE22, UE12 and UE21, and UE12 and UE22; do not determine UE11 and UE12 or UE21 and UE22 of the same site.

[0128] Mapping conditions: The time difference between the sensing data of two sensing targets is less than or equal to the time threshold, and the distance between the predicted locations is less than or equal to the distance threshold.

[0129] The possible outcomes of the processing result are shown in Table 3:

[0130] Table 3

[0131]

[0132] Step S13: Combine the matching results of sensing targets in multiple grid areas to obtain the inter-station sensing target association results.

[0133] For each group of perceived targets (regardless of order, UE11&UE21 and UE21&UE11 are in the same group), the judgment is based on the rules in Table 4, or the judgment can be made by referring to the above-mentioned embodiment of the perceived target matching result determination. UE11&UE21 and UE12&UE22 both satisfy Case 1.

[0134] Table 4

[0135] 1: Exact Match 2: Fuzzy matching 3: Mismatch Mapping results Case 1 √ - × Mappable, e.g., UE11 = UE21 Case 2 × - √ Unmappable, e.g., UE11 ≠ UE21 Case 3 √ - √ Further assessment is needed. Case 4 × √ × Further assessment is needed.

[0136] After uniqueness verification, the mapping relationship is obtained: UE11 = UE21, UE12 = UE22.

[0137] Step S14: Optimize and integrate the sensing data in multiple grid areas corresponding to the same sensing target.

[0138] Perceptual data optimization: Taking UE11 = UE21 as an example, the grid contains data points {gNB1, UE11, time11, loc11} and {gNB2, UE21, time21, loc21} for both UE11 and UE21. After merging, the output data point is {UE identifier, (time11+time21) / 2, (loc11+loc21) / 2}. The new data point is retained, and the two original data points are removed.

[0139] Based on all processed sensing data, a complete tracking trajectory of the sensed target is presented. For example, UE11 = UE21 is denoted as UE. A UE12 = UE22 is denoted as UE B Then the interface will only display UE. A and UE B Two tracking trajectories.

[0140] In one scenario embodiment, three sites have overlapping coverage and two sensing targets are perceived. The sites are gNB1, gNB2, and gNB3. The two sensing targets are denoted as {UE11, UE12}, {UE21, UE22}, and {UE31, UE32}, respectively.

[0141] Figure 6 This is a flowchart of target matching according to an embodiment of the present invention, such as... Figure 6 As shown, the process includes the following steps:

[0142] Step S21: Divide the sensing area controlled by SF into spatial grids, and map the sensing data of the three stations to the divided grid areas. The specific method of dividing the spatial grids can be referred to in the above embodiment, and will not be repeated here.

[0143] Step S22: Perform target matching for target pairs within any grid area.

[0144] Step S23: Combine the matching results of sensing targets in multiple grid areas to obtain the inter-station sensing target association results.

[0145] Based on the judgment principle given in step S13, the following uniqueness verification is performed to obtain the mapping relationship: UE11 = UE21, UE12 = UE22.

[0146] Neither UE31 nor UE32 has any mapped sensing targets.

[0147] Step S24: Optimize and integrate the sensing data in multiple grid areas corresponding to the same sensing target.

[0148] For the identified mappable sensing targets: UE11=UE21 and UE12=UE22, grid-level data integration and overall trajectory presentation are performed respectively.

[0149] Step S25: Detect and process abnormal data.

[0150] Perform anomaly detection on gNB3 data. If the following conditions are met simultaneously, gNB3 is considered to be abnormal, triggering SF to send a control message to it, instructing it to adjust its sensing configuration.

[0151] 1) gNB3 has overlapping coverage areas with gNB1 / gNB2.

[0152] 2) The activation time of the sensing function of gNB3 overlaps with that of gNB1 / gNB2.

[0153] 3) During the overlapping time period of 2), 1) there are UE31 / UE32 perception data reported in the grid covering the overlapping area.

[0154] 4) In the perception data matching results of 3), the proportion of "3: mismatch" exceeds a certain threshold.

[0155] Through the above embodiments of the present invention, in the scenario of multi-station sensing network, the sensing data of a single station is subjected to grid-level mapping relationship judgment, and the inter-station sensing target association relationship is obtained by combining the results, thereby improving data processing efficiency and association accuracy, that is, faster, more complete and accurate sensing of targets, and improving the sensing effect.

[0156] Embodiments of the present invention also provide a computer-readable storage medium storing a computer program, wherein the computer program is configured to perform the steps in any of the above method embodiments when executed.

[0157] In some exemplary embodiments, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0158] Embodiments of the present invention also provide an electronic device including a memory and a processor, the memory storing a computer program and the processor being configured to run the computer program to perform the steps in any of the above method embodiments.

[0159] In some exemplary embodiments, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor and the input / output device is connected to the processor.

[0160] Specific examples in this embodiment can be found in the examples described in the above embodiments and exemplary implementations, and will not be repeated here.

[0161] It is obvious to those skilled in the art that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. They can be implemented using computer-executable program code, and thus can be stored in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those described herein, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular combination of hardware and software.

[0162] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for processing sensory data, characterized in that, Applied to sensing network elements, including: The sensed data is mapped to multiple grid regions to obtain the grid-level data corresponding to the multiple grid regions; The correlation between multiple sensing targets corresponding to the sensing data is determined based on the grid-level data; If, based on the aforementioned correlation, it is determined that there are two sensing targets that are the same sensing target, the sensing data corresponding to the two sensing targets are integrated and processed.

2. The method according to claim 1, characterized in that, in, The sensing data includes at least one of the following: sensing device identifier, sensing target identifier, sensing data acquisition time point, and sensing target location coordinate information.

3. The method according to claim 1, characterized in that, Before mapping the sensed data to multiple grids, the method further includes: The sensing area is divided into multiple grid areas by spatial grid division; wherein, the sensing area is determined according to the control area corresponding to the sensing network element; or, it is determined according to the position boundary corresponding to the sensing data of all sensing devices collected by the sensing network element.

4. The method according to claim 3, characterized in that, The spatial grid division of the sensing area includes one of the following: The sensing area is divided into spatial grids according to a pre-set first distance interval; or... The sensing area is divided into spatial grids based on the differences in latitude and longitude and the differences in altitude; or, The sensing area is divided into multiple sub-sensing areas, and each of the multiple sub-sensing areas is spatially gridded according to a pre-set second distance interval; or, The perception area is divided into spatial grids using machine learning.

5. The method according to claim 3, characterized in that, The method of dividing the sensing area into spatial grids further includes: Based on the mapping of the new sensed data across the multiple raster regions, determine the data volume and data density of the new raster-level data: In response to the situation where the amount of new raster-level data exceeds a first preset threshold and / or the data density of the new raster-level data exceeds a second preset threshold, the sensing area is re-divided into spatial raster sections.

6. The method according to claim 1, characterized in that, Determining the association between multiple sensing targets corresponding to the sensing data based on the raster-level data includes: For any given grid area, based on the grid-level data within that grid area, sensing target matching is performed on each pair of sensing targets corresponding to that grid area. Each pair of sensing targets includes two sensing targets, and the two sensing targets correspond to different sensing devices. The sensing target matching is used to determine whether the two sensing targets in each pair of sensing targets are the same sensing target. Based on the result of the target matching for each grid region, the association between the multiple target sensing objects is determined; wherein the target matching result is one of the following: exact matching result, no matching result, or fuzzy matching result.

7. The method according to claim 6, characterized in that, The step of performing perceptual target matching for each pair of perceptual targets corresponding to the grid area includes: If the raster-level data of the raster area contains only the sensing data of one sensing device, the sensing target matching is stopped. When the grid-level data corresponding to the grid area contains sensing data from multiple sensing devices, the sensing target matching is performed on the sensing target pair corresponding to any two sensing devices based on the sensing data from any two sensing devices in the grid area.

8. The method according to claim 7, characterized in that, When each of the plurality of sensing devices senses only one sensing target, the step of matching the sensing target pairs corresponding to any two sensing devices based on the sensing data of any two sensing devices in the grid area includes: Based on the sensing data of any two sensing devices, determine the sensing time difference and sensing distance difference for each pair of sensing targets corresponding to the sensing device that only senses one sensing target. For any pair of sensing targets corresponding to the sensing device that only senses one sensing target, if the sensing time difference corresponding to the pair of sensing targets is less than or equal to a third preset threshold and the sensing distance difference is less than or equal to a fourth preset threshold, the result of the sensing target matching is determined to be a precise matching result; wherein, the precise matching result is used to indicate that the two sensing targets in the pair of sensing targets are the same sensing target; If the perception time difference corresponding to the perception target pair is greater than a third preset threshold, and / or the perception distance difference is greater than a fourth preset threshold, the result of the perception target matching is determined to be a mismatch result; wherein, the mismatch result is used to indicate that the two perception targets in the perception target pair are different perception targets.

9. The method according to claim 7, characterized in that, When at least one of the two sensing devices can sense multiple sensing targets, the step of matching the sensing targets corresponding to the two sensing devices based on the sensing data of the two sensing devices in the grid area includes: Based on the sensing data of any two sensing devices, for any sensing target corresponding to any two sensing devices, determine the sensing time difference and sensing distance difference of all sensing target pairs corresponding to that sensing target; If, among all the perceived target pairs, there exists only one pair where the perception time difference is less than or equal to a third preset threshold and the perception distance difference is less than or equal to a fourth preset threshold, the result of the perceived target matching is determined to be a precise matching result; wherein, the precise matching result is used to indicate that the two perceived targets in the corresponding perceived target pair are the same perceived target; if, among all the perceived target pairs, there exist multiple pairs where the perception time difference is less than or equal to a third preset threshold and the perception distance difference is less than or equal to a fourth preset threshold, the result of the perceived target matching is determined to be a fuzzy matching result; wherein, the fuzzy matching result is used to indicate that it is impossible to determine that the two perceived targets in the multiple pairs of perceived target pairs are the same perceived target. If the perception time difference of all the perceived target pairs is greater than a third preset threshold, and / or the perception distance difference of all the perceived target pairs is greater than a fourth preset threshold, the result of the perceived target matching is determined to be a mismatch result; wherein, the mismatch result is used to indicate that all the perceived targets in the multiple sets of perceived target pairs are different perceived targets.

10. The method according to claim 6, characterized in that, The step of determining the association relationship between the multiple sensing targets based on the sensing target matching result of each grid region includes: Based on the matching results of each sensing target pair in the sensing area in different grid areas and the preset weight value, the matching degree value of each sensing target pair is determined, wherein the preset weight value corresponds to the matching results of the sensing targets in different grid areas; The uniqueness of each pair of perception targets is verified based on the matching degree value, and the two perception targets corresponding to the pair of perception targets that pass the uniqueness verification are determined to be the same perception target.

11. The method according to claim 10, characterized in that, Based on the correlation between the multiple sensing targets, the sensing data is integrated and processed, including: For any pair of sensing targets, if the two sensing targets corresponding to the pair are the same sensing target, the sensing data in the overlapping grid area corresponding to the pair of sensing targets is integrated; wherein, the overlapping grid area is the grid area that both sensing targets in the pair of sensing targets have reached. The integrated sensing data will be used as the final movement trajectory of the sensing target.

12. The method according to claim 6, characterized in that, After determining the correlation between multiple sensing targets corresponding to the sensing data, the method further includes: If, based on the correlation between the multiple sensing targets, it is determined that there is a sensing target pair with an abnormal sensing target match, a control message is sent to the corresponding sensing device, wherein the control message is used to instruct the sensing device to adjust the sensing configuration information.

13. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, wherein the computer program, when executed by a processor, implements the steps of the method described in any one of claims 1 to 12.

14. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method described in any one of claims 1 to 12.

15. A computer program product comprising a computer program that, when executed by a processor, implements the steps of the method described in any one of claims 1 to 12.

Images