A device relocation method, electronic device and computer readable storage medium

Abstract

The application discloses a device relocation method, an electronic device and a computer readable storage medium. The method comprises the following steps: obtaining first sensor data collected by a sensor of a target device in a target place, and obtaining a candidate pose set of the target device based on map data of the target place. The candidate pose set comprises a plurality of candidate pose points. Second sensor data of the sensor at the candidate pose points is simulated and generated based on the map data and device parameters of the sensor. A target pose point of the target device is selected from the candidate pose points based on matching results between the first sensor data and the second sensor data at the candidate pose points. In this way, the efficiency and accuracy of device relocation can be improved.

Description

Technical Field

[0001] This application relates to the field of device relocation technology, and in particular to a device relocation method, electronic device, and computer-readable storage medium. Background Technology

[0002] Existing devices typically require prior map data and up-to-date sensor data before relocalization. Taking LiDAR sensors as an example, the prior map data is usually a 2D or 3D grid map, and the sensor data is usually point cloud data. During relocalization, the pose is obtained by calculating the overlap rate between the point cloud data and the grid map boundaries. However, this relocalization method has low efficiency and accuracy. Summary of the Invention

[0003] The main technical problem addressed by this application is to provide a device repositioning method, electronic device, and computer-readable storage medium that can improve the efficiency and accuracy of device repositioning.

[0004] To solve the above-mentioned technical problems, one technical solution adopted in this application is: to provide a device repositioning method, comprising: First sensor data collected by sensors on the target device in the target location is acquired, and a candidate pose set of the target device is obtained based on the map data of the target location; wherein, the candidate pose set includes several candidate pose points; Based on map data and sensor device parameters, second sensor data at candidate pose points is simulated and generated. Based on the matching results between the first sensor data and the second sensor data at each candidate pose point, the candidate pose points are selected as the target pose points of the target device.

[0005] To solve the above-mentioned technical problems, another technical solution adopted in this application is: to provide an electronic device, including a processor and a memory coupled to each other, wherein the memory stores program instructions, and the processor is used to execute the program instructions to implement the above-mentioned device relocation method.

[0006] To solve the above-mentioned technical problems, another technical solution adopted in this application is: to provide a computer-readable storage medium for storing program instructions that can be executed to implement the above-mentioned device relocation method. The above scheme simulates and generates second sensor data at candidate pose points based on map data and sensor device parameters. It then matches the first and second sensor data. Since both data represent sensor data under the same device parameters and share similar data characteristics, they can be directly matched to determine the pose point. Compared to matching sensor data with map data, this eliminates the need to transform the sensor data from the sensor coordinate system to the map coordinate system. This reduces the additional computational load caused by coordinate transformation during relocalization, improving efficiency, and minimizes data accuracy loss due to coordinate transformation, thus enhancing accuracy. Therefore, it improves both the efficiency and accuracy of device relocalization. Attached Figure Description

[0007] Figure 1 This is a flowchart illustrating one embodiment of the device repositioning method provided in this application; Figure 2 yes Figure 1 A schematic diagram of the structure of an embodiment of step S11; Figure 3 yes Figure 1 A schematic diagram of the structure of an embodiment of step S12; Figure 4 yes Figure 1 A schematic diagram of the data processing process in step S12 of the embodiment; Figure 5 yes Figure 1 A flowchart illustrating one embodiment of step S12; Figure 6 yes Figure 1 A flowchart illustrating another embodiment of step S12; Figure 7 yes Figure 1 A flowchart illustrating one embodiment of step S13; Figure 8 yes Figure 1 A flowchart illustrating another embodiment of step S13; Figure 9 This is a flowchart illustrating another embodiment of the device repositioning method provided in this application; Figure 10 This is a flowchart illustrating another embodiment of the device repositioning method provided in this application; Figure 11 This is a schematic diagram of one embodiment of the electronic device provided in this application; Figure 12 This is a schematic diagram of the structure of the computer-readable storage medium provided in this application. Detailed Implementation

[0008] To make the purpose, technical solution and effects of this application clearer and more explicit, the following describes this application in further detail with reference to the accompanying drawings and embodiments.

[0009] It should be noted that if the embodiments of this application involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, features defined with "first" or "second" may explicitly or implicitly include at least one of those features. Descriptions involving "and / or," such as A and / or B, indicate that A, B, or both A and B are possible. Furthermore, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed in this application.

[0010] Please see Figure 1 , Figure 1 This is a flowchart illustrating one embodiment of the device relocation method provided in this application. It should be noted that if substantially the same result is achieved, this embodiment does not necessarily reflect that outcome. Figure 1 The illustrated process sequence is limited. For example... Figure 1 As shown, this embodiment includes: Step S11: Obtain the first sensor data collected by the sensor on the target device in the target location, and obtain the candidate pose set of the target device based on the map data of the target location, wherein the candidate pose set includes several candidate pose points.

[0011] The sensor can be a vision sensor, a lidar sensor, a millimeter-wave radar sensor, or other sensors.

[0012] The target device can be an automated terminal device, including but not limited to: a robot vacuum cleaner, an autonomous vehicle, a venue service robot, or other automated mobile devices.

[0013] The target location is the activity area of ​​the target equipment, including the movable range of the target equipment and the immovable range of the target equipment. The movable range of the target equipment can be the area in the target location where there is no target object, or it can be a range further defined based on the area in the target location where there is no target object. The immovable range can be the area in the target location where the target object is located, or it can be a set immovable range.

[0014] The target object is an object in the target location, specifically including but not limited to: buildings, appliances, furniture or other objects.

[0015] Map data includes location information and object information, specifically including but not limited to: vector map data, raster map data, or other electronic map data.

[0016] Based on map data of the target location, the target object and its location can be identified, thereby determining the movable range of the target device. Several candidate pose points within the movable range of the target device can then be selected to facilitate subsequent determination of the target device's location. These pose points include location information.

[0017] In one implementation, please refer to Figure 2 , Figure 2 yes Figure 1 A schematic diagram of the structure of step S11-one embodiment is shown. The bold black frame represents the surface of the target object. The gray area enclosed by the bold black frame is the movable range of the target device. The white area outside the bold black frame is the immovable range of the target device. The red grid lines include multiple horizontal and vertical lines. The intersection of a horizontal line and a vertical line is taken as a pose point. The distance between two adjacent horizontal lines is the same as the distance between two adjacent vertical lines, which can be determined according to the accuracy requirements of the pose points for device repositioning. The intersection points of the red grid lines within the movable range of the target device are selected as candidate pose points, and several candidate pose points form a candidate pose set.

[0018] Step S12: Based on map data and sensor device parameters, simulate and generate second sensor data at candidate pose points.

[0019] The device parameters of a sensor include the operating characteristics of the sensor in receiving and / or emitting signals, such as radar sensor parameters, including but not limited to: angular resolution, field of view, frame rate, point frequency, reflectivity or other parameters, and visual sensor parameters, including but not limited to: sensor size, total number of pixels, number of effective pixels, resolution or other parameters.

[0020] Based on the device parameters, sensor data of the target device in different activity locations can be simulated and generated.

[0021] In one implementation, please refer to Figure 3 , Figure 3 yes Figure 1 The schematic diagram of an embodiment of step S12 is shown. Taking a radar sensor as an example, the small red dot is the candidate pose point of the simulated target device, a green arrow line indicates the signal emitted from the sensor to the surface of the target object, and the black box is the target object. According to the angular resolution of the sensor or other device parameters, multiple signals are received and / or emitted in all directions around the candidate pose point to obtain simulated sensor data.

[0022] Step S13: Based on the matching results between the first sensor data and the second sensor data at each candidate pose point, select the candidate pose point as the target pose point of the target device.

[0023] The data from the first sensor can be matched one by one with the data from the second sensor at each candidate pose point, and the matching results can be statistically analyzed. Alternatively, mathematical calculations can be performed on the data from the first sensor and the data from the second sensor. The calculation results from the first sensor can then be matched with the calculation results from the second sensor at each candidate pose point to obtain the matching results.

[0024] Matching results include, but are not limited to: ratios, differences, growth rates, or other methods.

[0025] Candidate pose points are selected as target pose points for the target device. The selection methods include, but are not limited to: matching results meeting the matching threshold, selecting the first few matching results in order, or other methods.

[0026] In this embodiment, second sensor data at candidate pose points is simulated by using map data and sensor device parameters. The first sensor data and the second sensor data are then matched to ensure that the first sensor data and the second sensor data are sensor data under the same device parameters and have the same data characteristics. No feature extraction is required for intermediate conversion, thereby improving the efficiency and accuracy of device repositioning.

[0027] In one embodiment, to improve the accuracy of the second sensor data and thus enhance the efficiency and accuracy of device repositioning, step S12—the process of simulating the generation of second sensor data at candidate pose points based on map data and sensor device parameters—is further optimized. Please refer to [link to relevant documentation]. Figure 4 , Figure 4 yes Figure 1 The schematic diagram of the data processing process in step S12 is described in one embodiment.

[0028] Based on the sensor's working principle 41, sensor data under arbitrary poses is simulated and generated, i.e., the possible data of the sensor under no constraints, thus obtaining arbitrary pose simulation data 42. The sensor's working principle 41 corresponds to the sensor's device parameters.

[0029] Based on candidate poses 44, map data 43, and arbitrary pose simulation data 42, the simulated sensor, on the candidate poses 44, combines the map data 43 to filter and correct the arbitrary pose simulation data 42, obtaining corrected simulation data 45. The specific filtering and correction methods are related to the sensor device parameters and also to the map parameters. Map parameters include accuracy and map data format.

[0030] Based on the corrected simulation data 45, candidate pose simulation data 46 corresponding to candidate pose 43 is obtained. Candidate pose simulation data 46 is used as the second sensor data at the candidate pose point.

[0031] In this implementation, sensor data under arbitrary poses is first simulated and generated. Then, the simulated data under arbitrary poses is filtered and corrected based on candidate poses and map data. The corrected simulated data is used as the second sensor data at the candidate poses, which improves the accuracy of the second sensor data and thus improves the efficiency and accuracy of device repositioning.

[0032] In one embodiment, to improve the accuracy of the second sensor data and thus enhance the efficiency and accuracy of device repositioning, step S12—the process of simulating the generation of second sensor data at candidate pose points based on map data and sensor device parameters—is further optimized. Please refer to [link to relevant documentation]. Figure 5 , Figure 5 yes Figure 1 A flowchart illustrating one embodiment of step S12 is provided. Specifically, step S12 includes: Step S51: Based on the device parameters, simulate and generate third sensor data; wherein, the third sensor data is the data collected by the simulated sensor in an area without objects.

[0033] The object-free area is a preset range centered on the location of the sensor, and there are no objects within the preset range. The preset range is determined based on the maximum range of data emitted and / or returned by the sensor.

[0034] Step S52: Correct the third sensor data based on the candidate pose points and map data to obtain the second sensor data.

[0035] The correction method can be either a lookup table approach: construct a correction table, which is stored in key-value pairs. The key can be extracted and combined from information such as sensor parameters, map type, distance between candidate pose points and the target object, and type of the target object. The value can be a correction value, which is used to correct the third sensor data to obtain the second sensor data. Alternatively, a mathematical model can be used: construct and train a mathematical model, using candidate pose points, map data, and third sensor data as inputs. The output data of the mathematical model is used as the second sensor data.

[0036] In this embodiment, data collected by the sensor in an object-free area is first simulated based on the device parameters to generate third sensor data. Then, the third sensor data is corrected based on candidate pose points and map data to obtain second sensor data, which improves the accuracy of the second sensor data and thus improves the efficiency and accuracy of device repositioning.

[0037] In one embodiment, improving the accuracy of the third sensor data correction is considered, thereby improving the accuracy of the second sensor data, and ultimately improving the efficiency and accuracy of device repositioning. Further optimization is made to step S12: the process of simulating and generating second sensor data at candidate pose points based on map data and sensor device parameters. Please refer to [link to relevant documentation]. Figure 6 , Figure 6 yes Figure 1 A flowchart illustrating another embodiment of step S12 is provided. Specifically, step S12 includes: Step S61: Obtain the sample sensor data collected by the sensor on the sample device at the sample pose point in the sample location, and generate simulated sensor data based on the device parameters. The simulated sensor data is the data collected by the simulated sensor in the area without objects.

[0038] The sample location can be the target location or other locations. It can be a real location where the target equipment is active, or a location where the target equipment is active in an experimental environment.

[0039] The method for acquiring the sample pose points can be the same as or different from the method for acquiring the target pose points. The sample device can be the same as or different from the target device. The sensors at the sample location can include multiple types of sensors, but must include at least one type, whose device parameters are the same as those of the sensors at the target location.

[0040] Step S62: Using simulated sensor data, sample pose points, and sample map data of the sample location as model input data, and sample sensor data as expected output data, supervised training is performed on the data correction model until training converges.

[0041] The data correction model can employ existing mathematical models, specifically including at least one: machine learning model, deep learning model, neural network model, large model, or other models. The data correction model is trained using a model training method corresponding to the mathematical model.

[0042] Step S63: Based on the device parameters, simulate and generate third sensor data; wherein, the third sensor data is the data collected by the simulated sensor in an area without objects.

[0043] Step S64: Correct the third sensor data based on the data correction model reference candidate pose points and map data to obtain the second sensor data.

[0044] The third sensor data, candidate pose points, and map data are input into the data correction model, and the data output by the data correction model is used as the second sensor data.

[0045] In this embodiment, a mathematical correction model is generated by training sample data. The mathematical correction model is then used to correct the third sensor data based on the data correction model, the candidate pose points, and the map data. This greatly improves the accuracy of the third sensor data correction, thereby improving the accuracy of the second sensor data and ultimately enhancing the efficiency and accuracy of device repositioning.

[0046] In one embodiment, mapping relationships are considered to improve the accuracy of third sensor data correction, thereby improving the accuracy of second sensor data, and ultimately improving the efficiency and accuracy of device repositioning. Step S12, which simulates the generation of second sensor data at candidate pose points based on map data and sensor device parameters, is further optimized. Taking an experimental site as an example, step S12 includes: Extract the influencing factors of sensor data, calculate the deviation between real sensor data and simulated sensor data, and construct a mapping relationship between influencing factors and deviation values.

[0047] Based on the device parameters, third sensor data is generated through simulation. This third sensor data consists of data collected by the simulated sensor in an area without objects.

[0048] Based on candidate pose points and map data, the influencing factors of the current sensor data are extracted, the current deviation value is obtained according to the mapping relationship, the third sensor data is corrected, and the second sensor data is obtained.

[0049] Taking a radar sensor as an example, the sensor emits a signal, which is then transmitted to the surface of an object and returned. Ignoring environmental influences such as air pollution, the factors affecting the sensor data are extracted: the relative position of the experimental pose point and the experimental object, and the influence of the object's surface on the signal. The object's surface can be represented by map data corresponding to the object's location. Through extensive experimental data, the relationship between these influencing factors and the sensor data can be simulated.

[0050] A pose point is selected at the experimental site as the experimental pose point. An experimental area is formed by a preset range centered on the experimental pose point. This experimental area is divided into multiple sub-regions, and a set of experimental objects are placed in different sub-regions, forming multiple sets of experimental regions and corresponding experimental maps. Correspondingly, experimental sensor data collected by sensors on the experimental equipment located at the experimental pose point within the experimental regions is acquired, and experimental simulation sensor data is generated based on the equipment parameters.

[0051] The deviation between experimental sensor data and experimental simulated sensor data, the relative position information of experimental pose points and experimental objects, and the experimental object information are calculated. The relative position information and experimental object information are extracted based on the experimental map data. A mapping relationship between the domain and the range is established, using the relative position information and experimental object information as the domain and the deviation value as the range. This is achieved through mathematical model fitting and training based on multiple sets of experimental data.

[0052] In this embodiment, by extracting influencing factors and constructing mapping relationships, the third sensor data is corrected based on the mapping relationships, referencing candidate pose points and map data. This greatly improves the accuracy of the third sensor data correction, thereby improving the accuracy of the second sensor data and ultimately enhancing the efficiency and accuracy of device repositioning.

[0053] In one embodiment, improving matching efficiency is considered to enhance the efficiency and accuracy of device repositioning. Step S13, which involves selecting candidate pose points as target pose points for the target device based on the matching results between the first sensor data and the second sensor data at each candidate pose point, is further optimized. Sensor data includes sub-data formed by several sensor signals, and each sub-data involves several data attributes. Taking a radar sensor as an example, the sensor emits n signals around its location and receives m signals. Each signal includes angle data, distance data, and other data, with data attributes including angle and distance. Here, n and m are positive integers, and m is not greater than n.

[0054] Please see Figure 7 , Figure 7 yes Figure 1 A flowchart illustrating one embodiment of step S13 is provided. Specifically, step S13 includes: Step S71: Based on the statistical values ​​of the data attributes in each sub-data in the first sensor data, obtain the first sensor index, and based on the statistical values ​​of the data attributes in each sub-data in the second sensor data, obtain the second sensor index. The first sensor index is obtained by fusing the statistical values ​​of each data attribute in the first sensor data, and the second sensor index is obtained by fusing the statistical values ​​of each data attribute in the second sensor data.

[0055] Statistical values ​​can be statistical indicators related to the distribution of sensor data. Statistical indicators include mean, standard deviation, etc. The specific statistical indicators can be selected by preset or adjusted based on the sensor parameters to reference relocation efficiency and accuracy.

[0056] To integrate statistical values, mathematical methods for quantifying indicators can be used, including but not limited to: linear transformation, normalization, weighted summation, weighted average, and other methods.

[0057] Step S72: Match the first sensor index with the second sensor index corresponding to the candidate pose point to determine whether to select the candidate pose point as the target pose point.

[0058] The first sensor index is compared with the second sensor index corresponding to each candidate pose point. The comparison result is then used to determine if it meets a matching threshold. Candidate pose points that meet the matching threshold are selected as the target pose points. The matching threshold can be preset or adjusted based on sensor parameter references to improve relocalization efficiency and accuracy.

[0059] In this embodiment, matching is performed by calculating the first sensor index and the second sensor index. Compared with matching directly using sensor data features, this reduces the amount of matching data, increases the matching speed, and thus improves the accuracy of device repositioning.

[0060] In one embodiment, refer to the above. Figure 7 The corresponding implementation further optimizes the calculation of the first sensor index and the second sensor index in step S71, specifically including: The first sensor index is obtained by weighted summation of the statistical values ​​of each data attribute in the first sensor data; and / or, the second sensor index is obtained by weighted summation of the statistical values ​​of each data attribute in the second sensor data. For example, the sensor index calculation formula is:

[0061] Where R represents the sensor index, Indicates the first i The influence coefficient of each statistical value Indicates the first i One statistical value, S This indicates the number of statistical values.

[0062] The contribution of different statistical indicators to the sensor index can be controlled by the influence coefficient. For example, the matching process and results are highly sensitive to the standard deviation of distance, so the coefficient of the standard deviation index can be increased.

[0063] In this embodiment, matching is performed by calculating the first sensor index and the second sensor index. Compared with direct matching using sensor data features, this reduces the amount of matching data and thus improves the matching speed. The influence coefficient is used to control the degree of influence of different statistical values ​​on the sensor index, which more accurately reflects the sensitivity of repositioning to different statistical values, thereby improving the accuracy of device repositioning.

[0064] In one embodiment, further improvements in matching efficiency are considered, thereby enhancing the efficiency and accuracy of device repositioning. Step S13, the process of selecting candidate pose points as target pose points for the target device based on the matching results between the first sensor data and the second sensor data at each candidate pose point, is further optimized. Please refer to [link to relevant documentation]. Figure 8 , Figure 8 yes Figure 1A flowchart illustrating another embodiment of step S13 is provided. Specifically, step S13 includes: Step S81: Based on the statistical values ​​of the data attributes in each sub-data in the first sensor data, a first sensor index is obtained, and based on the statistical values ​​of the data attributes in each sub-data in the second sensor data, a second sensor index is obtained. The first sensor index is obtained by fusing the statistical values ​​of each data attribute in the first sensor data, and the second sensor index is obtained by fusing the statistical values ​​of each data attribute in the second sensor data.

[0065] Step S82: Based on the comparison results between the statistical values ​​of each data attribute in the second sensor data corresponding to the two candidate pose points in the candidate pose set, determine whether to merge the two candidate pose points in the candidate pose set into a new candidate pose point, wherein the two candidate pose points in the candidate pose set are adjacent candidate pose points.

[0066] From the candidate pose set, two adjacent candidate pose points are sequentially selected as a group to be merged. Alternatively, a merging range can be set, and multiple candidate pose points within the merging range can be grouped together. The candidate pose points within the group to be merged are compared with a merging threshold to determine whether to merge the candidate pose points into a new candidate pose point.

[0067] For example, if 100 candidate pose points are arranged in a matrix, with 10 in the horizontal direction and 10 in the vertical direction, and the merging range is 2 pose points, then the two adjacent candidate poses in the horizontal direction are taken as a group to be merged. If the merging range is 4 pose points, then the two adjacent candidate poses in the horizontal direction and the corresponding two adjacent candidate poses in the vertical direction are taken as a group to be merged.

[0068] The statistical values ​​of the data attributes of the second sensor data from two adjacent pose points in the group to be merged are compared one by one to determine whether they meet the merging threshold. The merging threshold can be a set of thresholds related to the data attributes, or it can be a single threshold. The comparison can be the difference, ratio, or other methods between two statistical values.

[0069] When the merging threshold is met, the center of the candidate pose points in the group to be merged can be calculated as a new candidate pose point, or the average value of the positions of the candidate pose points in the group to be merged can be calculated to obtain a new candidate pose point.

[0070] Step S83: Calculate the mean value based on the statistical values ​​of each data attribute in the second sensor data to obtain the second sensor data index of the new candidate pose point.

[0071] When the merging threshold is met, new candidate pose points and corresponding second sensor indices are calculated, and the new candidate pose points replace the candidate pose points in the group to be merged.

[0072] If the merging threshold is not met, retain the candidate pose points in the group to be merged.

[0073] Step S84: Match the first sensor index with the second sensor index corresponding to the candidate pose point to determine whether to select the candidate pose point as the target pose point.

[0074] In this embodiment, adjacent candidate pose points with similar statistical values ​​are identified and merged to reduce the number of candidate pose points, reduce redundant calculations in the matching process, improve matching efficiency, and thus improve the efficiency and accuracy of device repositioning.

[0075] In one implementation, improving matching efficiency is considered, thereby improving the efficiency and accuracy of device repositioning. Please refer to [link / reference]. Figure 9 , Figure 9 This is a flowchart illustrating another embodiment of the device relocation method provided in this application, specifically including: Step S91: Obtain the first sensor data collected by the sensor on the target device in the target location, and obtain the candidate pose set of the target device based on the map data of the target location. The candidate pose set includes several candidate pose points, which are obtained by sampling the map data with the target step size.

[0076] Step S92: Based on map data and sensor device parameters, simulate and generate second sensor data at candidate pose points.

[0077] Step S93: Based on the matching results between the first sensor data and the second sensor data at each candidate pose point, select the candidate pose point as the target pose point of the target device.

[0078] When the target device shifts or after a period of time, a repositioning process may need to be performed again. Previous repositioning processes (or the first few) serve as historical repositioning processes. The efficiency and accuracy of these historical repositioning processes are monitored, and the target step size is adjusted based on the monitoring results. A smaller step size results in more candidate pose points and a slower matching speed; a larger step size results in fewer candidate pose points and a faster matching speed. Monitoring results can include: the number of target pose points, the number of candidate pose points, matching speed, and matching accuracy. Target step size adjustment rules can be set, and the target step size is adjusted based on these rules and the monitoring results. Specifically, this includes: Step S94: In response to the displacement of the target device, obtain the first number of target pose points during historical repositioning and the second number of several candidate pose points in the candidate pose set.

[0079] Step S95: Based on the first and second quantities, adjust the target step size and return to step S91 to initiate this relocation. Further, this includes: In response to a first quantity satisfying a first condition and a second quantity satisfying a second condition, the target step size is increased; wherein the first condition includes: the first quantity is greater than a first threshold, and the second condition includes: the second quantity is greater than a second threshold.

[0080] For example, if the first threshold is 10 and the second threshold is 10000, when the number of target pose points is greater than 10 and the number of candidate pose points is greater than 10000, it indicates that the target step size is too small, resulting in a large number of target pose points and candidate pose points. Therefore, the target step size should be increased. The data in this example is for illustrative purposes only and is not intended to limit specific thresholds.

[0081] In response to the first quantity satisfying the third condition and the second quantity satisfying the fourth condition, the target step size is reduced; wherein the third condition includes: the first quantity is less than the third threshold, and the fourth condition includes: the second quantity is less than the fourth threshold, and the first threshold is greater than the third threshold and the second threshold is greater than the fourth threshold.

[0082] For example, if the third threshold is 5 and the fourth threshold is 5000, when the number of target pose points is less than 5 and the number of candidate pose points is less than 5000, it means that the target step size is too large and the number of target pose points and candidate pose points obtained is too small. Therefore, the target step size should be reduced.

[0083] In one embodiment, where there is a historical relocation process, the current relocation steps include: The system acquires the first sensor data collected by the sensors on the target device in the target location, and identifies whether a historical candidate pose set is stored. If yes, the historical candidate pose set is used as the candidate pose set for the target device; otherwise, the candidate pose set for the target device is obtained based on the map data of the target location.

[0084] If historical second sensor data corresponding to the historical candidate pose set is stored, then the historical second sensor data is used as the second sensor data at the current candidate pose point; otherwise, the second sensor data at the candidate pose point is simulated based on the map data and the sensor's device parameters.

[0085] Based on the matching results between the first sensor data and the second sensor data at each candidate pose point, the candidate pose points are selected as the target pose points of the target device.

[0086] In this embodiment, by monitoring historical relocation data and adjusting the target step size based on the monitoring results, matching efficiency can be improved, thereby improving the efficiency and accuracy of device relocation.

[0087] In one implementation, improving matching accuracy is considered, thereby improving the efficiency and accuracy of device repositioning. Please refer to... Figure 10 , Figure 10 This is a flowchart illustrating another embodiment of the device relocation method provided in this application, specifically including: Step S101: Obtain the first sensor data collected by the sensor on the target device in the target location, and obtain the candidate pose set of the target device based on the map data of the target location, wherein the candidate pose set includes several candidate pose points.

[0088] Step S102: Based on map data and sensor device parameters, simulate and generate second sensor data at candidate pose points.

[0089] Step S103: Based on the matching results between the first sensor data and the second sensor data at each candidate pose point, select the candidate pose point as the target pose point of the target device.

[0090] In complex scenarios, multiple target pose points may be selected based on the matching results, making it difficult to determine the accurate pose point. Further selection of the target pose points specifically includes: Step S104: Based on the target pose point, obtain a local map containing the target pose point from the map data.

[0091] Step S105: Based on the data from the first sensor, match it with the local map to determine the first pose point in the local map as the final target pose point.

[0092] The first sensor data and local map matching adopts a general sensor data and map matching process, and there are no restrictions on the sensor data and map matching process here.

[0093] In this embodiment, unlike the existing technology that directly uses sensor data to match the map, the target pose point is selected by matching the first sensor data and the simulated second sensor data, and a local map is obtained based on the target pose point. This narrows the search range for matching sensor data with the map and improves the efficiency and accuracy of device repositioning.

[0094] In one embodiment, improving matching accuracy is considered to improve the efficiency and accuracy of device repositioning. The device repositioning method provided in this application, after selecting candidate pose points as target pose points for the target device based on the matching results between first sensor data and second sensor data at each candidate pose point, and treating this repositioning process as a historical repositioning process, further includes: Based on the historical relocation, the first sensor data and map data are matched to determine the second pose point in the map data, which is used as the reference target pose point. A general sensor data and map matching process is adopted, and no limitation is made to the sensor data and map matching process here.

[0095] The pose point mapping relationship is obtained by fitting the reference target pose point and the target pose point determined by the first sensor data during historical relocation.

[0096] Based on the pose point mapping relationship, the target pose points determined in this relocalization are mapped to obtain the final target pose points for this relocalization.

[0097] For example, after acquiring four target pose points through historical relocalization, a reference target pose point is obtained using a common sensor data and map matching process. The relative positional relationship between the four target pose points and the reference target pose point is recorded as a pose point mapping relationship. After acquiring four target pose points during the current relocalization, the target pose points determined in this relocalization are mapped based on the pose point mapping relationship to obtain the final target pose points for this relocalization.

[0098] This implementation differs from the existing technology that directly uses sensor data to match the map. Instead, it obtains the pose point mapping relationship through one or fewer sensor data to map matching processes. The pose point mapping relationship is then used to determine the final target pose point of multiple target pose points, reducing the number of sensor data to map matching processes and improving the efficiency and accuracy of device repositioning.

[0099] Please see Figure 11 , Figure 11 This is a schematic diagram of an embodiment of the electronic device provided in this application. In this embodiment, the electronic device 110 includes a processor 112 and a memory 111 that are coupled to each other.

[0100] The processor 112 can be a general-purpose processor, including but not limited to: CPU (Central Processing Unit), NP (Network Processor), etc.; it can also be DSP (Digital Signal Processor), ASIC (Application Specific Integrated Circuit), FPGA (Field-Programmable Gate Array) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0101] The memory 111 can be a non-volatile memory, such as ROM (Read-Only Memory), RAM (Random Access Memory), or other memory.

[0102] The memory 111 stores program instructions, and the processor 112 executes the program instructions to implement the method provided by any embodiment and any non-conflicting combination of the above-described device relocation method.

[0103] Please see Figure 12 , Figure 12 This is a schematic diagram of the structure of the computer-readable storage medium provided in this application. In this embodiment, the computer-readable storage medium 120 stores program instructions 121, which, when executed, implement the method provided by any embodiment of the device relocation method of this application and any non-conflicting combination thereof.

[0104] The program instructions 121 can be formed into a program file and stored in the aforementioned computer-readable storage medium 120 in the form of a software product, so that a computer device (which may be a personal computer, server, or network device, etc.) can execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned computer-readable storage medium 120 includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory, random access memory, magnetic disks, or optical disks, or terminal devices such as computers, servers, mobile phones, and tablets.

[0105] The above description is merely an embodiment of this application and does not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method for repositioning equipment, characterized in that, include: First sensor data collected by sensors on a target device in a target location is acquired, and a candidate pose set of the target device is obtained based on map data of the target location; wherein, the candidate pose set includes several candidate pose points; Based on the map data and the device parameters of the sensor, second sensor data of the sensor at the candidate pose point is simulated and generated. Based on the matching results between the first sensor data and the second sensor data at each of the candidate pose points, the candidate pose points are selected as the target pose points of the target device.

2. The method according to claim 1, characterized in that, The process of simulating and generating second sensor data at the candidate pose point based on the map data and the sensor's device parameters includes: Based on the device parameters, third sensor data is simulated and generated; wherein, the third sensor data is data collected by the sensor in a non-object area, the non-object area is a preset range centered on the location of the sensor, and there are no objects within the preset range; The third sensor data is corrected based on the candidate pose points and the map data to obtain the second sensor data.

3. The method according to claim 2, characterized in that, Before correcting the third sensor data based on the candidate pose points and the map data to obtain the second sensor data, the method further includes: The sample sensor data collected by the sensor on the sample device at the sample pose point in the sample location is acquired, and simulated sensor data is generated based on the device parameters; wherein, the simulated sensor data is data that simulates the sensor's collection in the objectless area; The simulated sensor data, the sample pose points, and the sample map data of the sample location are used as model input data, and the sample sensor data is used as the expected output data. The data correction model is trained in a supervised manner until the training converges. The step of correcting the third sensor data based on the candidate pose points and the map data to obtain the second sensor data includes: The third sensor data is corrected based on the data correction model, referencing the candidate pose points and the map data, to obtain the second sensor data.

4. The method according to claim 1, characterized in that, The sensor data includes sub-data formed by several sensor signals, and the sub-data involves several data attributes. The step of selecting the candidate pose points as the target pose points of the target device based on the matching results between the first sensor data and the second sensor data at each of the candidate pose points includes: A first sensor index is obtained based on the statistical values ​​of the data attributes in each of the sub-data in the first sensor data, and a second sensor index is obtained based on the statistical values ​​of the data attributes in each of the sub-data in the second sensor data; wherein, the first sensor index is obtained by fusing the statistical values ​​of each of the data attributes in the first sensor data, and the second sensor index is obtained by fusing the statistical values ​​of each of the data attributes in the second sensor data; Based on the matching of the first sensor index with the second sensor index corresponding to the candidate pose point, it is determined whether to select the candidate pose point as the target pose point.

5. The method according to claim 4, characterized in that, The first sensor index is obtained by weighted summation of the statistical values ​​of each data attribute in the first sensor data; And / or, the second sensor index is obtained by weighted summation of the statistical values ​​of each of the data attributes in the second sensor data.

6. The method according to claim 4, characterized in that, After obtaining the second sensor index based on the statistical values ​​of the data attributes in each of the sub-data in the second sensor data, and before determining whether to select the candidate pose point as the target pose point by matching the first sensor index with the second sensor index corresponding to the candidate pose point, the method further includes: Based on the comparison results between the statistical values ​​of each data attribute in the second sensor data corresponding to the two candidate pose points in the candidate pose set, it is determined whether to merge the two candidate pose points in the candidate pose set into a new candidate pose point; wherein, the two candidate pose points in the candidate pose set are adjacent candidate pose points; The mean value of each data attribute in the second sensor data is calculated to obtain the second sensor data index of the new candidate pose point.

7. The method according to claim 1, characterized in that, The candidate pose points within the candidate pose set are obtained by sampling the map data using a target step size. After selecting a candidate pose point as the target pose point of the target device based on the matching results between the first sensor data and the second sensor data at each candidate pose point, the method further includes: In response to the displacement of the target device, a first number of target pose points and a second number of candidate pose points in the candidate pose set are obtained during historical repositioning. Based on the first quantity and the second quantity, the target step size is adjusted, and the process returns to the step of acquiring the first sensor data collected by the sensor on the target device in the target location to initiate this repositioning.

8. The method according to claim 1, characterized in that, When the target pose point is determined, after selecting the candidate pose point as the target pose point of the target device based on the matching results between the first sensor data and the second sensor data at each of the candidate pose points, the method further includes: Based on the target pose point, obtain a local map containing the target pose point from the map data; Based on the matching of the first sensor data with the local map, the first pose point in the local map is determined as the final target pose point.

9. The method according to claim 1, characterized in that, When the target pose point is determined, after selecting the candidate pose point as the target pose point of the target device based on the matching results between the first sensor data and the second sensor data at each of the candidate pose points, the method further includes: Based on the historical relocation, the first sensor data is matched with the map data to determine the second pose point in the map data, which is used as the reference target pose point; The pose point mapping relationship is obtained by fitting the reference target pose point and the target pose point determined by the first sensor data during historical relocation. Based on the pose point mapping relationship, the target pose points determined in this relocalization are mapped to obtain the final target pose points for this relocalization.

10. An electronic device, characterized in that, It includes a processor and a memory coupled to each other, the memory storing program instructions, and the processor executing the program instructions to implement the device relocation method as described in any one of claims 1-9.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store program instructions that can be executed to implement the device relocation method as described in any one of claims 1-9.

Images