Edge path determination method, device, robot and storage medium

Through computer programs, a computer program processes obstacle point cloud data of multiple sensors is generated to generate accurate paths along the edge, solving the path inaccuracy problem caused by sensor errors and improving the effect of robots along the edge cleaning.

CN118425980BActive Publication Date: 2024-11-22SHENZHEN PUDU TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410897174.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-07-05
Publication Date
2024-11-22
Estimated Expiration
2044-07-05

AI Technical Summary

Technical Problem

During the cleaning process of traditional robots along the edge, the accuracy of the edge path is reduced due to the accumulated error of the sensor, which affects the cleaning effect.

Method used

By acquiring the obstacle point cloud of at least two sensors, the target observation income and cumulative observation income at the target moment are calculated. If the cumulative observation income is greater than the threshold, the obstacle point cloud is determined and a path along the edge is generated to improve the accuracy of the path.

Benefits of technology

Improves the accuracy of the edge path, ensures the cleaning effect of the robot in the edge cleaning mode, and reduces positioning jitter.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118425980B_ABST
    Figure CN118425980B_ABST
Patent Text Reader

Abstract

The present application relates to a method, device, robot and storage medium for determining a path along the edge. The method comprises: obtaining obstacle point clouds corresponding to at least two sensors at a target time; determining a target observation benefit corresponding to the target time based on the obstacle point clouds corresponding to at least two sensors; summing a plurality of historical observation benefits and a target observation benefit to obtain a cumulative observation benefit corresponding to the target time; the historical time corresponding to the historical observation benefit is before the target time; when the cumulative observation benefit is greater than an observation benefit threshold, determining the obstacle point clouds corresponding to at least two sensors as target obstacle point clouds; and determining a path along the edge corresponding to the target time based on the target obstacle point cloud. The use of this method can improve the accuracy of the path along the edge.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a method, device, robot and storage medium for determining an edge path. Background Art

[0002] With the development of artificial intelligence technology, robots with cleaning functions are widely used in life and production. Robots have multiple cleaning modes to meet different cleaning needs. For example, robots use edge cleaning mode to clean the edges of walls.

[0003] In traditional technology, the robot collects position information through sensors such as vision or lidar to clean along the edge of the wall. However, as the robot walks along the wall, the cumulative error of the sensor will become larger and larger, and the robot's edge cleaning function will be greatly adversely affected, reducing the accuracy of the robot's edge cleaning. Summary of the invention

[0004] Based on this, it is necessary to provide a method, device, robot and computer-readable storage medium for determining an edge path that can improve the accuracy of the edge path in order to address the above technical problems.

[0005] In a first aspect, the present application provides a method for determining an edge path. The method comprises:

[0006] Obtain obstacle point clouds corresponding to at least two sensors at target times;

[0007] Determining a target observation benefit corresponding to the target moment based on obstacle point clouds corresponding to at least two of the sensors respectively;

[0008] The cumulative observation benefit corresponding to the target moment is obtained by summing up multiple historical observation benefits and the target observation benefit; the historical moment corresponding to the historical observation benefit is before the target moment;

[0009] When the accumulated observation benefit is greater than the observation benefit threshold, determining the obstacle point clouds corresponding to at least two sensors as target obstacle point clouds;

[0010] Based on the target obstacle point cloud, determine the edge path corresponding to the target time.

[0011] In a second aspect, the present application also provides a device for determining an edge path. The device comprises:

[0012] An acquisition module, used to acquire obstacle point clouds corresponding to at least two sensors at target times;

[0013] A calculation module, used for determining the target observation benefit corresponding to the target moment based on the obstacle point clouds respectively corresponding to at least two of the sensors;

[0014] A summing module, used for summing a plurality of historical observation benefits and the target observation benefit to obtain a cumulative observation benefit corresponding to the target moment; the historical moment corresponding to the historical observation benefit is before the target moment;

[0015] A comparison module, configured to determine the obstacle point clouds corresponding to at least two sensors as target obstacle point clouds when the cumulative observation benefit is greater than an observation benefit threshold;

[0016] A determination module is used to determine the edge path corresponding to the target moment based on the target obstacle point cloud.

[0017] In a third aspect, the present application further provides a robot comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the methods described in the first aspect when executing the computer program.

[0018] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of any one of the methods described in the first aspect.

[0019] In a fifth aspect, the present application further provides a computer program product, comprising a computer program, which, when executed by a processor, implements the steps of any one of the methods described in the first aspect.

[0020] The above-mentioned edge path determination method, device, robot and storage medium determine the target observation gain corresponding to the target moment through the obstacle point clouds corresponding to at least two sensors respectively, sum up multiple historical observation gains and target observation gains, and obtain the cumulative observation gain corresponding to the target moment. The cumulative observation gain represents the total amount of information of the obstacle point clouds acquired by multiple sensors within a period of time including the target moment. If the cumulative observation gain is greater than the observation gain threshold, it means that the amount of information of the obstacle point clouds acquired by multiple sensors within the period of time including the target moment is large, that is, the accuracy of the obstacle point clouds acquired by multiple sensors is high. Then, the obstacle point clouds corresponding to at least two sensors respectively are determined as the target obstacle point clouds, and the edge path corresponding to the target moment is determined based on the target obstacle point cloud, thereby improving the accuracy of the edge path. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0022] Figure 1 A diagram of an application environment of a method for determining an edge path in an embodiment;

[0023] Figure 2 A schematic diagram of a flow chart of a method for determining an edge path in an embodiment;

[0024] Figure 3 A schematic diagram of a process for determining a target observation benefit in one embodiment;

[0025] Figure 4 A schematic diagram of a flow chart of a candidate observation benefit determination step in one embodiment;

[0026] Figure 5 A schematic diagram of a flow chart of a step of determining an edge path in an embodiment;

[0027] Figure 6 A structural block diagram of an edge path determination device according to an embodiment;

[0028] Figure 7 1 is a diagram showing the internal structure of a robot in one embodiment. DETAILED DESCRIPTION

[0029] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0030] The edge path determination method provided in the embodiment of the present application can be applied to Figure 1In the application environment shown. Among them, the terminal 102 communicates with the robot 104 through the network. The terminal 102 and the robot 104 can be used together to execute the edge path determination method provided in the embodiment of the present application, and the robot 104 can also be used alone to execute the edge path determination method provided in the embodiment of the present application. Among them, the robot 104 can be various self-moving devices that can perform edge tasks, such as: cleaning robots, delivery robots, guide robots, automatic guided vehicles (Automatic Guided Vehicle, AGV) and sweepers. The terminal 102 can be but is not limited to various personal computers, laptops, smart phones, tablets, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart TVs and smart car-mounted devices, etc., and the portable wearable devices can be smart watches, etc. Of course, the robot 104 can also cooperate with the server to implement the edge path determination method provided in the embodiment of the present application, wherein the server can be an independent physical server or a service node in the blockchain system, and each service node in the blockchain system forms a peer-to-peer (P2P) network, and the P2P protocol is an application layer protocol running on the Transmission Control Protocol (TCP) protocol. In addition, the server can also be a server cluster composed of multiple physical servers, which can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (Content Delivery Network, CDN), and big data and artificial intelligence platforms. The robot 104 can be connected to the server through Bluetooth, USB (Universal Serial Bus) or network communication connection methods, and this application does not limit this.

[0031] In one embodiment, Figure 2 As shown, a method for determining an edge path is provided. This embodiment is described by taking the method applied to a robot that cleans the floor as an example, and includes steps 202 to 210.

[0032] Step 202: Obtain obstacle point clouds corresponding to at least two sensors at target times.

[0033] Among them, the sensor refers to a device used to sense and collect environmental information. The sensor is a part of the robot. The sensor includes but is not limited to a lidar sensor, a depth camera, and an ultrasonic sensor. At least two sensors can be a lidar sensor and a depth camera, or a lidar sensor and an ultrasonic sensor, or a lidar sensor, a depth camera, and an ultrasonic sensor. The obstacle point cloud refers to a point set containing the obstacle position obtained by the sensor. The target time refers to the time corresponding to the obstacle point cloud, which can be understood as the time when the obstacle point cloud corresponding to at least two sensors is obtained.

[0034] Exemplarily, when the robot performs a cleaning task in the edge cleaning mode, it obtains an obstacle point cloud corresponding to a target time through at least two sensors. The edge cleaning mode is a cleaning mode of the robot. In the edge cleaning mode, the robot performs cleaning tasks along the edge of a room wall, the bottom edge of furniture or other obstacles.

[0035] Step 204: determining the target observation benefit corresponding to the target time based on the obstacle point clouds corresponding to at least two sensors.

[0036] Among them, the observation gain represents the amount of useful information obtained through observation. The larger the observation gain, the larger the amount of useful information obtained through observation. It can be understood as the amount of useful information of the obstacle point cloud obtained by the sensor. The target observation gain refers to the sum of the useful information of the obstacle point cloud corresponding to at least two sensors, which can be understood as the sum of the useful information corresponding to at least two obstacle point clouds.

[0037] Exemplarily, the robot determines the observation benefit corresponding to each obstacle point cloud, sums up multiple observation benefits, and obtains the target observation benefit corresponding to the target moment. It is understandable that multiple observation benefits can be directly summed, or the weight coefficient of the corresponding observation benefit can be determined according to the accuracy of the sensor, and then the multiple observation benefits are weighted and summed. The weight coefficient is proportional to the accuracy of the sensor.

[0038] Step 206, summing up multiple historical observation benefits and target observation benefits to obtain the cumulative observation benefit corresponding to the target moment; the historical moment corresponding to the historical observation benefit is before the target moment.

[0039] The historical observation benefit refers to the target observation benefit corresponding to the historical moment, and the historical moment refers to the moment before the target moment. The multiple historical observation benefits can be a preset number of historical observation benefits adjacent to the target moment, for example, the multiple historical observation benefits can be 9 historical observation benefits, or can also be 4 historical observation benefits.

[0040] Exemplarily, the robot obtains the historical observed benefits corresponding to a preset number of historical moments before the target moment, sums up the multiple historical observed benefits and the target observed benefits, and obtains the cumulative observed benefits corresponding to the target moment. It can be understood that the summation can be a direct summation or a weighted summation. A weighted summation can be performed based on the weight coefficients corresponding to the historical observed benefits and the target observed benefits to obtain the cumulative observed benefits. Among them, the weight coefficient can be determined based on the interval between the historical moment and the target moment. The longer the interval, the smaller the weight coefficient.

[0041] Step 208: When the accumulated observation benefit is greater than the observation benefit threshold, the obstacle point clouds corresponding to at least two sensors are determined as target obstacle point clouds.

[0042] Among them, the observation gain threshold refers to the minimum observation gain for determining the obstacle point cloud as the target obstacle point cloud. The observation gain threshold can be set according to actual needs. For example, the observation gain threshold can be determined according to the number of sensors and the measurement accuracy. The observation gain threshold can be proportional to the number of sensors and the measurement accuracy of the sensors. That is, the more sensors and the higher the measurement accuracy of the sensors, the larger the observation gain threshold, and the fewer sensors and the lower the measurement accuracy of the sensors, the smaller the observation gain threshold.

[0043] Exemplarily, the robot compares the accumulated observation gain with an observation gain threshold, and if the accumulated observation gain is greater than the observation gain threshold, the obstacle point clouds corresponding to at least two sensors are determined as target obstacle point clouds.

[0044] Step 210: Determine the edge path corresponding to the target time based on the target obstacle point cloud.

[0045] Among them, the edge path refers to the local route of the robot, which can be understood as the local route of the robot running along the obstacle in the edge cleaning mode, for example, the local route of the robot running along the wall in the edge cleaning mode, or the local route of the robot running along the sofa in the edge cleaning mode.

[0046] Exemplarily, the robot generates a target grid map based on the target obstacle point cloud, and generates an edge path corresponding to the target time based on the target grid map.

[0047] In one embodiment, the robot obtains the local pose corresponding to the target moment, multiplies the target obstacle point cloud by the local pose to obtain the target obstacle grid point cloud corresponding to the target obstacle point cloud, generates a target grid map based on the target obstacle grid point cloud, and generates an edge path corresponding to the target moment based on the target grid map.

[0048] In the above-mentioned method for determining the edge path, the target observation gain corresponding to the target moment is determined by the obstacle point clouds corresponding to at least two sensors respectively, and multiple historical observation gains and target observation gains are summed to obtain the cumulative observation gain corresponding to the target moment. The cumulative observation gain represents the total amount of information of the obstacle point clouds acquired by multiple sensors within a period of time including the target moment. If the cumulative observation gain is greater than the observation gain threshold, it means that the amount of information of the obstacle point clouds acquired by multiple sensors within the period of time including the target moment is large, that is, the accuracy of the obstacle point clouds acquired by multiple sensors is high. Then, the obstacle point clouds corresponding to at least two sensors respectively are determined as the target obstacle point clouds, and the edge path corresponding to the target moment is determined based on the target obstacle point cloud, thereby improving the accuracy of the edge path.

[0049] In one embodiment, Figure 3 As shown, based on the obstacle point clouds corresponding to at least two sensors respectively, determining the target observation benefit corresponding to the target time includes:

[0050] Step 302: determine a reference point cloud and at least one candidate point cloud from obstacle point clouds corresponding to at least two sensors respectively; the reference point cloud is the obstacle point cloud corresponding to the sensor with the highest measurement accuracy among the at least two sensors at the target time.

[0051] The reference point cloud refers to the obstacle point cloud corresponding to the sensor with the highest measurement accuracy among at least two sensors at the target time, and the candidate point cloud refers to other obstacle point clouds excluding the reference point cloud.

[0052] Exemplarily, the robot obtains the measurement accuracy corresponding to each sensor, compares the measurement accuracy of multiple sensors, determines the obstacle point cloud corresponding to the sensor with the highest measurement accuracy at the target time as the reference point cloud, and determines the remaining obstacle point clouds as candidate point clouds.

[0053] Step 304 : for each candidate point cloud, based on the candidate point cloud and the reference point cloud, determine the candidate observation gain of the sensor corresponding to the candidate point cloud.

[0054] Among them, the candidate observation benefit represents the useful information amount of the candidate point cloud.

[0055] Exemplarily, for each candidate point cloud, the robot determines the candidate observation benefit corresponding to the candidate point cloud based on the candidate point cloud and the reference point cloud.

[0056] In one embodiment, for each candidate point cloud, the candidate position of at least one grid where the candidate point cloud is located in the grid map and the reference position of at least one grid where the reference point cloud is located in the grid map are determined; based on the straight-line distance between the reference position and the candidate position, the influence weight coefficient corresponding to the candidate position is determined; for each candidate position, based on the straight-line distance between the candidate position and the robot, the distance weight coefficient corresponding to the candidate position is determined; based on the local initial observation yield, influence weight coefficient and distance weight coefficient corresponding to the candidate position, the local adjusted observation yield corresponding to the candidate position is determined; based on the local adjusted observation yields corresponding to multiple candidate positions, the candidate observation yield of the sensor corresponding to the candidate point cloud is obtained.

[0057] Step 306: Based on the reference point cloud, determine the reference observation gain of the sensor corresponding to the reference point cloud.

[0058] Among them, the reference observation gain represents the useful information amount of the reference point cloud.

[0059] Exemplarily, the robot determines the reference position of at least one grid where the reference point cloud is located in the grid map, and for each reference position, determines the straight-line distance between the reference position and the robot, and determines the distance weight coefficient corresponding to the reference position based on the straight-line distance, where the distance weight coefficient is inversely proportional to the straight-line distance. Based on the local initial observation gain, influence weight coefficient and distance weight coefficient corresponding to the reference position, the local adjusted observation gain corresponding to the reference position is obtained, and the local adjusted observation gains corresponding to multiple reference positions are adjusted to obtain the reference observation gain of the sensor corresponding to the reference point cloud.

[0060] In one embodiment, the reference observation yield Gc is as follows:

[0061] Formula (1)

[0062] Formula (2)

[0063] Where n is the total number of grids where the reference point cloud is located in the grid map; a i is the local initial observation gain of the ith grid. Since there is a mapping point corresponding to the reference point cloud in the ith grid, then a i The local initial observation benefit is 1; w a To influence the weight coefficient, since the measurement accuracy of the sensor corresponding to the reference point cloud is the highest, the weight coefficient w a =1;d i is the straight-line distance between the reference position of the i-th grid and the robot; w(d i ) is the distance weight coefficient corresponding to the reference position of the i-th grid; D ais the distance threshold of the sensor corresponding to the reference point cloud. The distance threshold can be set according to the effective measurement distance of the sensor. For example, if the effective measurement distance of the lidar sensor is 1.5 meters, the distance threshold can be set to 1.5 meters. is the preset coefficient of the sensor corresponding to the reference point cloud, and the preset coefficient of the sensor corresponding to the reference point cloud is smaller than the preset coefficient of the sensor corresponding to the candidate point cloud.

[0064] Step 308: sum the at least one candidate observation benefit and the reference observation benefit to obtain the target observation benefit corresponding to the target moment.

[0065] Exemplarily, the robot sums multiple candidate observation benefits and reference observation benefits to obtain the target observation benefit corresponding to the target moment. It can be understood that the summation can be a direct summation or a weighted summation based on the weight coefficients of the candidate observation benefits and the reference observation benefits.

[0066] In this embodiment, the robot determines the reference observation gain of the reference point cloud and the candidate observation gain of the candidate point cloud, and then sums the reference observation gain and multiple candidate observation gains to obtain the target observation gain corresponding to the target moment, providing accurate basic data for the subsequent determination of the cumulative observation gain corresponding to the target moment.

[0067] In one embodiment, Figure 4 As shown, based on the candidate point cloud and the reference point cloud, the candidate observation gain of the sensor corresponding to the candidate point cloud is determined, including:

[0068] Step 402: Determine a reference position of at least one grid where a reference point cloud is located in a grid map, and a candidate position of at least one grid where a candidate point cloud is located in a grid map.

[0069] Among them, the grid map refers to a map composed of multiple grids, that is, the grid map includes multiple grids, and each grid corresponds to a small area in the environment. The reference position refers to the position coordinates of the grid containing the reference mapping point in the grid map. It can be understood that the reference point in the reference point cloud is mapped to the grid map, and a reference mapping point corresponding to the reference point is formed in the grid map. The reference mapping point will exist in some grids of the grid map. The position coordinates of the grid with the reference mapping point are the reference position, and the reference position can be expressed in two-dimensional coordinates. Similarly, the candidate position refers to the position coordinates of the grid containing the candidate mapping point in the grid map.

[0070] Exemplarily, the robot maps the reference point cloud into a grid map, determines the position coordinates of the grid including the reference mapping points as the reference positions, maps the candidate point cloud into the grid map, and determines the position coordinates of the grid including the candidate mapping points as the candidate positions.

[0071] Step 404: for each candidate position, determine the minimum straight-line distance between the candidate position and a plurality of reference positions.

[0072] Exemplarily, for each candidate position, the robot determines the straight-line distance between the candidate position and each reference position, and then compares multiple straight-line distances to obtain the minimum straight-line distance.

[0073] Step 406: Determine the influence weight coefficient corresponding to the candidate position based on the minimum straight-line distance.

[0074] Among them, the influence weight coefficient represents the importance of the candidate mapping point in the grid corresponding to the candidate position, and the influence weight coefficient is proportional to the minimum straight-line distance, that is, the smaller the minimum straight-line distance is, the smaller the influence weight coefficient is, and the larger the minimum straight-line distance is, the larger the influence weight coefficient is.

[0075] Exemplarily, the robot obtains the influence threshold of the sensor corresponding to the candidate point cloud, compares the minimum straight-line distance with the influence threshold, if the minimum straight-line distance is less than the influence threshold, then the first coefficient of the sensor corresponding to the candidate point cloud is determined as the influence weight coefficient, if the minimum straight-line distance is greater than or equal to the influence threshold, then the second coefficient of the sensor corresponding to the candidate point cloud is determined as the influence weight coefficient, and the first coefficient is less than the second coefficient.

[0076] The impact threshold can be determined based on the measurement error of the sensor corresponding to the candidate point cloud. For example, if the measurement error of the sensor is ±2 cm, the impact threshold can be set to 4 cm. The higher the measurement accuracy of the sensor, the larger the first coefficient and the second coefficient of the sensor. It can be understood that when at least two sensors include a first sensor, a second sensor, and a third sensor, the measurement accuracy of the first sensor is greater than the measurement accuracy of the second sensor, and the measurement accuracy of the second sensor is greater than the measurement accuracy of the third sensor, then the first coefficient of the second sensor is greater than the first coefficient of the third sensor, and the second coefficient of the second sensor is greater than the second coefficient of the third sensor. For example, the first coefficient of the second sensor is 0.5, the second coefficient is 1, and the first coefficient of the third sensor is 0.3, and the second coefficient is 0.8.

[0077] Step 408, determining a distance weight coefficient corresponding to the candidate position based on the straight-line distance between the candidate position and the robot.

[0078] The distance weight coefficient represents the accuracy of the candidate mapping point in the grid corresponding to the candidate position.

[0079] Exemplarily, the robot grid position in the grid map at the target moment of the robot is obtained, the straight-line distance between the robot grid position and the candidate position is calculated, and the distance weight coefficient corresponding to the candidate position is determined based on the straight-line distance. The distance weight coefficient corresponding to the candidate position is inversely proportional to the straight-line distance between the candidate position and the robot.

[0080] Step 410, based on the local initial observation benefit, influence weight coefficient and distance weight coefficient corresponding to the candidate position, determine the local adjusted observation benefit corresponding to the candidate position.

[0081] Among them, the local initial observation gain refers to the initial observation gain of the grid corresponding to the candidate position. If there is a candidate mapping point in the grid, the initial observation gain is 1; if there is no candidate mapping point in the grid, the initial observation gain is 0.

[0082] Exemplarily, the robot multiplies the local initial observation benefit, the influence weight coefficient, and the distance weight coefficient corresponding to the candidate position to obtain the local adjusted observation benefit corresponding to the candidate position.

[0083] Step 412, summing up the local adjusted observation gains corresponding to the multiple candidate positions to obtain the candidate observation gain of the sensor corresponding to the candidate point cloud.

[0084] Exemplarily, the robot sums the local adjustment observation benefits corresponding to multiple candidate positions to obtain the candidate observation benefits of the sensor corresponding to the candidate point cloud. Similarly, the summation can be a direct summation or a weighted summation.

[0085] In this embodiment, the influence weight coefficient is determined by the minimum straight-line distance between the candidate position and multiple reference positions. The influence weight coefficient represents the importance of the candidate mapping point in the grid corresponding to the candidate position. It can be understood that the accuracy of the candidate point cloud is lower than that of the reference point cloud. If there is a grid corresponding to the reference position near the grid corresponding to the candidate position, that is, the reference mapping point in the grid corresponding to the reference position is mainly used, then the importance of the candidate mapping point in the grid corresponding to the candidate position is relatively low; the distance weight coefficient corresponding to the candidate position is determined by the straight-line distance between the candidate position and the robot. The distance weight coefficient represents the accuracy of the candidate mapping point in the grid corresponding to the candidate position. It is understood that the smaller the straight-line distance between the candidate position and the robot, the closer the candidate point in the candidate point cloud corresponding to the candidate mapping point is to the robot, the higher the accuracy of the candidate point, and thus the higher the accuracy of the candidate mapping point; the local adjusted observation benefit corresponding to the candidate position is obtained by multiplying the local initial observation benefit, the influence weight coefficient and the distance weight coefficient corresponding to the candidate position, that is, the importance and accuracy of the candidate mapping point in the grid corresponding to the candidate position are comprehensively considered, thereby improving the accuracy of the local adjusted observation benefit, and the local adjusted observation benefits corresponding to multiple candidate positions are summed to obtain the candidate observation benefit of the sensor corresponding to the candidate point cloud, thereby improving the accuracy of the candidate observation benefit.

[0086] In one embodiment, based on the straight-line distance between the candidate position and the robot, determining the distance weight coefficient corresponding to the candidate position includes:

[0087] When the straight-line distance is less than or equal to zero, the distance weight coefficient corresponding to the candidate position is determined to be 1; when the straight-line distance is greater than zero, the distance threshold of the sensor corresponding to the candidate point cloud is obtained, and the straight-line distance is compared with the distance threshold; when the straight-line distance is greater than the distance threshold, the distance weight coefficient corresponding to the candidate position is 0; when the straight-line distance is greater than zero and less than or equal to the distance threshold, the preset coefficient of the sensor corresponding to the candidate point cloud is obtained, and the distance weight coefficient corresponding to the candidate position is determined based on the preset coefficient and the straight-line distance; the distance weight coefficient is inversely proportional to the straight-line distance.

[0088] Among them, the distance threshold refers to a pre-set threshold, and the distance threshold corresponds to the sensor one-to-one, that is, different sensors can correspond to different distance thresholds, and the distance threshold can be set according to the effective measurement distance of the sensor. For example, if the effective measurement distance of the lidar sensor is 1.5 meters, the distance threshold of the lidar sensor can be set to 1.5 meters, the effective measurement distance of the depth camera is 1.0 meters, the distance threshold of the depth camera can be set to 1.0 meters, and the effective measurement distance of the ultrasonic sensor is 0.5 meters, then the distance threshold of the ultrasonic sensor can be set to 0.5 meters. The preset coefficient refers to a pre-set coefficient, and the preset coefficient corresponds to the sensor one-to-one, that is, different sensors can correspond to different preset coefficients. The preset coefficient is inversely proportional to the measurement accuracy of the sensor, that is, the higher the measurement accuracy of the sensor, the smaller the preset coefficient corresponding to the sensor. For example, if the measurement accuracy of the lidar sensor is greater than the measurement accuracy of the depth camera, and the measurement accuracy of the depth camera is greater than the measurement accuracy of the ultrasonic sensor, then the preset coefficient of the lidar sensor can be 0.3, the preset coefficient of the depth camera can be 0.5, and the preset coefficient of the ultrasonic sensor can be 0.8.

[0089] Exemplarily, the robot compares the straight-line distance corresponding to the candidate position with zero. If the straight-line distance is less than or equal to zero, the distance weight coefficient corresponding to the candidate position is determined to be 1; if the straight-line distance is greater than zero, the distance threshold of the sensor corresponding to the candidate point cloud is obtained, and the straight-line distance is compared with the distance threshold. If the straight-line distance is greater than the distance threshold, the distance weight coefficient corresponding to the candidate position is 0; if the straight-line distance is greater than zero and less than or equal to the distance threshold, the preset coefficient of the sensor corresponding to the candidate point cloud is obtained, and the distance weight coefficient corresponding to the candidate position is determined based on the preset coefficient and the straight-line distance.

[0090] In one embodiment, the distance weight factor is as follows:

[0091] Formula (3)

[0092] Wherein, x is the sensor identifier, for example, x=1 indicates the first sensor, x=2 indicates the second sensor, and x=3 indicates the third sensor; d xi is the straight-line distance between the reference position of the ith grid in the grid map where the obstacle point cloud corresponding to the xth sensor is located and the robot; w(d xi ) is the distance weight coefficient corresponding to the reference position of the ith grid in the grid map where the obstacle point cloud corresponding to the xth sensor is located; D xa is the distance threshold of the x-th sensor; α x is the preset coefficient of the xth sensor, α x It is inversely proportional to the measurement accuracy of the sensor.

[0093] In this embodiment, the distance weight coefficient corresponding to the candidate position is determined by the straight-line distance between the candidate position and the robot. The distance weight coefficient is inversely proportional to the straight-line distance. It can be understood that the smaller the straight-line distance between the candidate position and the robot, the closer the grid corresponding to the candidate position is to the robot, the higher the accuracy of the candidate mapping points in the grid, and the larger the amount of available information. In this way, a larger distance weight coefficient is used when calculating the local adjustment observation benefit of the grid. Similarly, the larger the straight-line distance between the candidate position and the robot, the farther the grid corresponding to the candidate position is from the robot, the lower the accuracy of the candidate mapping points in the grid, and the smaller the amount of available information. In this way, a smaller distance weight coefficient is used when calculating the local adjustment observation benefit of the grid, thereby improving the accuracy of the local adjustment observation benefit.

[0094] In one embodiment, at least two sensors include a first sensor, a second sensor, and a third sensor, the measurement accuracy of the first sensor is greater than the measurement accuracy of the second sensor, and the measurement accuracy of the second sensor is greater than the measurement accuracy of the third sensor, then the distance threshold of the first sensor is greater than the distance threshold of the second sensor, the distance threshold of the second sensor is greater than the distance threshold of the third sensor, the preset coefficient of the first sensor is less than the preset coefficient of the second sensor, and the preset coefficient of the second sensor is less than the preset coefficient of the third sensor.

[0095] Exemplarily, in the case where at least two sensors include a first sensor, a second sensor, and a third sensor, the measurement accuracy of the first sensor is greater than the measurement accuracy of the second sensor, and the measurement accuracy of the second sensor is greater than the measurement accuracy of the third sensor, then the distance threshold of the first sensor is greater than the distance threshold of the second sensor, the distance threshold of the second sensor is greater than the distance threshold of the third sensor, the preset coefficient of the first sensor is smaller than the preset coefficient of the second sensor, and the preset coefficient of the second sensor is smaller than the preset coefficient of the third sensor.

[0096] In this embodiment, the distance threshold corresponding to the sensor with higher measurement accuracy is larger and the preset coefficient is smaller. For the candidate position whose straight-line distance is greater than zero and less than or equal to the distance threshold, the distance weight coefficient corresponding to the candidate position is determined based on the preset coefficient of the corresponding sensor and the straight-line distance. For the sensor with high measurement accuracy, the smaller the preset coefficient corresponding to the sensor, the slower the distance weight coefficient decreases with the increase of the straight-line distance. For the sensor with low measurement accuracy, the larger the preset coefficient corresponding to the sensor, the faster the distance weight coefficient decreases with the increase of the straight-line distance. Therefore, under the condition of the same straight-line distance, the distance weight coefficient of the candidate position corresponding to the sensor with higher measurement accuracy is greater than the distance weight coefficient of the candidate position corresponding to the sensor with lower measurement accuracy.

[0097] In one embodiment, Figure 5 As shown, based on the target obstacle point cloud, the edge path corresponding to the target time is determined, including:

[0098] Step 502, obtaining the local pose and global pose corresponding to the target moment.

[0099] Among them, the local pose refers to the pose of the robot in the robot coordinate system, and the global pose refers to the pose of the robot in the world coordinate system.

[0100] Exemplarily, the robot obtains the local pose corresponding to the target time through the odometer and IMU (Inertial Measurement Unit), and obtains the global pose corresponding to the target time through SLAM (Simultaneous Localization and Mapping) or GPS (Global Positioning System).

[0101] Step 504, based on the ratio between the global pose and the local pose, determine the transformation matrix corresponding to the target moment.

[0102] Exemplarily, the robot calculates the ratio between the global pose and the local pose to obtain a transformation matrix corresponding to the target moment.

[0103] In one embodiment, the transformation matrix T offset As shown below:

[0104] T global= T local* T offset Formula (4)

[0105] Among them, T global is the global pose; T local is the local pose.

[0106] Step 506: Based on the target obstacle point cloud and the local pose, a target obstacle grid point cloud corresponding to the target obstacle point cloud is obtained.

[0107] Exemplarily, the robot multiplies the target obstacle point cloud with the local pose to obtain a target obstacle grid point cloud corresponding to the target obstacle point cloud.

[0108] In one embodiment, the target obstacle point cloud is P i ={p i1 ,p i2 ,……,p in}, i is the identifier of the obstacle point cloud, P inRepresents the point of the nth obstacle point cloud in the obstacle point cloud of the i-th frame. For example, P1 represents the obstacle point cloud of the first frame, and the target obstacle grid point cloud As shown below:

[0109] Formula (5)

[0110] Among them, T local is the local pose.

[0111] Step 508: Based on the global position and the transformation matrix corresponding to the target point, the grid position corresponding to the target point is obtained.

[0112] The target point refers to the specific location that the robot needs to reach. The target point can be defined in the robot coordinate system or in the world coordinate system. The target point can be a pre-set location, which can be set by the user or according to the cleaning task. The global position refers to the position of the target point in the world coordinate system.

[0113] Exemplarily, the robot obtains the global position of the target point corresponding to the cleaning task, divides the global position by the change matrix, and obtains the grid position corresponding to the target point.

[0114] In one embodiment, the number of target points corresponding to the cleaning task is multiple, and the robot obtains a target point set, the target point set P global ={p global1 ,p global2 ,……,p globalm},p globalm is the global position of the mth target point. The target point set is divided by the change matrix to obtain the grid position set corresponding to the target point set. The grid position set is as follows:

[0115] Formula (6)

[0116] Among them, T offset is the transformation matrix.

[0117] Step 510, based on the target obstacle grid point cloud and grid position, determine the edge path corresponding to the target time.

[0118] Exemplarily, in the absence of a virtual obstacle, the robot projects the target obstacle grid point cloud and grid position into the grid map to obtain a target grid map, and generates an edge path corresponding to the target time based on the target grid map.

[0119] In this embodiment, in the absence of virtual obstacles, the target obstacle point cloud is converted into the corresponding target obstacle grid point cloud in the robot coordinate system, and the global position corresponding to the target point is converted into the corresponding grid position in the robot coordinate system. An edge path is generated according to the corresponding target obstacle grid point cloud and grid position in the robot coordinate system, so that the edge path has local smoothness and avoids positioning jitter.

[0120] In one embodiment, before determining the edge path corresponding to the target time based on the target obstacle grid point cloud and the grid position, and determining the edge path corresponding to the target time based on the target obstacle grid point cloud and the grid position, including:

[0121] Based on the virtual obstacle point cloud and the transformation matrix corresponding to the virtual obstacle, a virtual obstacle grid point cloud corresponding to the virtual obstacle is obtained; based on the target obstacle grid point cloud, the grid position and the virtual obstacle grid point cloud, the edge path corresponding to the target time is determined.

[0122] Among them, virtual obstacles refer to obstacles used as constraints in the robot's navigation and path planning process. Virtual obstacles are used to define areas that the robot cannot enter or obstacles that need to be avoided. Virtual obstacles include but are not limited to virtual walls, restricted areas, or obstacle avoidance areas. Virtual obstacle point cloud refers to a collection of virtual obstacles. The position coordinates in the virtual obstacle point cloud are in the world coordinate system.

[0123] Exemplarily, in the presence of a virtual obstacle, the robot obtains a virtual obstacle point cloud corresponding to the virtual obstacle, divides the virtual obstacle point cloud by the transformation matrix to obtain a virtual obstacle grid point cloud corresponding to the virtual obstacle, maps the target obstacle grid point cloud, grid position and virtual obstacle grid point cloud to a grid map to obtain a target grid map, and generates an edge path corresponding to the target time based on the target grid map.

[0124] In one embodiment, the virtual obstacle point cloud is H global , virtual obstacle grid point cloud

[0125] As shown below:

[0126] Formula (7)

[0127] Among them, T offset is the transformation matrix.

[0128] In this embodiment, in the presence of virtual obstacles, the target obstacle point cloud, the global position corresponding to the target point, and the virtual obstacle point cloud are converted into the corresponding target obstacle grid point cloud, the grid position corresponding to the target point, and the virtual obstacle grid point cloud in the robot coordinate system. An edge path is generated according to the corresponding target obstacle grid point cloud, the grid position corresponding to the target point, and the virtual obstacle grid point cloud in the robot coordinate system, so that the edge path has local smoothness and avoids positioning jitter.

[0129] In an exemplary embodiment, when the robot performs a cleaning task using the edge cleaning mode, it obtains obstacle point clouds corresponding to the lidar sensor, the depth camera, and the ultrasonic sensor at the target time. Since the measurement accuracy of the lidar sensor is greater than that of the depth camera, and the measurement accuracy of the depth camera is greater than that of the ultrasonic sensor, the obstacle point cloud corresponding to the lidar sensor is determined as the reference point cloud, and the obstacle point cloud corresponding to the depth camera and the obstacle point cloud corresponding to the ultrasonic sensor are determined as candidate point clouds.

[0130] The robot maps the reference point cloud to the grid map, determines the position coordinates of the grid including the reference mapping point as the reference position, maps the candidate point cloud to the grid map, and determines the position coordinates of the grid including the candidate mapping point as the candidate position. For each reference position, determine the straight-line distance between the reference position and the robot, and use formula (3) to determine the distance weight coefficient corresponding to the reference position, where D xa =1.5 m, α x =0.3; multiply the local initial observation benefit, influence weight coefficient and distance weight coefficient corresponding to the reference position to obtain the local adjusted observation benefit corresponding to the reference position, where the influence weight coefficient is 1; adjust the local adjusted observation benefits corresponding to multiple reference positions to obtain the reference observation benefit of the sensor corresponding to the reference point cloud.

[0131] For each candidate position corresponding to the candidate point cloud, determine the straight-line distance between the candidate position and each reference position, and then compare multiple straight-line distances to obtain the minimum straight-line distance; compare the minimum straight-line distance with the influence threshold, if the minimum straight-line distance is less than the influence threshold, the first coefficient of the sensor corresponding to the candidate point cloud is determined as the influence weight coefficient, if the minimum straight-line distance is greater than or equal to the influence threshold, the second coefficient of the sensor corresponding to the candidate point cloud is determined as the influence weight coefficient, and the first coefficient is less than the second coefficient. Among them, the influence threshold is 4 cm, if the candidate point cloud is an obstacle point cloud corresponding to the depth camera, the first coefficient is 0.5, and the second coefficient is 1; if the candidate point cloud is an obstacle point cloud corresponding to the ultrasonic sensor, the first coefficient is 0.3, and the second coefficient is 0.8.

[0132] Get the robot grid position in the grid map at the target moment, calculate the straight-line distance between the robot grid position and the candidate position, and use formula (3) to determine the distance weight coefficient corresponding to the reference position. If the candidate point cloud is the obstacle point cloud corresponding to the depth camera, then D xa = 1.0 m, α x =0.5; if the candidate point cloud is the obstacle point cloud corresponding to the ultrasonic sensor, then D xa = 0.5 m, α x =0.8.

[0133] The robot multiplies the local initial observation benefit, influence weight coefficient and distance weight coefficient corresponding to the candidate position to obtain the local adjusted observation benefit corresponding to the candidate position; the local adjusted observation benefits corresponding to multiple candidate positions are summed to obtain the candidate observation benefit of the sensor corresponding to the candidate point cloud.

[0134] The reference observation gain corresponding to the lidar sensor, the candidate observation gain corresponding to the depth camera, and the candidate observation gain corresponding to the ultrasonic sensor are summed to obtain the cumulative observation gain corresponding to the target moment. The robot obtains the historical observation gain corresponding to a preset number of historical moments before the target moment, sums the multiple historical observation gains and the target observation gain, and obtains the cumulative observation gain corresponding to the target moment; the cumulative observation gain is compared with the observation gain threshold, and if the cumulative observation gain is greater than the observation gain threshold, the obstacle point cloud corresponding to at least two sensors is determined as the target obstacle point cloud.

[0135] The robot obtains the local pose corresponding to the target time through the odometer and IMU, and obtains the global pose corresponding to the target time through SLAM or GPS; calculates the ratio between the global pose and the local pose to obtain the transformation matrix corresponding to the target time; the robot multiplies the target obstacle point cloud with the local position to obtain the target obstacle grid point cloud corresponding to the target obstacle point cloud; obtains the global position of the target point corresponding to the cleaning task, divides the global position by the change matrix, and obtains the grid position corresponding to the target point; obtains the virtual obstacle point cloud corresponding to the virtual obstacle, divides the virtual obstacle point cloud by the transformation matrix, and obtains the virtual obstacle grid point cloud corresponding to the virtual obstacle; maps the target obstacle grid point cloud, grid position and virtual obstacle grid point cloud to the grid map to obtain the target grid map, and generates the edge path corresponding to the target time based on the target grid map.

[0136] The above-mentioned method for determining the edge path determines the target observation gain corresponding to the target moment through the obstacle point clouds corresponding to at least two sensors respectively, and sums multiple historical observation gains and target observation gains to obtain the cumulative observation gain corresponding to the target moment. The cumulative observation gain represents the total amount of information of the obstacle point clouds acquired by multiple sensors within a period of time including the target moment. If the cumulative observation gain is greater than the observation gain threshold, it means that the amount of information of the obstacle point clouds acquired by multiple sensors within the period of time including the target moment is large, that is, the accuracy of the obstacle point clouds acquired by multiple sensors is high. Then, the obstacle point clouds corresponding to at least two sensors respectively are determined as the target obstacle point clouds, and the edge path corresponding to the target moment is determined based on the target obstacle point cloud, thereby improving the accuracy of the edge path. Moreover, the target obstacle point cloud, the global position corresponding to the target point and the virtual obstacle point cloud are converted into the corresponding target obstacle grid point cloud, the grid position corresponding to the target point and the virtual obstacle grid point cloud in the robot coordinate system, and an edge path is generated according to the corresponding target obstacle grid point cloud, the grid position corresponding to the target point and the virtual obstacle grid point cloud in the robot coordinate system, so that the edge path has local smoothness and avoids positioning jitter.

[0137] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0138] Based on the same inventive concept, the embodiment of the present application also provides an edge path determination device for implementing the edge path determination method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more edge path determination device embodiments provided below can refer to the limitations of the edge path determination method above, and will not be repeated here.

[0139] In one embodiment, Figure 6 As shown, a device for determining an edge path is provided, comprising: an acquisition module 602, a calculation module 604, a summation module 606, a comparison module 608 and a determination module 610, wherein:

[0140] An acquisition module 602 is used to acquire obstacle point clouds corresponding to at least two sensors at a target time;

[0141] A calculation module 604, configured to determine a target observation benefit corresponding to a target time based on obstacle point clouds corresponding to at least two sensors respectively;

[0142] A summing module 606 is used to sum a plurality of historical observation benefits and a target observation benefit to obtain a cumulative observation benefit corresponding to a target time; the historical time corresponding to the historical observation benefit is before the target time;

[0143] A comparison module 608 is used to determine the obstacle point clouds corresponding to at least two sensors as target obstacle point clouds when the accumulated observation benefit is greater than the observation benefit threshold;

[0144] The determination module 610 is used to determine the edge path corresponding to the target time based on the target obstacle point cloud.

[0145] In one embodiment, the calculation module 604 is also used to: determine a reference point cloud and at least one candidate point cloud from the obstacle point clouds corresponding to at least two sensors respectively; the reference point cloud is the obstacle point cloud corresponding to the sensor with the highest measurement accuracy among the at least two sensors at the target time; for each candidate point cloud, determine the candidate observation gain of the sensor corresponding to the candidate point cloud based on the candidate point cloud and the reference point cloud; determine the reference observation gain of the sensor corresponding to the reference point cloud based on the reference point cloud; sum at least one candidate observation gain and the reference observation gain to obtain the target observation gain corresponding to the target time.

[0146] In one embodiment, the calculation module 604 is also used to: determine the reference position of at least one grid where the reference point cloud is located in the grid map, and the candidate positions of multiple grids where the candidate point cloud is located in the grid map; for each candidate position, determine the minimum straight-line distance between the candidate position and at least one reference position; based on the minimum straight-line distance, determine the influence weight coefficient corresponding to the candidate position; based on the straight-line distance between the candidate position and the robot, determine the distance weight coefficient corresponding to the candidate position; based on the local initial observation benefit, influence weight coefficient and distance weight coefficient corresponding to the candidate position, determine the local adjusted observation benefit corresponding to the candidate position; sum the local adjusted observation benefits corresponding to multiple candidate positions to obtain the candidate observation benefit of the sensor corresponding to the candidate point cloud.

[0147] In one embodiment, the calculation module 604 is also used to: when the straight-line distance is less than or equal to zero, determine that the distance weight coefficient corresponding to the candidate position is 1; when the straight-line distance is greater than zero, obtain the distance threshold of the sensor corresponding to the candidate point cloud, and compare the straight-line distance with the distance threshold; when the straight-line distance is greater than the distance threshold, set the distance weight coefficient corresponding to the candidate position to 0; when the straight-line distance is greater than zero and less than or equal to the distance threshold, obtain the preset coefficient of the sensor corresponding to the candidate point cloud, and determine the distance weight coefficient corresponding to the candidate position based on the preset coefficient and the straight-line distance; the distance weight coefficient is inversely proportional to the straight-line distance.

[0148] In one embodiment, at least two sensors include a first sensor, a second sensor, and a third sensor, the measurement accuracy of the first sensor is greater than the measurement accuracy of the second sensor, and the measurement accuracy of the second sensor is greater than the measurement accuracy of the third sensor, then the distance threshold of the first sensor is greater than the distance threshold of the second sensor, the distance threshold of the second sensor is greater than the distance threshold of the third sensor, the preset coefficient of the first sensor is less than the preset coefficient of the second sensor, and the preset coefficient of the second sensor is less than the preset coefficient of the third sensor.

[0149] In one embodiment, the determination module 610 is also used to: obtain the local pose and the global pose corresponding to the target moment; determine the transformation matrix corresponding to the target moment based on the ratio between the global pose and the local pose; obtain the target obstacle grid point cloud corresponding to the target obstacle point cloud based on the target obstacle point cloud and the local pose; obtain the grid position corresponding to the target point based on the global position and the transformation matrix corresponding to the target point; determine the edge path corresponding to the target moment based on the target obstacle grid point cloud and the grid position.

[0150] In one embodiment, the determination module 610 is further used to: obtain a virtual obstacle grid point cloud corresponding to the virtual obstacle based on the virtual obstacle point cloud and the transformation matrix corresponding to the virtual obstacle; and determine the edge path corresponding to the target time based on the target obstacle grid point cloud, the grid position and the virtual obstacle grid point cloud.

[0151] Each module in the above-mentioned edge path determination device can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of the processor in the robot in the form of hardware, or can be stored in the memory in the robot in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0152] In one embodiment, a robot is provided. The robot may be a terminal, and its internal structure diagram may be as follows: Figure 7As shown. The robot includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. The processor of the robot is used to provide computing and control capabilities. The memory of the robot includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the robot is used to exchange information between the processor and an external device. The communication interface of the robot is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be realized through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a method for determining an edge path is implemented. The display unit of the robot is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the robot can be a touch layer covering the display screen, or a button, trackball or touchpad set on the robot shell, or an external keyboard, touchpad or mouse.

[0153] Those skilled in the art will understand that Figure 7 The structure shown in the figure is merely a block diagram of a partial structure related to the scheme of the present application, and does not constitute a limitation on the robot to which the scheme of the present application is applied. The specific robot may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.

[0154] In one embodiment, a robot is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0155] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0156] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0157] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0158] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0159] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0160] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for determining an edge path, characterized in that: The method comprises: In the process of performing a cleaning task using the edge cleaning mode, obtaining obstacle point clouds corresponding to at least two sensors at target times; Determine a reference point cloud and at least one candidate point cloud from obstacle point clouds corresponding to at least two sensors respectively; the reference point cloud is an obstacle point cloud corresponding to the sensor with the highest measurement accuracy among the at least two sensors at the target time; For each of the candidate point clouds, determining a reference position of at least one grid where the reference point cloud is located in the grid map, and a candidate position of at least one grid where the candidate point cloud is located in the grid map; For each of the candidate positions, determining a minimum straight-line distance between the candidate position and a plurality of reference positions; Based on the minimum straight-line distance, determining an influence weight coefficient corresponding to the candidate position; Determine a distance weight coefficient corresponding to the candidate position based on a straight-line distance between the candidate position and the robot; Determine the local adjusted observation benefit corresponding to the candidate position based on the local initial observation benefit corresponding to the candidate position, the influence weight coefficient and the distance weight coefficient; Summing the local adjusted observation gains corresponding to the multiple candidate positions to obtain the candidate observation gain of the sensor corresponding to the candidate point cloud; Based on the reference point cloud, determining a reference observation gain of a sensor corresponding to the reference point cloud; Summing at least one of the candidate observation benefits and the reference observation benefit to obtain a target observation benefit corresponding to the target moment; The cumulative observation benefit corresponding to the target moment is obtained by summing up multiple historical observation benefits and the target observation benefit; the historical moment corresponding to the historical observation benefit is before the target moment; When the accumulated observation benefit is greater than the observation benefit threshold, determining the obstacle point clouds corresponding to at least two sensors as target obstacle point clouds; Based on the target obstacle point cloud, determine the edge path corresponding to the target time.

2. The method according to claim 1, characterized in that The step of determining a reference point cloud and at least one candidate point cloud from obstacle point clouds respectively corresponding to at least two sensors comprises: Obtaining the measurement accuracy corresponding to each of the sensors; Comparing the measurement accuracy of a plurality of said sensors; The obstacle point cloud corresponding to the sensor with the highest measurement accuracy at the target time is determined as the reference point cloud, and the remaining obstacle point clouds are determined as candidate point clouds.

3. The method according to claim 1, characterized in that The determining, based on the reference point cloud, a reference observation gain of a sensor corresponding to the reference point cloud comprises: Determining a reference position of at least one grid where the reference point cloud is located in the grid map; For each of the reference positions, determine the straight-line distance between the reference position and the robot, and determine a distance weight coefficient corresponding to the reference position based on the straight-line distance; the distance weight coefficient is inversely proportional to the straight-line distance; Based on the local initial observation benefit corresponding to the reference position, the influence weight coefficient and the distance weight coefficient, a local adjusted observation benefit corresponding to the reference position is obtained; The local adjustment observation gains corresponding to the multiple reference positions are counted to obtain the reference observation gain of the sensor corresponding to the reference point cloud.

4. The method according to claim 1, characterized in that: The determining of the distance weight coefficient corresponding to the candidate position based on the straight-line distance between the candidate position and the robot includes: When the straight-line distance is less than or equal to zero, determining the distance weight coefficient corresponding to the candidate position to be 1; When the straight-line distance is greater than zero, obtaining a distance threshold of the sensor corresponding to the candidate point cloud, and comparing the straight-line distance with the distance threshold; When the straight-line distance is greater than the distance threshold, the distance weight coefficient corresponding to the candidate position is set to 0; When the straight-line distance is greater than zero and less than or equal to the distance threshold, the preset coefficient of the sensor corresponding to the candidate point cloud is obtained, and the distance weight coefficient corresponding to the candidate position is determined based on the preset coefficient and the straight-line distance; the distance weight coefficient is inversely proportional to the straight-line distance.

5. The method according to claim 4, characterized in that The at least two sensors include a first sensor, a second sensor and a third sensor. The measurement accuracy of the first sensor is greater than that of the second sensor, and the measurement accuracy of the second sensor is greater than that of the third sensor. Then, the distance threshold of the first sensor is greater than that of the second sensor, the distance threshold of the second sensor is greater than that of the third sensor, the preset coefficient of the first sensor is less than that of the second sensor, and the preset coefficient of the second sensor is less than that of the third sensor.

6. The method according to claim 1, characterized in that The determining, based on the target obstacle point cloud, the edge path corresponding to the target time comprises: Obtaining the local pose and global pose corresponding to the target moment; Determining a transformation matrix corresponding to the target moment based on a ratio between the global pose and the local pose; Based on the target obstacle point cloud and the local pose, obtaining a target obstacle grid point cloud corresponding to the target obstacle point cloud; Based on the global position corresponding to the target point and the transformation matrix, a grid position corresponding to the target point is obtained; Based on the target obstacle grid point cloud and the grid position, determine the edge path corresponding to the target time.

7. The method according to claim 6, characterized in that Before determining the edge path corresponding to the target time based on the target obstacle grid point cloud and the grid position, the method further includes: Based on the virtual obstacle point cloud corresponding to the virtual obstacle and the transformation matrix, obtaining a virtual obstacle grid point cloud corresponding to the virtual obstacle; The determining, based on the target obstacle grid point cloud and the grid position, the edge path corresponding to the target time, comprises: Based on the target obstacle grid point cloud, the grid position and the virtual obstacle grid point cloud, a path along the edge corresponding to the target time is determined.

8. A device for determining a path along an edge, characterized in that: The device comprises: An acquisition module, used for acquiring obstacle point clouds corresponding to at least two sensors at target times in the process of performing a cleaning task using the edge cleaning mode; A calculation module is used to determine a reference point cloud and at least one candidate point cloud from obstacle point clouds corresponding to at least two sensors respectively; the reference point cloud is the obstacle point cloud corresponding to the sensor with the highest measurement accuracy among the at least two sensors at the target time; for each candidate point cloud, determine the reference position of at least one grid in the grid map where the reference point cloud is located, and the candidate position of at least one grid in the grid map where the candidate point cloud is located; for each candidate position, determine the minimum straight-line distance between the candidate position and a plurality of reference positions; based on the minimum straight-line distance, determine the shadow position corresponding to the candidate position The influence weight coefficient is determined; based on the straight-line distance between the candidate position and the robot, the distance weight coefficient corresponding to the candidate position is determined; based on the local initial observation benefit corresponding to the candidate position, the influence weight coefficient and the distance weight coefficient, the local adjusted observation benefit corresponding to the candidate position is determined; the local adjusted observation benefits corresponding to multiple candidate positions are summed to obtain the candidate observation benefit of the sensor corresponding to the candidate point cloud; based on the reference point cloud, the reference observation benefit of the sensor corresponding to the reference point cloud is determined; at least one candidate observation benefit and the reference observation benefit are summed to obtain the target observation benefit corresponding to the target moment; A summing module, used for summing a plurality of historical observation benefits and the target observation benefit to obtain a cumulative observation benefit corresponding to the target moment; the historical moment corresponding to the historical observation benefit is before the target moment; A comparison module, configured to determine the obstacle point clouds corresponding to at least two sensors as target obstacle point clouds when the cumulative observation benefit is greater than an observation benefit threshold; A determination module is used to determine the edge path corresponding to the target moment based on the target obstacle point cloud.

9. A robot comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.