Obstacle avoidance method, device, equipment and medium for sanitation robots in complex scenarios

Abstract

The present application provides a method, device, equipment and medium for obstacle avoidance of a sanitation robot in complex scenarios. The method includes: separating noise data from environmental data to obtain multi-source preliminary signals; extracting the spatial coordinates and velocity vectors of each dynamic obstacle, determining the position of the dynamic obstacle and the target motion trend, and the position of the static obstacle, and determining the obstacle avoidance sequence of the sanitation robot; extracting the key trigger points and their characteristic vectors for obstacle avoidance; determining whether the sanitation robot has a potential deadlock area; if there is a potential deadlock area, adjusting the path weight to obtain an optimized planned path; extracting adjustment parameters from the optimized planned path to determine the relative displacement trend; prioritizing each obstacle, determining the target obstacle avoidance path, and controlling the movement of the sanitation robot. In this way, by separating noise data from environmental data and formulating a collaborative obstacle avoidance strategy for dynamic and static obstacles, the accuracy of obstacle avoidance in complex scenarios can be improved.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving control, and in particular to a method, device, equipment and medium for obstacle avoidance of a sanitation robot in complex scenarios. Background Art

[0002] Sanitation robots operate in complex environments, such as plazas and scenic areas, where there is heavy foot traffic, static trash cans and roadblocks, and dynamic pedestrians and vehicles. These obstacles can hinder the robots' cleaning performance. Traditional path planning algorithms can become locked due to obstacles or closed passages, or repeatedly adjust the path, leading to ineffective detours and reduced cleaning efficiency. Therefore, the aforementioned shortcomings of existing sanitation robots in obstacle avoidance control are a pressing technical issue that needs to be addressed. Summary of the Invention

[0003] The purpose of the embodiments of the present application is to provide a method for sanitation robots to avoid obstacles in complex scenarios, so as to improve the accuracy of obstacle avoidance in complex scenarios.

[0004] In a first aspect, an embodiment of the present application provides a method for avoiding obstacles in complex scenarios for a sanitation robot, the method comprising:

[0005] Separating noise data from environmental data within a preset range of the sanitation robot to obtain a multi-source preliminary signal, wherein the environmental data includes at least the speed and distance of dynamic obstacles and the distance of static obstacles within the preset range;

[0006] Extracting the spatial coordinates and velocity vector of each of the dynamic obstacles from the multi-source preliminary signals, and determining the position and target motion trend of each of the dynamic obstacles, as well as the position of each of the static obstacles based on a spatial mapping algorithm;

[0007] Determining an obstacle avoidance sequence for the sanitation robot according to the position and target motion trend of each of the dynamic obstacles, and the position of each of the static obstacles;

[0008] Extracting key triggering points for obstacle avoidance of the sanitation robot and feature vectors of the key triggering points from the obstacle avoidance sequence;

[0009] Based on the A-star algorithm, according to the characteristic vector of the key trigger point and the preset operation time of the sanitation robot, it is determined whether the sanitation robot has a potential deadlock area;

[0010] In response to the presence of a deadlock potential area for the sanitation robot, adjusting a path weight according to the deadlock potential area to obtain an optimized planned path, wherein the path weight includes at least a dynamic obstacle weight and a deadlock risk weight;

[0011] Extracting adjustment parameters from the optimized planned path to determine the relative displacement trend of each of the dynamic obstacles and each of the static obstacles, wherein the adjustment parameters include the speed of the sanitation robot;

[0012] According to the relative displacement trend, the obstacles are prioritized and a target obstacle avoidance path is determined;

[0013] The movement of the sanitation robot is controlled based on the target obstacle avoidance path.

[0014] In some embodiments, separating noise data from environmental data within a preset range of the sanitation robot to obtain a multi-source preliminary signal includes:

[0015] Dividing the environmental data based on multiple independent signal channels to obtain multiple channel data;

[0016] Performing filtering processing on the plurality of sub-channel data by using a parallel filter to obtain a filtered signal;

[0017] The noise data of the filtered signal is separated by an adaptive weighted average algorithm to obtain a multi-source preliminary signal.

[0018] In some embodiments, extracting the spatial coordinates and velocity vector of each of the dynamic obstacles from the multi-source preliminary signals, and determining the position and target motion trend of each of the dynamic obstacles, as well as the position of each of the static obstacles based on a spatial mapping algorithm, includes:

[0019] Decomposing the multi-source preliminary signals to obtain the spatial coordinates and velocity vector of each of the dynamic obstacles;

[0020] By using a grid division method, a mapping structure is constructed based on the spatial coordinates and velocity vectors of each of the dynamic obstacles to obtain initial spatial distribution data;

[0021] Extracting the position corresponding to each of the static obstacles and the position of each of the dynamic obstacles from the initial spatial distribution data;

[0022] The target motion trend of each of the dynamic obstacles is determined according to the velocity vector.

[0023] In some embodiments, determining the target motion trend of each of the dynamic obstacles based on the velocity vector includes:

[0024] determining an initial motion trend of each of the dynamic obstacles according to the velocity vector;

[0025] determining an overlapping area between each of the dynamic obstacles and each of the static obstacles based on an initial movement trend of each of the dynamic obstacles;

[0026] For the mapping structure within the overlapping area, separate the static obstacles from the dynamic obstacles by adjusting the grid to obtain an optimized obstacle point distribution, wherein the obstacle points include the static obstacles and the dynamic obstacles;

[0027] According to the optimized obstacle point distribution, the velocity vectors are integrated to determine the target motion trend of each dynamic obstacle point;

[0028] The optimized obstacle point distribution is:

[0029]

[0030] in, represents the obstacle distribution density function, Indicates the number of sampling points, represents the weight coefficient, represents the attenuation coefficient, Indicates the reference point position, Indicates the location of the obstacle point.

[0031] In some embodiments, determining an obstacle avoidance sequence for the sanitation robot based on the position and target motion trend of each of the dynamic obstacles, and the position of each of the static obstacles, includes:

[0032] Based on the time window sliding method, determining the overlapping area of ​​each dynamic obstacle according to the position of each dynamic obstacle and the target movement trend;

[0033] In the case where the overlapping area contains static obstacles, determining the priority index of each dynamic obstacle by a vector decomposition method;

[0034] Based on the priority index of each of the dynamic obstacles, the dynamic obstacles are sorted to determine the obstacle avoidance sequence of the sanitation robot;

[0035] The priority index is:

[0036]

[0037] in, Represents the priority index, represents the priority coefficient, represents the weight factor, Indicates the obstacle distance, represents the scale parameter, Indicates the number of obstacles.

[0038] In some embodiments, extracting adjustment parameters from the optimized planned path and determining the relative displacement trend of each of the dynamic obstacles and each of the static obstacles includes:

[0039] Extracting adjustment parameters from the optimized planned path, wherein the adjustment parameters include the speed of the sanitation robot;

[0040] According to the distribution changes of the dynamic obstacles and the static obstacles, a Kalman filter is used to fuse the environmental data to obtain the preliminary displacement distribution of each dynamic obstacle;

[0041] Extracting the overlapping area of ​​the dynamic obstacle and the static obstacle according to the preliminary displacement distribution, and determining the boundary range of each dynamic obstacle;

[0042] Update the path planning parameters according to the boundary range to obtain an optimized path sequence;

[0043] The relative displacement trends of the dynamic obstacles and the static obstacles are determined according to the optimized path sequence.

[0044] In some embodiments, prioritizing obstacles based on the relative displacement trend and determining a target obstacle avoidance path includes:

[0045] Performing filtering and smoothing processing on the relative displacement trend to obtain a stable displacement trend distribution;

[0046] extracting the velocity of each of the dynamic obstacles according to the stable displacement trend distribution;

[0047] Marking the dynamic obstacle whose speed exceeds a preset speed threshold as a high priority obstacle;

[0048] For each of the high priority obstacles, determining an absolute difference between a speed and the preset speed threshold;

[0049] Sort the absolute differences corresponding to the high-priority obstacles in descending order to obtain a target sorting sequence;

[0050] Determining a path angle adjustment value for each of the dynamic obstacles based on the target sorting sequence;

[0051] A target obstacle avoidance path is determined according to the path deflection angle adjustment value.

[0052] In a second aspect, an embodiment of the present application provides an obstacle avoidance device for a sanitation robot in complex scenarios, the device comprising:

[0053] a separation module, configured to separate noise data from environmental data within a preset range of the sanitation robot to obtain a multi-source preliminary signal, wherein the environmental data includes at least the speed and distance of dynamic obstacles and the distance of static obstacles within the preset range;

[0054] a first extraction module, configured to extract the spatial coordinates and velocity vector of each of the dynamic obstacles from the multi-source preliminary signals, and determine the position and target motion trend of each of the dynamic obstacles, as well as the position of each of the static obstacles based on a spatial mapping algorithm;

[0055] A first determination module is configured to determine an obstacle avoidance sequence for the sanitation robot based on the position and target motion trend of each of the dynamic obstacles and the position of each of the static obstacles;

[0056] A second extraction module is used to extract the key triggering points of the sanitation robot's obstacle avoidance and the feature vectors of the key triggering points from the obstacle avoidance sequence;

[0057] a judgment module, configured to judge whether the sanitation robot has a deadlock potential area based on the A-star algorithm and the characteristic vector of the key trigger point and the preset operation duration of the sanitation robot;

[0058] an adjustment module, configured to, in response to a deadlock potential area existing in the sanitation robot, adjust a path weight according to the deadlock potential area to obtain an optimized planned path, wherein the path weight includes at least a dynamic obstacle weight and a deadlock risk weight;

[0059] a second determination module, configured to extract adjustment parameters from the optimized planned path and determine a relative displacement trend between each of the dynamic obstacles and each of the static obstacles, wherein the adjustment parameters include a speed of the sanitation robot;

[0060] a third determination module, configured to prioritize obstacles according to the relative displacement trend and determine a target obstacle avoidance path;

[0061] A control module is used to control the movement of the sanitation robot based on the target obstacle avoidance path.

[0062] In a third aspect, an embodiment of the present application provides an electronic device, including:

[0063] a memory configured to store instructions;

[0064] The processor is configured to call the instructions from the memory and to implement the complex scene sanitation robot obstacle avoidance method provided by the first aspect of the embodiment of the present application when executing the instructions.

[0065] On the fourth aspect, the present application provides a machine-readable storage medium on which instructions are stored, and the instructions are used to enable the machine to execute the above-mentioned complex scenario sanitation robot obstacle avoidance method.

[0066] In an embodiment of the present application, noise data is separated from environmental data within a preset range of the sanitation robot to obtain a multi-source preliminary signal, wherein the environmental data includes at least the speed and distance of dynamic obstacles and the distance of static obstacles within the preset range; the spatial coordinates and velocity vector of each dynamic obstacle are extracted from the multi-source preliminary signal, and the position and target motion trend of each dynamic obstacle, as well as the position of each static obstacle, are determined based on a spatial mapping algorithm; an obstacle avoidance sequence for the sanitation robot is determined based on the position and target motion trend of each dynamic obstacle, as well as the position of each static obstacle; and key trigger points for the sanitation robot's obstacle avoidance, as well as feature vectors of the key trigger points, are extracted from the obstacle avoidance sequence. Based on the A-star algorithm, the system determines whether the sanitation robot has a potential deadlock area based on the characteristic vectors of key trigger points and the robot's preset operating duration. In response to the presence of a potential deadlock area, the system adjusts the path weight based on the deadlock area to obtain an optimized planned path, where the path weight includes at least a dynamic obstacle weight and a deadlock risk weight. Adjustment parameters are extracted from the optimized planned path to determine the relative displacement trends of each dynamic obstacle and each static obstacle, where the adjustment parameters include the robot's speed. Based on the relative displacement trends, the obstacles are prioritized and a target obstacle avoidance path is determined. The robot's motion is then controlled based on the target obstacle avoidance path. In this way, by separating noise from environmental data and developing a collaborative obstacle avoidance strategy for dynamic and static obstacles, the accuracy of obstacle avoidance in complex scenarios can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 This is a flow chart of the obstacle avoidance method for a sanitation robot in complex scenarios provided by an embodiment of the present application;

[0068] Figure 2 This is a trajectory diagram of a sanitation robot avoiding obstacles in complex scenarios provided by a specific embodiment of the present application;

[0069] Figure 3 This is a schematic diagram of the structure of the obstacle avoidance device for the sanitation robot in complex scenarios provided by an embodiment of the present application;

[0070] Figure 4 It is a structural diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0071] The following will be combined with the accompanying drawings in the embodiments of the present application to clearly describe the technical solutions in the embodiments of the present application. Obviously, the embodiments described are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field are within the scope of protection of this application.

[0072] The terms "first," "second," and the like in the specification and claims of this application are used to distinguish similar objects, and are not used to describe a specific order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments of this application can be implemented in an order other than that illustrated or described herein, and that the objects distinguished by "first," "second," and the like are generally of the same type, and do not limit the number of objects; for example, the first object can be one or more. In addition, the term "and / or" in the specification and claims represents at least one of the connected objects, and the character " / " generally indicates that the objects associated with each other are in an "or" relationship.

[0073] Below, in conjunction with the accompanying drawings, the complex scene sanitation robot obstacle avoidance method, device, equipment and medium provided by the embodiments of the present application are described in detail through specific embodiments and their application scenarios.

[0074] See Figure 1 , is a flow chart of a complex scene sanitation robot obstacle avoidance method provided by an embodiment of the present application, which is applied to electronic equipment. Figure 1 As shown, the complex scene sanitation robot obstacle avoidance method includes the following steps S100 to S900.

[0075] Step S100: Separating noise data from environmental data within a preset range of the sanitation robot to obtain a multi-source preliminary signal, wherein the environmental data at least includes the speed and distance of dynamic obstacles and the distance of static obstacles within the preset range.

[0076] In the embodiments of the present application, sanitation robots may include but are not limited to unmanned cleaning robots and manned cleaning robots. The preset range of the sanitation robot can be understood as the pre-set data collection range of the current location of the sanitation robot, for example, it can be within 5 meters from the sanitation robot. Environmental data can be understood as multi-source data collected by sensors such as lidar and ultrasonic sensors. Sanitation data may include but is not limited to data such as the speed and distance of dynamic obstacles and the distance of static obstacles within a preset range. Dynamic obstacles can be understood as obstacles that can move, such as pedestrians and vehicles. Static obstacles can be understood as obstacles that remain stationary, such as road posts.

[0077] After collecting the environmental data within the preset range of the sanitation robot, the environmental data within the preset range of the sanitation robot can be filtered using an adaptive weighted average algorithm to separate the noise data to obtain a multi-source preliminary signal with noise reduction.

[0078] Step S200: extracting the spatial coordinates and velocity vector of each of the dynamic obstacles from the multi-source preliminary signals, and determining the position and target motion trend of each of the dynamic obstacles, as well as the position of each of the static obstacles based on a spatial mapping algorithm.

[0079] In embodiments of the present application, spatial coordinates and velocity vectors can be obtained by decomposing multi-source preliminary signals. In one example, the relative velocity of a dynamic obstacle is calculated by analyzing the frequency changes of the multi-source preliminary signals. The velocity vector of the dynamic obstacle can be estimated based on the change in the position of the dynamic obstacle over time. In another example, the spatial coordinates of a dynamic obstacle can be estimated using geometric positioning or the Doppler effect based on information such as the strength, time difference, or frequency offset of the multi-source preliminary signals.

[0080] After obtaining the spatial coordinates and velocity vectors of each dynamic obstacle, a grid-based spatial mapping algorithm can be used to construct a mapping structure to obtain preliminary spatial distribution data. The positions of each static obstacle and the positions of each dynamic obstacle are extracted from this initial spatial distribution data. The target motion trend of each dynamic obstacle is determined based on its velocity vector.

[0081] Step S300: determining an obstacle avoidance sequence for the sanitation robot according to the position and target motion trend of each of the dynamic obstacles, and the position of each of the static obstacles.

[0082] In an embodiment of the present application, after determining the position and target motion trend of each dynamic obstacle, the overlapping region of each dynamic obstacle can be determined based on the position and target motion trend of each dynamic obstacle using a time window sliding method. If the overlapping region includes a static obstacle, the priority index of each dynamic obstacle can be determined using a vector decomposition method. Based on the priority index of each dynamic obstacle, the dynamic obstacles are then sorted to determine the obstacle avoidance sequence of the sanitation robot.

[0083] Step S400: extracting key trigger points for the sanitation robot to avoid obstacles and feature vectors of the key trigger points from the obstacle avoidance sequence.

[0084] In an embodiment of the present application, the obstacle avoidance trigger points of the sanitation robot are extracted from the obstacle avoidance sequence, and a preset lightweight convolutional network is called to extract event features to obtain an initial feature set. Among them, time features may include but are not limited to time features and spatial features (coordinates). The initial feature set is then compressed using a sparse matrix to reduce computational latency and determine a streamlined feature set. For the trigger points in the streamlined feature set, it is determined in the embedded environment whether the resource constraints exceed a preset threshold to obtain constraint conditions. Among them, resource constraints may include but are not limited to device processor performance and memory. If the threshold is exceeded, the sparse matrix structure is adjusted using compression technology to obtain an optimized feature set. Based on the correspondence between the optimized feature set and the trigger points, a distribution sequence of feature vectors is generated. Key trigger points are extracted from the distribution sequence, and the output of the lightweight convolutional network is fused to determine the feature vectors of the key trigger points.

[0085] Step S500: Based on the A-star algorithm, determine whether the sanitation robot has a potential deadlock area according to the feature vector of the key trigger point and the preset operation time of the sanitation robot.

[0086] In this embodiment of the present application, the characteristic vectors of the key trigger points are combined with the preset operation duration of the sanitation robot, and the A-star algorithm is used to evaluate the path deadlock risk to obtain a risk distribution. The potential deadlock area of ​​the sanitation robot is then extracted from the risk distribution.

[0087] Step S600: In response to the existence of a potential deadlock area of ​​the sanitation robot, the path weight is adjusted according to the potential deadlock area to obtain an optimized planned path, wherein the path weight includes at least a dynamic obstacle weight and a deadlock risk weight.

[0088] In an embodiment of the present application, when a sanitation robot has a potential deadlock area, the boundary of the potential deadlock area is detected to obtain a boundary division. Based on the boundary division, a dynamic adjustment technology is used to update the path weight to obtain a weight distribution. The change trend of the local path is judged by the weight distribution, and an adjusted path set is obtained. For the key trigger points in the path set, combined with a multiple parallel feature fusion strategy, an optimized path sequence can be determined. Among them, the path weight may include but is not limited to the path cost (distance), dynamic obstacle weight, and deadlock risk weight, etc., and the initial path sequence is finally determined by adjusting different weight parameters. The key area under the time constraint is then extracted from the initial path sequence, and a preset threshold is used for judgment to obtain the optimized planned path.

[0089] Step S700: extracting adjustment parameters from the optimized planned path to determine the relative displacement trend between each of the dynamic obstacles and each of the static obstacles, wherein the adjustment parameters include the speed of the sanitation robot.

[0090] In an embodiment of the present application, adjustment parameters are extracted from the optimized planned path, wherein the adjustment parameters include the speed of the sanitation robot, and the speed of the sanitation robot may include the angular velocity of the sanitation robot.

[0091] To account for the changes in dynamic obstacle interference and static obstacle distribution, a Kalman filter is used to fuse multi-source data to obtain a preliminary displacement distribution of dynamic obstacles. Based on this preliminary displacement distribution, the overlapping impact area of ​​dynamic and static obstacles is extracted. Based on the relative positions of dynamic and static obstacles, a preset threshold is used to determine the boundary range of dynamic interference.

[0092] Path planning parameters are updated based on the boundaries of dynamic obstacles. Based on the changing trends of overlapping influences, an adjusted path weight distribution is obtained. Key nodes along the path are extracted based on this weight distribution, and an optimized path sequence is derived by combining real-time adjustment information from multi-source data. The optimized path sequence is used to determine the motion trends of dynamic obstacles. Based on the static constraints of the static obstacle distribution, a Kalman filter is used to update the relative displacement trends of each dynamic obstacle and each static obstacle.

[0093] In the embodiment of the present application, the intensity distribution of dynamic obstacles can be extracted according to the relative displacement trend, and combined with the adjustment parameters of the path to obtain the final path optimization solution.

[0094] Step S800: Prioritizing obstacles according to the relative displacement trend and determining a target obstacle avoidance path;

[0095] In an embodiment of the present application, after obtaining the relative displacement trends of each dynamic obstacle and each static obstacle, a filtering method is used for smoothing to obtain a stable displacement trend distribution. The dynamic obstacle characteristics are extracted based on the stable displacement trend distribution. If the speed exceeds the preset speed threshold, the dynamic obstacle is marked as a high-priority target. In one example, if the speed of the dynamic obstacle exceeds 1 meter per second, the dynamic obstacle is marked as a high-priority target. The adjustment requirement range is determined based on each high-priority target. By adjusting the requirement range in combination with the preset threshold, each obstacle is prioritized to obtain a target obstacle avoidance sequence, and then the target obstacle avoidance path can be determined.

[0096] In the embodiment of the present application, after determining the target obstacle avoidance sequence, dynamic obstacles with excessive speed can be extracted according to the target obstacle avoidance sequence, and the parameters are adjusted to calculate the path deflection angle to obtain the deflection angle adjustment value. The deflection angle adjustment value can be expressed as:

[0097]

[0098] in, Indicates the path angle adjustment value, represents the deflection adjustment coefficient, and Indicates the target position coordinates, and Indicates the current position coordinates, represents the speed correction factor, represents the target speed, Indicates the speed threshold.

[0099] In this embodiment, the path plan is updated using the deflection adjustment value, generating a decision instruction sequence to obtain a preliminary obstacle avoidance plan. This preliminary obstacle avoidance plan is then integrated with trend data, and a Kalman filter is used to optimize the adjustment parameters and determine the final decision instruction. The final decision instruction is then matched to the dynamic target position, generating a real-time path adjustment sequence and determining the target obstacle avoidance path.

[0100] Step S900: Control the movement of the sanitation robot based on the target obstacle avoidance path.

[0101] Through the above steps S100 to S900, in the embodiment of the present application, the noise data is separated from the environmental data within the preset range of the sanitation robot to obtain a multi-source preliminary signal, wherein the environmental data at least includes the speed and distance of the dynamic obstacles and the distance of the static obstacles within the preset range; the spatial coordinates and velocity vector of each dynamic obstacle are extracted from the multi-source preliminary signal, and the position and target motion trend of each dynamic obstacle, as well as the position of each static obstacle, are determined based on the spatial mapping algorithm; the obstacle avoidance sequence of the sanitation robot is determined according to the position and target motion trend of each dynamic obstacle, as well as the position of each static obstacle; the key triggering point of the sanitation robot's obstacle avoidance, as well as the key triggering point of the obstacle avoidance are extracted from the obstacle avoidance sequence. The system uses the A-star algorithm to determine whether the sanitation robot has a potential deadlock area based on the characteristic vector of the key trigger point and the preset operation duration of the sanitation robot. In response to the presence of a potential deadlock area, the system adjusts the path weight based on the potential deadlock area to obtain an optimized planned path, where the path weight includes at least a dynamic obstacle weight and a deadlock risk weight. Adjustment parameters are extracted from the optimized planned path to determine the relative displacement trend between each dynamic obstacle and each static obstacle, where the adjustment parameters include the sanitation robot's speed. Based on the relative displacement trend, the obstacles are prioritized to determine the target obstacle avoidance path. The sanitation robot's motion is controlled based on the target obstacle avoidance path. In this way, by separating noise data from environmental data and formulating a collaborative obstacle avoidance strategy for dynamic and static obstacles, the accuracy of obstacle avoidance in complex scenarios can be improved.

[0102] In some embodiments, separating noise data from environmental data within a preset range of the sanitation robot to obtain a multi-source preliminary signal includes:

[0103] Dividing the environmental data based on multiple independent signal channels to obtain multiple channel data;

[0104] Performing filtering processing on the plurality of sub-channel data by using a parallel filter to obtain a filtered signal;

[0105] The noise data of the filtered signal is separated by an adaptive weighted average algorithm to obtain a multi-source preliminary signal.

[0106] Specifically, the environmental data can be divided into multiple channels to form multiple independent signal channels, and multiple sub-channel data can be obtained. Among them, multiple channels can be understood as one acquisition device corresponding to one channel. The multiple sub-channel data are then processed by parallel filters, and the preset filter parameters are used to separate high-frequency interference to obtain filtered signals. For example, the displacement change of 0.5 meters / second during cleaning operations is tracked in real time. The noise data is then separated from the filtered signal by an adaptive weighted average algorithm. Specifically, for the filtered signal, the displacement changes of dynamic obstacles at adjacent time points are calculated, and the displacement change trend is determined by time series analysis. The displacement change trend is processed using an adaptive weighted average algorithm, and the weight is dynamically adjusted according to historical data to obtain a smooth displacement signal. If there are still abnormal fluctuations in the smooth displacement signal, the noise residue is determined by local mean detection to obtain an optimized signal. The multi-channel data is fused according to the optimized signal to obtain a multi-source preliminary signal with noise reduction.

[0107] In this embodiment of the present application, displacement tracking results can be determined under low-speed cruising conditions based on multi-source preliminary signals. By updating the multi-source preliminary signals in real time and adjusting the filter parameters in combination with real-time collected environmental data, a continuous tracking signal of dynamic obstacle displacement can be obtained.

[0108] In some embodiments, extracting the spatial coordinates and velocity vector of each of the dynamic obstacles from the multi-source preliminary signals, and determining the position and target motion trend of each of the dynamic obstacles, as well as the position of each of the static obstacles based on a spatial mapping algorithm, includes:

[0109] Decomposing the multi-source preliminary signals to obtain the spatial coordinates and velocity vector of each of the dynamic obstacles;

[0110] By using a grid division method, a mapping structure is constructed based on the spatial coordinates and velocity vectors of each of the dynamic obstacles to obtain initial spatial distribution data;

[0111] Extracting the position corresponding to each of the static obstacles and the position of each of the dynamic obstacles from the initial spatial distribution data;

[0112] The target motion trend of each of the dynamic obstacles is determined according to the velocity vector.

[0113] Specifically, by decomposing the multi-source preliminary signals, the spatial coordinates and velocity vector of each dynamic obstacle can be obtained. In one example, the relative velocity of a dynamic obstacle is calculated by analyzing the frequency changes of the multi-source preliminary signals. The velocity vector of the dynamic obstacle can be estimated based on the change in the dynamic obstacle's position over time. In another example, the spatial coordinates of a dynamic obstacle can be estimated using geometric positioning or the Doppler effect based on information such as the strength, time difference, or frequency offset of the multi-source preliminary signals.

[0114] Specifically, grid partitioning methods divide a spatial region into grid cells of equal or unequal sizes, performing data analysis or calculations within each grid cell, ultimately mapping the entire spatial region. Each grid cell may contain data from different sensors or results calculated according to specific rules. Through appropriate grid partitioning, effective spatial analysis and processing can be achieved.

[0115] After obtaining the spatial coordinates and velocity vectors of each dynamic obstacle, we can use a grid-based spatial mapping algorithm to construct a mapping structure and obtain preliminary spatial distribution data. From this initial spatial distribution data, we can extract the corresponding position distribution of each static obstacle. Furthermore, we can analyze the target motion trend of each dynamic obstacle using its velocity vector.

[0116] In some embodiments, determining the target motion trend of each of the dynamic obstacles based on the velocity vector includes:

[0117] determining an initial motion trend of each of the dynamic obstacles according to the velocity vector;

[0118] determining an overlapping area between each of the dynamic obstacles and each of the static obstacles based on an initial movement trend of each of the dynamic obstacles;

[0119] For the mapping structure within the overlapping area, separate the static obstacles from the dynamic obstacles by adjusting the grid to obtain an optimized obstacle point distribution, wherein the obstacle points include the static obstacles and the dynamic obstacles;

[0120] According to the optimized obstacle point distribution, the velocity vectors are integrated to determine the target motion trend of each dynamic obstacle point;

[0121] The optimized obstacle point distribution is:

[0122]

[0123] in, represents the obstacle distribution density function, Indicates the number of sampling points, represents the weight coefficient, represents the attenuation coefficient, Indicates the reference point position, Indicates the location of the obstacle point.

[0124] Specifically, the velocity vectors of each dynamic obstacle can be used to determine the initial motion trend of each dynamic obstacle. Based on the initial motion trend of each dynamic obstacle and the position of each static obstacle, the overlapping area of ​​each dynamic obstacle and each static obstacle can be determined. For the mapping structure within the overlapping area, a local grid adjustment method can be used to separate the static and dynamic obstacles, resulting in an optimized obstacle point distribution, namely the static and dynamic obstacle distribution.

[0125] Specifically, based on the optimized obstacle point distribution and integrating the velocity vectors of dynamic obstacles, the target motion trend of each dynamic obstacle can be calculated. The velocity vector of the dynamic obstacle can be expressed as:

[0126]

[0127] in, represents the velocity vector of the obstacle, represents the spatial position vector, 、 、 Represent the velocity components in the x, y and z directions respectively.

[0128] The real-time position coordinates of dynamic obstacles can be expressed as:

[0129]

[0130] in, Indicates the real-time position coordinates of the obstacle, 、 and represents the initial coordinates, Indicates the coordinate correction amount.

[0131] Specifically, the optimized obstacle point distribution can be detected by the obstacle detection function. The obstacle detection function can be expressed as:

[0132]

[0133] in, represents the obstacle detection function, represents the detection threshold, represents the detection weight, Indicates the number of detection points, represents the spatial mapping function, Represents the coordinates of the detection point.

[0134] In this embodiment, the dynamic change direction of each dynamic obstacle is determined based on its target motion trend. If the dynamic change direction exceeds a preset threshold, the gridding parameters are adjusted based on historical data comparison to generate an updated mapping structure. The spatial coordinate change sequence of the dynamic obstacle is then extracted from this updated mapping structure. Multi-source signals are then integrated using a weighted fusion method to determine the final motion trend distribution. By updating the velocity vectors and spatial coordinates using real-time multi-source signals and combining them with the adjusted mapping structure, the continuously tracked obstacle point location distribution can be obtained.

[0135] In some embodiments, determining an obstacle avoidance sequence for the sanitation robot based on the position and target motion trend of each of the dynamic obstacles, and the position of each of the static obstacles, includes:

[0136] Based on the time window sliding method, determining the overlapping area of ​​each dynamic obstacle according to the position of each dynamic obstacle and the target movement trend;

[0137] In the case where the overlapping area contains static obstacles, determining the priority index of each dynamic obstacle by a vector decomposition method;

[0138] Based on the priority index of each of the dynamic obstacles, the dynamic obstacles are sorted to determine the obstacle avoidance sequence of the sanitation robot;

[0139] The priority index is:

[0140]

[0141] in, Represents the priority index, represents the priority coefficient, represents the weight factor, Indicates the obstacle distance, represents the scale parameter, Indicates the number of obstacles.

[0142] Specifically, after generating the positions and target motion trends of dynamic obstacles using a gridding method, the time window sliding method can be used to determine the overlapping regions of each dynamic obstacle based on the positions and target motion trends of each dynamic obstacle. A determination is then made as to whether a static obstacle exists within the overlapping region. If a static obstacle is included in the overlapping region, a vector decomposition method can be used to determine the priority index of each dynamic obstacle. The priority indexes of each dynamic obstacle are then sorted in descending order to determine the obstacle avoidance sequence for the sanitation robot.

[0143] In this embodiment, the sanitation robot's obstacle avoidance sequence determines the direction of motion trend adjustment. Based on the adjusted motion trend, the grid structure is updated to obtain optimized position data. By integrating this optimized position data with the obstacle avoidance sequence and using a time window sliding method, a corrected trajectory for dynamic obstacles is obtained. Based on the corrected trajectory and the distribution of static obstacles, the change in motion trend after the dynamic obstacle is separated is determined, and the sanitation robot's obstacle avoidance sequence can be updated accordingly.

[0144] In some embodiments, extracting adjustment parameters from the optimized planned path and determining the relative displacement trend of each of the dynamic obstacles and each of the static obstacles includes:

[0145] Extracting adjustment parameters from the optimized planned path, wherein the adjustment parameters include the speed of the sanitation robot;

[0146] According to the distribution changes of the dynamic obstacles and the static obstacles, a Kalman filter is used to fuse the environmental data to obtain the preliminary displacement distribution of each dynamic obstacle;

[0147] extracting the overlapping area of ​​the dynamic obstacle and the static obstacle according to the preliminary displacement distribution, and determining the boundary range of each dynamic obstacle;

[0148] Update the path planning parameters according to the boundary range to obtain an optimized path sequence;

[0149] The relative displacement trends of the dynamic obstacles and the static obstacles are determined according to the optimized path sequence.

[0150] Specifically, in an embodiment of the present application, adjustment parameters are extracted from the optimized planned path, including the sanitation robot's speed. A Kalman filter is used to fuse multi-source environmental data to determine the initial displacement distribution of dynamic obstacles, considering the changes in dynamic obstacle interference and static obstacle distribution. Based on this initial displacement distribution, the overlapping impact area of ​​dynamic and static obstacles is extracted. A preset threshold is used to determine the boundary range of dynamic interference based on the relative positions of dynamic and static obstacles.

[0151] Path planning parameters are updated based on the boundaries of dynamic obstacles. Based on the changing trends of overlapping influences, an adjusted path weight distribution is obtained. Key nodes along the path are extracted based on this weight distribution, and an optimized path sequence is derived by combining real-time adjustment information from multi-source data. The optimized path sequence is used to determine the motion trends of dynamic obstacles. Based on the static constraints of the static obstacle distribution, a Kalman filter is used to update the relative displacement trends of each dynamic obstacle and each static obstacle.

[0152] In some embodiments, prioritizing obstacles based on the relative displacement trend and determining a target obstacle avoidance path includes:

[0153] Performing filtering and smoothing processing on the relative displacement trend to obtain a stable displacement trend distribution;

[0154] extracting the velocity of each of the dynamic obstacles according to the stable displacement trend distribution;

[0155] Marking the dynamic obstacle whose speed exceeds a preset speed threshold as a high priority obstacle;

[0156] For each of the high priority obstacles, determining an absolute difference between a speed and the preset speed threshold;

[0157] Sort the absolute differences corresponding to the high-priority obstacles in descending order to obtain a target sorting sequence;

[0158] Determining a path angle adjustment value for each of the dynamic obstacles based on the target sorting sequence;

[0159] A target obstacle avoidance path is determined according to the path deflection angle adjustment value.

[0160] Specifically, in an embodiment of the present application, after obtaining the relative displacement trend of each dynamic obstacle and each static obstacle, a filtering method is used for smoothing to obtain a stable displacement trend distribution. The dynamic obstacle characteristics are extracted based on the stable displacement trend distribution. If the speed exceeds the preset speed threshold, the dynamic obstacle is marked as a high-priority target. In one example, if the speed of the dynamic obstacle exceeds 1 meter per second, the dynamic obstacle is marked as a high-priority target. The adjustment requirement range is determined based on each high-priority target. By adjusting the requirement range in combination with the preset threshold, each obstacle is prioritized to obtain a target obstacle avoidance sequence, and then the target obstacle avoidance path can be determined.

[0161] In the embodiment of the present application, after determining the target obstacle avoidance sequence, dynamic obstacles with excessive speed can be extracted according to the target obstacle avoidance sequence, and the parameters are adjusted to calculate the path angle to obtain the angle adjustment value. The target speed can be expressed as:

[0162]

[0163] in, represents the target speed, M represents the number of observation data points, Indicates the current location, Indicates the position at the previous moment, represents the time interval. This formula can be used to calculate the average speed of a dynamic obstacle;

[0164] The deflection adjustment value can be expressed as:

[0165]

[0166] in, Indicates the path angle adjustment value, represents the deflection adjustment coefficient, and Indicates the target position coordinates, and Indicates the current position coordinates, represents the speed correction factor, represents the target speed, Indicates the speed threshold.

[0167] In this embodiment, the path plan is updated using the deflection adjustment value, generating a decision instruction sequence to obtain a preliminary obstacle avoidance plan. This preliminary obstacle avoidance plan is then integrated with trend data, and a Kalman filter is used to optimize the adjustment parameters and determine the final decision instruction. The final decision instruction is then matched to the dynamic target position, generating a real-time path adjustment sequence and determining the target obstacle avoidance path.

[0168] See Figure 2 , is a trajectory diagram of a sanitation robot avoiding obstacles in a complex scene provided by a specific embodiment of the present application, such as Figure 2 As shown in the figure, squares represent static obstacles, circles represent dynamic obstacles, and the line connecting the starting point and the end point represents the obstacle avoidance path of the sanitation robot. The position of dynamic obstacles changes in real time, and the sanitation robot avoids them by adjusting the path angle.

[0169] See Figure 3 , is a structural diagram of a complex scene sanitation robot obstacle avoidance device provided by an embodiment of the present application. A second aspect of an embodiment of the present application provides a complex scene sanitation robot obstacle avoidance device 30, and the complex scene sanitation robot obstacle avoidance device 30 includes:

[0170] A separation module 31 is configured to separate noise data from environmental data within a preset range of the sanitation robot to obtain a multi-source preliminary signal, wherein the environmental data includes at least the speed and distance of dynamic obstacles and the distance of static obstacles within the preset range;

[0171] a first extraction module 32 configured to extract the spatial coordinates and velocity vector of each of the dynamic obstacles from the multi-source preliminary signals, and determine the position and target motion trend of each of the dynamic obstacles, as well as the position of each of the static obstacles based on a spatial mapping algorithm;

[0172] A first determining module 33 is configured to determine an obstacle avoidance sequence for the sanitation robot according to the position and target motion trend of each of the dynamic obstacles, and the position of each of the static obstacles;

[0173] A second extraction module 34 is configured to extract key triggering points for the sanitation robot to avoid obstacles and feature vectors of the key triggering points from the obstacle avoidance sequence;

[0174] A judgment module 35 is configured to determine whether the sanitation robot has a potential deadlock area based on the A-star algorithm according to the characteristic vector of the key trigger point and the preset operation duration of the sanitation robot;

[0175] an adjustment module 36 for adjusting a path weight according to a potential deadlock area of ​​the sanitation robot to obtain an optimized planned path, wherein the path weight includes at least a dynamic obstacle weight and a deadlock risk weight;

[0176] a second determining module 37, configured to extract adjustment parameters from the optimized planned path to determine a relative displacement trend between each of the dynamic obstacles and each of the static obstacles, wherein the adjustment parameters include a speed of the sanitation robot;

[0177] A third determination module 38 is configured to prioritize the obstacles according to the relative displacement trend and determine a target obstacle avoidance path;

[0178] The control module 39 is used to control the movement of the sanitation robot based on the target obstacle avoidance path.

[0179] The complex scene sanitation robot obstacle avoidance device 30 provided in the second aspect of the embodiment of the present application can implement the various processes implemented in the above method embodiment and achieve the same beneficial effects. To avoid repetition, it will not be repeated here.

[0180] See Figure 4 , is a structural diagram of an electronic device provided in an embodiment of the present application. The third aspect of an embodiment of the present application provides an electronic device 4000, including a processor 4100 and a memory 4200. The memory 4200 stores machine-executable instructions that can be executed by the processor 4100. The processor 4100 can execute the machine-executable instructions to implement the above-mentioned complex scene sanitation robot obstacle avoidance method.

[0181] The fourth aspect of an embodiment of the present application provides a machine-readable storage medium, on which instructions are stored. When the instructions are executed by a processor, the processor implements the above-mentioned complex scene sanitation robot obstacle avoidance method.

[0182] In some embodiments, the embodiments of the present application also provide a computer program product, including a computer program, which, when executed by a processor, implements the complex scene sanitation robot obstacle avoidance method according to the above-mentioned embodiment.

[0183] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0184] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 These computer program instructions can also be stored in a computer-readable memory that can guide a computer or other programmable data processing device to work in a specific way, so that the instructions stored in the computer-readable memory produce a product including the instruction device, which implements the function specified in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0185] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0186] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0187] Computer-readable media includes permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0188] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0189] The above are merely embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

[0190] In addition, the various embodiments of the present invention may be arbitrarily combined, and as long as they do not violate the concept of the present invention, they should also be regarded as the contents disclosed by the present invention.

Claims

1. A complex scene sanitation robot obstacle avoidance method, characterized in that: The method comprises: Separating noise data from environmental data within a preset range of the sanitation robot to obtain a multi-source preliminary signal, wherein the environmental data includes at least the speed and distance of dynamic obstacles and the distance of static obstacles within the preset range; Extracting the spatial coordinates and velocity vector of each of the dynamic obstacles from the multi-source preliminary signals, and determining the position and target motion trend of each of the dynamic obstacles, as well as the position of each of the static obstacles based on a spatial mapping algorithm; Determining an obstacle avoidance sequence for the sanitation robot according to the position and target motion trend of each of the dynamic obstacles, and the position of each of the static obstacles; Extracting key triggering points for obstacle avoidance of the sanitation robot and feature vectors of the key triggering points from the obstacle avoidance sequence; Based on the A-star algorithm, according to the characteristic vector of the key trigger point and the preset operation time of the sanitation robot, it is determined whether the sanitation robot has a potential deadlock area; In response to the presence of a deadlock potential area for the sanitation robot, adjusting a path weight according to the deadlock potential area to obtain an optimized planned path, wherein the path weight includes at least a dynamic obstacle weight and a deadlock risk weight; Extracting adjustment parameters from the optimized planned path to determine the relative displacement trend of each of the dynamic obstacles and each of the static obstacles, wherein the adjustment parameters include at least the speed and angular velocity of the sanitation robot; According to the relative displacement trend, the obstacles are prioritized and a target obstacle avoidance path is determined; Controlling the movement of the sanitation robot based on the target obstacle avoidance path; The step of extracting the key triggering points for the sanitation robot to avoid obstacles from the obstacle avoidance sequence includes: Extracting the obstacle avoidance trigger point of the sanitation robot from the obstacle avoidance sequence; Extracting event features of each obstacle avoidance trigger point according to a preset lightweight convolutional network to obtain an initial feature set; The initial feature set is compressed using a sparse matrix, and when resource constraints in the embedded environment exceed a preset threshold, the sparse matrix structure is adjusted using compression technology to obtain an optimized feature set; Matching the optimized feature set with each of the obstacle avoidance trigger points to obtain a distribution sequence of feature vectors; The key triggering points for the sanitation robot to avoid obstacles are determined according to the distribution sequence.

2. The method according to claim 1, characterized in that The method of separating noise data from environmental data within a preset range of the sanitation robot to obtain a multi-source preliminary signal includes: Dividing the environmental data based on multiple independent signal channels to obtain multiple channel data; Performing filtering processing on the plurality of sub-channel data by using a parallel filter to obtain a filtered signal; The noise data of the filtered signal is separated by an adaptive weighted average algorithm to obtain a multi-source preliminary signal.

3. The method according to claim 1, characterized in that Extracting the spatial coordinates and velocity vector of each of the dynamic obstacles from the multi-source preliminary signals, and determining the position and target motion trend of each of the dynamic obstacles, as well as the position of each of the static obstacles based on a spatial mapping algorithm, includes: Decomposing the multi-source preliminary signals to obtain the spatial coordinates and velocity vector of each of the dynamic obstacles; By using a grid division method, a mapping structure is constructed based on the spatial coordinates and velocity vectors of each of the dynamic obstacles to obtain initial spatial distribution data; Extracting the position corresponding to each of the static obstacles and the position of each of the dynamic obstacles from the initial spatial distribution data; The target motion trend of each of the dynamic obstacles is determined according to the velocity vector.

4. The method according to claim 3, characterized in that Determining the target motion trend of each of the dynamic obstacles according to the velocity vector includes: determining an initial motion trend of each of the dynamic obstacles according to the velocity vector; determining an overlapping area between each of the dynamic obstacles and each of the static obstacles based on an initial movement trend of each of the dynamic obstacles; For the mapping structure within the overlapping area, separate the static obstacles from the dynamic obstacles by adjusting the grid to obtain an optimized obstacle point distribution, wherein the obstacle points include the static obstacles and the dynamic obstacles; According to the optimized obstacle point distribution, the velocity vectors are integrated to determine the target motion trend of each dynamic obstacle point; The optimized obstacle point distribution is: in, represents the obstacle distribution density function, Indicates the number of sampling points, represents the weight coefficient, represents the attenuation coefficient, Indicates the reference point position, Indicates the location of the obstacle point.

5. The method according to claim 1, wherein The determining of the obstacle avoidance sequence of the sanitation robot according to the position and target motion trend of each of the dynamic obstacles and the position of each of the static obstacles includes: Based on the time window sliding method, determining the overlapping area of ​​each dynamic obstacle according to the position of each dynamic obstacle and the target movement trend; In the case where the overlapping area contains static obstacles, determining the priority index of each dynamic obstacle by a vector decomposition method; Based on the priority index of each of the dynamic obstacles, the dynamic obstacles are sorted to determine the obstacle avoidance sequence of the sanitation robot; The priority index is: in, Represents the priority index, represents the priority coefficient, represents the weight factor, Indicates the obstacle distance, represents the scale parameter, Indicates the number of obstacles.

6. The method according to claim 1, characterized in that Extracting adjustment parameters from the optimized planned path to determine the relative displacement trend of each of the dynamic obstacles and each of the static obstacles includes: Extracting adjustment parameters from the optimized planned path, wherein the adjustment parameters include at least the speed and angular velocity of the sanitation robot; According to the distribution changes of the dynamic obstacles and the static obstacles, a Kalman filter is used to fuse the environmental data to obtain the preliminary displacement distribution of each dynamic obstacle; Extracting the overlapping area of ​​the dynamic obstacle and the static obstacle according to the preliminary displacement distribution, and determining the boundary range of each dynamic obstacle; Update the path planning parameters according to the boundary range to obtain an optimized path sequence; The relative displacement trends of the dynamic obstacles and the static obstacles are determined according to the optimized path sequence.

7. The method according to claim 1, characterized in that Prioritizing the obstacles according to the relative displacement trend and determining a target obstacle avoidance path includes: Performing filtering and smoothing processing on the relative displacement trend to obtain a stable displacement trend distribution; extracting the velocity of each of the dynamic obstacles according to the stable displacement trend distribution; Marking the dynamic obstacle whose speed exceeds a preset speed threshold as a high priority obstacle; For each of the high priority obstacles, determining an absolute difference between a speed and the preset speed threshold; Sort the absolute differences corresponding to the high-priority obstacles in descending order to obtain a target sorting sequence; Determining a path angle adjustment value for each of the dynamic obstacles based on the target sorting sequence; A target obstacle avoidance path is determined according to the path deflection angle adjustment value.

8. An obstacle avoidance device for a sanitation robot in complex scenarios, characterized in that: The device comprises: a separation module, configured to separate noise data from environmental data within a preset range of the sanitation robot to obtain a multi-source preliminary signal, wherein the environmental data includes at least the speed and distance of dynamic obstacles and the distance of static obstacles within the preset range; a first extraction module, configured to extract the spatial coordinates and velocity vector of each of the dynamic obstacles from the multi-source preliminary signals, and determine the position and target motion trend of each of the dynamic obstacles, as well as the position of each of the static obstacles based on a spatial mapping algorithm; A first determination module is configured to determine an obstacle avoidance sequence for the sanitation robot based on the position and target motion trend of each of the dynamic obstacles and the position of each of the static obstacles; A second extraction module is used to extract the key triggering points of the sanitation robot's obstacle avoidance and the feature vectors of the key triggering points from the obstacle avoidance sequence; a judgment module, configured to judge whether the sanitation robot has a deadlock potential area based on the A-star algorithm and the characteristic vector of the key trigger point and the preset operation duration of the sanitation robot; an adjustment module, configured to, in response to a deadlock potential area existing in the sanitation robot, adjust a path weight according to the deadlock potential area to obtain an optimized planned path, wherein the path weight includes at least a dynamic obstacle weight and a deadlock risk weight; a second determination module, configured to extract adjustment parameters from the optimized planned path and determine a relative displacement trend between each of the dynamic obstacles and each of the static obstacles, wherein the adjustment parameters include at least a speed and an angular velocity of the sanitation robot; a third determination module, configured to prioritize obstacles according to the relative displacement trend and determine a target obstacle avoidance path; A control module is used to control the movement of the sanitation robot based on the target obstacle avoidance path.

9. An electronic device, characterized in that: include: a memory configured to store instructions; The processor is configured to call the instructions from the memory and to implement the complex scene sanitation robot obstacle avoidance method according to any one of claims 1 to 7 when executing the instructions.

10. A machine-readable storage medium, characterized in that The machine-readable storage medium stores instructions for enabling a machine to execute the complex scenario sanitation robot obstacle avoidance method according to any one of claims 1 to 7.

Images