Robot home track self-adaptive generation method based on intelligent control

Abstract

The invention discloses a robot home track self-adaptive generation method based on intelligent control, and relates to the technical field of intelligent control, and the method comprises the steps: collecting environment control data; establishing a corresponding relation between a local coordinate system and a global occupation grid map coordinate system, generating an initial rectangular home area, and performing unilateral contraction and opposite-side compensation in combination with obstacle conflicts to obtain a course coupling home area; fusing wall shielding, gradient change and signal attenuation to execute overall offset correction on the course coupling home-returning area, and screening candidate entrance doors on the corrected boundary to obtain an optimal entrance door; constructing a multi-factor cost map, and generating a self-adaptive home-returning track in a segmented manner; and controlling the robot to go home based on the self-adaptive going-home track, and detecting pose deviation, dynamic obstacles and signal changes in real time until precise parking is completed. The stability and the parking accuracy of the robot in the home returning process in a complex environment are improved through integrated control from home returning area generation to precise parking of the robot.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology, and in particular to an adaptive method for generating a robot's home trajectory based on intelligent control. Background Technology

[0002] With the development of autonomous mobile devices such as yard service robots, cleaning robots, and lawnmower robots, the ability for robots to return to their docking stations and autonomously dock after completing their tasks has become an important research topic in the field of mobile robot control. Existing conventional methods typically generate a home route based on localization results, environmental maps, and preset docking point information, using a combination of fixed return routes, global path planning, or local obstacle avoidance control. Furthermore, they control the final docking process by considering the orientation of the charging station, obstacle distribution, and communication status, thereby enabling the robot to navigate, approach, and dock from the work area to the docking area.

[0003] However, the above-mentioned conventional methods still face two main adaptation problems in practical applications: First, home routes and parking areas are mostly generated according to fixed boundaries or static channels, and the consistency between the path direction and the end-entry direction is insufficient when there are wall obstructions, slope changes and irregular obstacles; Second, conventional path planning has weak consideration for communication reliability, remaining power and attitude adjustment costs, resulting in limited adaptability to dynamic environmental changes during trajectory execution. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides an adaptive generation method for robot home trajectory based on intelligent control to solve the problems of insufficient coupling between the path and the end-positioning direction during the existing robot home process and the limited adaptive adjustment capability of the trajectory under multiple constraints.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] This invention provides an adaptive method for generating a robot's home trajectory based on intelligent control. The method includes: collecting the robot's current position, docking point position, base station orientation, yard boundary, obstacle distribution, real-time signal distribution, and robot motion constraints to form environmental control data; establishing a correspondence between the local coordinate system and the global occupancy grid map coordinate system based on the environmental control data to generate an initial rectangular home region, and performing unilateral contraction and contralateral compensation in conjunction with obstacle conflicts to obtain a heading-coupled home region; performing overall offset correction on the heading-coupled home region by incorporating wall occupancy, slope changes, and signal attenuation, and selecting candidate entry gates on the corrected boundary to obtain the optimal entry gate; constructing a multi-factor cost map by incorporating obstacle risk, attitude adjustment cost, communication reliability, and remaining battery power based on the optimal entry gate, and generating an adaptive home trajectory in segments; controlling the robot to return home based on the adaptive home trajectory, detecting pose deviation, dynamic obstacles, and signal changes in real time until precise docking is achieved.

[0008] As a preferred embodiment of the robot home trajectory adaptive generation method based on intelligent control described in this invention, the formation of environmental control data refers to uniformly writing the robot's current position, docking point position, base station orientation, courtyard boundary, obstacle distribution, real-time signal distribution, and body motion constraints into a global occupancy grid map to form environmental control data.

[0009] As a preferred embodiment of the robot home trajectory adaptive generation method based on intelligent control described in this invention, the step of establishing a correspondence between the local coordinate system and the global occupied grid map coordinate system based on the environmental control data, and generating an initial rectangular home trajectory area, includes the following steps:

[0010] Based on environmental control data, a local coordinate system is established, and a rotation matrix is ​​used to establish the correspondence between the local coordinate system and the global occupied raster map coordinate system;

[0011] Based on the correspondence, as well as the courtyard boundary, obstacle distribution, real-time signal distribution, and body motion constraints, the main direction half-length constraint and the lateral half-width constraint are determined.

[0012] Based on the main direction half-length constraint and the lateral half-width constraint, the half-length and half-width of the initial rectangular home region are determined, and the size parameters are obtained.

[0013] Based on the size parameters, the four vertices in the local coordinate system are mapped to the global occupied raster map coordinate system using a rotation matrix, and then connected to form a closed rectangular boundary, thus obtaining the initial rectangular home area.

[0014] As a preferred embodiment of the robot home trajectory adaptive generation method based on intelligent control described in this invention, the step of combining obstacle conflict to perform unilateral contraction and contralateral compensation to obtain the heading-coupled home region includes the following steps:

[0015] The initial rectangular home area is discretized into multiple grid cells. Each grid cell is checked to see if it meets the passage conditions. Grid cells that do not meet the passage conditions are marked as conflict grids.

[0016] Based on the conflict grid, the rectangular edges of the set of conflict grids are identified as conflict edges, and the conflict depth of each conflict edge is calculated.

[0017] Based on the conflict edge and conflict depth, the conflict edge is gradually moved towards the stop point along a direction perpendicular to the conflict edge until the moved rectangular boundary no longer corresponds to the conflict grid, thus obtaining the middle home area;

[0018] Calculate the size difference between the intermediate home region and the initial rectangular home region, and expand outwards along the rectangular side opposite the conflict side at the same grid step size until the compensated rectangular boundary reaches the expansion limit or restores the preset size requirement to obtain the candidate home region;

[0019] Based on the candidate home region, the rectangular boundary after unilateral contraction and contralateral compensation is determined as the final boundary, thus obtaining the heading coupled home region.

[0020] As a preferred embodiment of the robot home trajectory adaptive generation method based on intelligent control described in this invention, the step of performing overall offset correction on the heading-coupled home region based on the heading-coupled home region, incorporating wall occlusion, slope changes, and signal attenuation, includes the following steps:

[0021] Extract wall distribution, slope distribution and signal strength distribution around the heading-coupled home area to determine the degree of imbalance in the heading-coupled home area in the directions to the sides and front and back of the base station, and determine the restricted direction;

[0022] Based on the restricted direction, the heading-coupled home area is gradually shifted away from the restricted direction until a corrected home area that meets the requirements of wall avoidance, slope adaptation, and signal reliability is obtained.

[0023] As a preferred embodiment of the robot home trajectory adaptive generation method based on intelligent control described in this invention, the step of filtering candidate entry gates on the corrected boundary to obtain the optimal entry gate includes the following steps:

[0024] Based on the corrected home area, the rectangular side of the base station's orientation and the two rectangular sides adjacent to the base station's orientation are used as candidate boundaries.

[0025] Divide the candidate boundary into multiple continuous boundary segments of a preset length, and determine the passage conditions of the adjacent grids corresponding to each continuous boundary segment to obtain the candidate entry gate;

[0026] Based on the candidate entry gates, the travel distance from the robot's current position to each candidate entry gate, the orientation adjustment amount from the candidate entry gate to the stopping point, the obstacle clearance in the neighborhood of the candidate entry gate, and the signal reliability in the neighborhood of the candidate entry gate are calculated and sorted to obtain the optimal entry gate.

[0027] As a preferred embodiment of the robot home trajectory adaptive generation method based on intelligent control described in this invention, the construction of the multi-factor cost map refers to determining the starting grid, the first target grid, and the second target grid based on the optimal entry gate and environmental control data, and assigning a cost value to the passable grids in the global occupied grid map based on obstacle risk, posture adjustment cost, communication reliability, and remaining power, thereby obtaining the multi-factor cost map.

[0028] As a preferred embodiment of the robot home trajectory adaptive generation method based on intelligent control described in this invention, the segmented generation of the adaptive home trajectory includes the following steps:

[0029] Based on the multi-factor cost map, the A× algorithm is used to search from the current position to the optimal entry gate to obtain the first trajectory segment.

[0030] Based on the first trajectory segment, the second trajectory segment is generated grid by grid within the heading coupled home area, and it is determined whether the second trajectory segment meets the passage conditions corresponding to the environmental control data, until the second trajectory segment that meets the passage conditions is obtained.

[0031] Based on the first trajectory segment and the second trajectory segment that meets the traffic conditions, the trajectory is spliced ​​together according to the driving sequence to obtain the adaptive home route.

[0032] As a preferred embodiment of the robot home trajectory adaptive generation method based on intelligent control described in this invention, the step of controlling the robot to return home based on the adaptive home trajectory, and detecting pose deviation, dynamic obstacles, and signal changes in real time until precise docking is achieved, includes the following steps:

[0033] Based on the adaptive home trajectory, current location, optimal entry door and stopping point location, determine the trajectory tracking sequence, entry switching area and stopping endpoint grid;

[0034] Based on the trajectory tracking sequence, the robot is controlled to move along an adaptive home trajectory, and after entering the neighborhood of the current target grid at the current position, it switches to the next target grid to obtain the trajectory tracking result;

[0035] Based on the trajectory tracking results, pose deviation, dynamic obstacles and signal changes are continuously detected to obtain the trajectory validity determination results;

[0036] Based on the trajectory validity determination result, the driving trajectory near the stop point is obtained or an adaptive home trajectory is regenerated until a precise stop is completed.

[0037] As a preferred embodiment of the intelligent control-based adaptive home trajectory generation method for robots described in this invention, the step of obtaining a driving trajectory close to the stopping point or regenerating an adaptive home trajectory based on the trajectory validity determination result means that when the trajectory is valid, the robot continues to run along the adaptive home trajectory and enters the stopping point along the subsequent trajectory of the optimal entry door; when the trajectory fails, the robot stops moving, re-selects the optimal entry door, reconstructs the multi-factor cost map, and regenerates the adaptive home trajectory.

[0038] The beneficial effects of this invention are as follows: By combining obstacle conflict to perform unilateral contraction and contralateral compensation, a heading-coupled homecoming area is obtained, realizing the directional correction of the homecoming area boundary with the courtyard boundary, obstacle distribution, real-time signal distribution and aircraft motion constraints, so that the area shape is consistent with the base station entry orientation, which facilitates providing stable spatial constraints for subsequent entry paths; by constructing a multi-factor cost map and generating adaptive homecoming trajectories in segments, the integrated organization of the homecoming path and the final docking process is realized, so that trajectory tracking, trajectory validity determination and replanning processes can be executed around the same multi-factor cost constraints, thereby improving the adaptability of path planning and the stability of accurate docking. Attached Figure Description

[0039] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 This is a flowchart of a method for adaptively generating a robot's home trajectory based on intelligent control.

[0041] Figure 2 The flowchart shows the offset correction and optimal door selection process. Detailed Implementation

[0042] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0043] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0044] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0045] Reference Figure 1 and Figure 2 This is one embodiment of the present invention, which provides an adaptive generation method for a robot's home trajectory based on intelligent control, comprising the following steps:

[0046] S1. Collect the robot's current position, docking point position, base station orientation, courtyard boundary, obstacle distribution, real-time signal distribution, and robot motion constraints to form environmental control data.

[0047] The robot receives the "return home" command and basic configuration parameters input or confirmed by the user via a mobile terminal. These parameters include at least the safe docking distance threshold and the preferred aspect ratio of the rectangular area. If the user does not input a preferred aspect ratio, the default value of 1:1 is used. After receiving the basic configuration parameters, the robot simultaneously reads the SLAM positioning results, LiDAR scan data, ultrasonic sensor ranging data, inertial measurement unit attitude data, and the remaining battery power from the battery management data.

[0048] Furthermore, the robot's current position is represented by the mapped coordinates of the SLAM localization results in the global occupancy grid map; the docking point position uses the pre-stored charging base station area in the global occupancy grid map as the initial search area, and uses LiDAR scan data to perform local matching on the outer contour of the charging base station, determining the midpoint of the charging base station's leading edge as the docking point position; the base station's orientation is represented by the opposite direction of the charging base station's leading edge normal. The current LiDAR scan data is projected onto the global occupancy grid map and compared segment by segment with the historically saved courtyard boundaries. Line segments consistent with historical boundaries are taken as valid courtyard boundaries, and new line segments that are stable and located inside the courtyard are added to the courtyard boundary to obtain the current courtyard boundary. The occupancy grid formed by LiDAR scan data is superimposed with the near-field occupancy area formed by ultrasonic sensor ranging data, and the continuously occupied grid areas that do not belong to the courtyard boundary are taken as obstacle distribution. Based on the measured RSSI value of the robot's current position, the docking point position, and the obstacle distribution, the real-time signal distribution is calculated. The real-time signal distribution uses the logarithmic distance path loss model to calculate the predicted signal strength of each grid in the courtyard to the docking point position, the expression of which is:

[0049] ;

[0050] in, Indicates the distance from the stop point. The predicted signal strength corresponding to the grid. Indicates reference distance Reference signal strength at that location, This represents the path loss index. This indicates the actual distance from the grid to the docking point. Indicates the reference distance. This indicates the occlusion loss caused by the distribution of obstacles on the propagation path.

[0051] Based on real-time signal distribution, pre-stored robot body width, body length, minimum turning radius, docking safety distance threshold, and inertial measurement unit attitude data, robot motion constraints are calculated. These constraints include at least passable width, minimum clearance constraint, and slope constraint. The docking safety distance threshold is determined jointly based on the robot's body width and length from the parameter table, the positioning error radius corresponding to the SLAM positioning results, and the grid edge length of the global occupancy grid map. The body width and length characterize the robot's outline range during docking and attitude adjustment; the positioning error radius compensates for the deviation between the current position and the actual position; and the grid edge length compensates for boundary errors caused by map discretization. The calculated results are compared with the minimum engineering safety value, and the larger value is taken as the docking safety distance threshold. Finally, using the global occupancy grid map as the spatial basis, the robot's current position, docking point position, base station orientation, courtyard boundary, obstacle distribution, real-time signal distribution, and robot motion constraints are associated, labeled, and uniformly organized to form environmental control data.

[0052] S2. Based on the environmental control data, establish the correspondence between the local coordinate system and the global occupied grid map coordinate system, generate an initial rectangular home area, and combine obstacle conflict to perform unilateral contraction and contralateral compensation to obtain the heading coupled home area.

[0053] S2.1. Based on environmental control data, establish a local coordinate system and use a rotation matrix to establish the correspondence between the local coordinate system and the global occupied raster map coordinate system;

[0054] Furthermore, based on environmental control data, the docking point location, base station orientation, courtyard boundary, obstacle distribution, real-time signal distribution, and machine motion constraints are read from the global occupancy grid map. The grid corresponding to the docking point location is determined as the center grid, and the base station orientation is determined as the main direction of the initial rectangular home area, generating a center reference and direction reference. Based on the center reference and direction reference, a local coordinate system for the initial rectangular home area is established with the docking point location as the origin, the base station orientation as the positive axis of the local coordinate system, and the direction perpendicular to the base station orientation as the lateral axis of the local coordinate system. A rotation matrix is ​​used to establish the correspondence between the local coordinate system and the global occupancy grid map coordinate system. The expression for the rotation matrix is:

[0055] ;

[0056] in, and Represents the coordinate values ​​in the local coordinate system. and This represents the coordinate value of the globally occupied raster map coordinate system. Indicates the base station's orientation angle. and This indicates the coordinates of the docking point in the global grid map coordinate system.

[0057] S2.2. Based on the correspondence, courtyard boundaries, obstacle distribution, real-time signal distribution, and body motion constraints, determine the main direction half-length constraint and the lateral half-width constraint;

[0058] Specifically, based on the correspondence between the local coordinate system and the global occupied grid map coordinate system, starting from the docking point, the system performs centrally symmetrical sampling along the positive and negative axes of the local coordinate system with a fixed grid step size. For each sampling grid, four conditions are sequentially determined: whether the sampling grid is located within the courtyard boundary, whether the sampling grid coincides with the obstacle distribution, whether the real-time signal distribution corresponding to the sampling grid meets the high-reliability signal gain range, and whether the robot's motion constraints corresponding to the sampling grid allow the robot to enter. When any sampling grid does not meet any of the four conditions, sampling in the corresponding direction is stopped, and the distance from the previous valid sampling grid to the docking point is determined as the effective length in the corresponding direction, resulting in the main direction positive effective length and the main direction negative effective length. The smaller value between the main direction positive effective length and the main direction negative effective length is determined as the main direction half-length constraint.

[0059] Starting from the docking point, center-symmetric sampling is performed along the positive and negative lateral directions of the local coordinate system with the same fixed grid step size. For each sampling grid, four conditions—courtyard boundary, obstacle distribution, real-time signal distribution, and body motion constraints—are repeatedly evaluated. If any sampling grid fails to meet any of these four conditions, sampling in the corresponding direction is stopped, and the distance from the previous valid sampling grid to the docking point is determined as the effective width in that direction. This yields the effective width in the positive and negative lateral directions. The smaller of these two effective widths is determined as the lateral half-width constraint. If the first sampling grid in the starting direction fails to meet the condition, the effective length or width in that direction is recorded as zero.

[0060] It should be noted that the use of centrally symmetric sampling and taking the smaller value as the initial constraint is to first ensure that the initial rectangular home area centered on the docking point and with the base station's orientation as the main direction has basic symmetry and consistent orientation, and then achieve environmental adaptability correction through subsequent unilateral contraction and contralateral compensation.

[0061] It should also be noted that the predicted signal strength of all valid grids within the current decision area is extracted from the real-time signal distribution. A valid grid is defined as a grid that simultaneously meets the following conditions: it is located within the courtyard boundary, does not overlap with the obstacle distribution, and satisfies the corresponding passage conditions for the body's motion constraints. The predicted signal strength of all valid grids within the current decision area is averaged to obtain the regional average signal strength. The deviation of the predicted signal strength of all valid grids within the current decision area from the regional average signal strength is statistically analyzed to obtain the regional signal dispersion. For each valid grid, valid grids in its adjacent grids are extracted, and the average signal difference between the current grid and its adjacent grids is calculated to obtain the local signal fluctuation. Valid grids that simultaneously meet the conditions of "predicted signal strength not less than the sum of regional average signal strength and regional signal dispersion" and "local signal fluctuation not greater than regional signal dispersion" are identified as high-reliability signal gain grids. Finally, interconnected high-reliability signal gain grids are merged to obtain a high-reliability signal gain interval.

[0062] S2.3. Based on the main direction half-length constraint and the lateral half-width constraint, determine the half-length and half-width of the initial rectangular home area to obtain the size parameters; according to the size parameters, use the rotation matrix to map the four vertices in the local coordinate system to the global occupied grid map coordinate system, and connect them to form a closed rectangular boundary to obtain the initial rectangular home area.

[0063] Specifically, based on the main direction half-length constraint and the lateral half-width constraint, combined with the length-to-width ratio preference of the rectangular region, the half-length and half-width of the initial rectangular home region are determined, and the expression is:

[0064] ;

[0065] b= ;

[0066] in, Let represent the half-length of the initial rectangular home area, and b represent the half-width of the initial rectangular home area. This indicates a half-length constraint in the main direction. Indicates lateral half-width constraint. This indicates the preferred aspect ratio of the rectangular area.

[0067] Based on the size parameters, with the docking point as the center point, the system expands along half the length of the base station's orientation direction and the opposite direction, and along half the width of the base station's orientation direction and the opposite direction. Using a rotation matrix, the four vertices in the local coordinate system are mapped to the global occupancy grid map coordinate system to obtain the coordinates of the four vertices of the initial rectangular home area. Then, the coordinates of the four vertices are connected in the global occupancy grid map to form a closed rectangular boundary, thereby generating the initial rectangular home area centered on the docking point and with the orientation direction as the main direction.

[0068] S2.4 Discretize the initial rectangular home area into multiple grid cells, determine whether each grid cell meets the passage conditions, and mark the grid cells that do not meet the passage conditions as conflict grids;

[0069] Specifically, the initial rectangular home area is projected onto the global occupied grid map. The initial rectangular home area is discretized into grids to be detected according to the grid side length. At the same time, the four rectangular sides of the initial rectangular home area are determined as the front side, back side, left side, and right side. Based on the grids to be detected and the four rectangular sides, each grid to be detected is judged one by one to determine whether it meets the passage conditions. Grids to be detected that do not meet the passage conditions are marked as conflict grids.

[0070] It should be noted that the passability conditions for each grid cell to be detected refer to whether each grid cell is located within the courtyard boundary, whether it coincides with the obstacle distribution, whether it meets the reliable communication conditions corresponding to the real-time signal distribution, and whether it meets the passability conditions corresponding to the robot's motion constraints. The passability conditions corresponding to the robot's motion constraints mean that the passable width of the area where the grid cell to be judged is located is not less than the robot's body width, the minimum net distance from the grid cell to the nearest obstacle distribution is not less than the parking safety distance threshold, and the area where the grid cell to be judged is located meets the turning requirements corresponding to the minimum turning radius. The reliable communication conditions corresponding to the real-time signal distribution mean that the predicted signal strength of the grid cell to be judged is not less than the sum of the regional average signal strength and the regional signal dispersion of the effective grid cells in the current area to be judged, and the local signal fluctuation of the grid cell to be judged is not greater than the regional signal dispersion.

[0071] Based on the conflict grid, the rectangular edges of the set of conflicting grids are identified as conflicting edges, and the conflict depth of each conflicting edge is calculated. Specifically, based on the conflict grid, the conflicting grid is assigned to the corresponding rectangular edge according to the minimum vertical distance from the conflicting grid to the four rectangular edges, thus obtaining the distribution range of conflicting grids corresponding to each rectangular edge; the rectangular edges of the set of conflicting grids are identified as conflicting edges, and the conflict depth of each conflicting edge is calculated. The conflict depth is represented by the maximum vertical distance from the conflicting edge to the center point of the corresponding conflicting grid.

[0072] Based on the conflict edges and conflict depths, the conflict edges are gradually moved towards the stopping point in order of decreasing conflict depth along a direction perpendicular to the conflict edges. After each move, the updated rectangular boundary is re-evaluated to see if it meets the passage conditions, until the moved rectangular boundary no longer corresponds to the conflict grid, thus obtaining the middle home area.

[0073] Based on the middle home region, the size difference between the middle home region and the initial rectangular home region is calculated, and the region is expanded outward in the direction of the rectangular side opposite to the conflict side at the same grid step size until the compensated rectangular boundary reaches the expansion limit or restores the preset size requirement, thus obtaining the candidate home region.

[0074] Furthermore, based on the central homecoming area, the size difference between the central homecoming area and the initial rectangular homecoming area in the main direction and lateral direction is compared. The opposite side of the conflicting side where the size decreases is determined as the compensation side. Then, the compensation side is gradually moved outward along the direction away from the stopping point with the same step size as the unilateral contraction. After each outward movement, it is re-evaluated whether the updated rectangular boundary meets the passage conditions. The most recent outward movement is withdrawn when the compensated rectangular boundary first fails to meet the passage conditions, or the outward movement stops when the compensated rectangular boundary returns to the preset size requirement of the corresponding axis. The candidate homecoming area is obtained. The preset size requirement is the size of the initial rectangular homecoming area determined based on the main direction half-length constraint, the lateral half-width constraint, and the length-width ratio preference of the rectangular area. The expansion limit is the maximum grid range that the candidate homecoming area boundary is allowed to expand under the conditions of not exceeding the courtyard boundary, not entering the obstacle distribution, and meeting the body movement constraints.

[0075] Based on the candidate home region, the rectangular boundary after unilateral contraction and contralateral compensation is determined as the final boundary, thus obtaining the heading coupled home region.

[0076] S3. Based on the heading coupled home area, the overall offset correction of the heading coupled home area is performed by integrating wall obstruction, slope change and signal attenuation, and candidate entrance gates are screened on the corrected boundary to obtain the optimal entrance gate.

[0077] S3.1 Extract the wall distribution, slope distribution and signal strength distribution around the heading coupling home area, determine the degree of imbalance in the heading coupling home area on both sides and in the front and back directions of the base station entry orientation, and determine the restricted direction;

[0078] Specifically, wall distribution, slope distribution, and signal strength distribution are extracted around the heading-coupled home area. The wall distribution is extracted through continuous and approximately linear obstacle grids in the obstacle distribution; the slope distribution is extracted through the body motion constraints in the environmental control data; and the signal strength distribution is extracted through the real-time signal distribution in the environmental control data.

[0079] The heading-coupled home area is divided into four directional regions: front, rear, left, and right, according to the base station's orientation. The minimum wall distance, average slope, and average predicted RSSI value are calculated for each of the four directional regions. The direction with the smaller minimum wall distance, larger average slope, or lower average predicted RSSI value is determined as the restricted direction.

[0080] S3.2. Based on the restricted direction, the heading-coupled home area is gradually translated along a direction away from the restricted direction. After each translation, the rectangular boundary after translation is rechecked to see if it meets the passage conditions, until a corrected home area that meets the requirements of wall avoidance, slope adaptation and signal reliability is obtained.

[0081] S3.3 Based on the corrected home area, the rectangular side of the base station's orientation and the two rectangular sides adjacent to the base station's orientation are used as candidate boundaries. Multiple continuous boundary segments are divided along the candidate boundaries according to a preset length. The passage conditions of the adjacent grids corresponding to each continuous boundary segment are judged, and the boundary segments that continuously meet the passage conditions are determined as candidate entrance gates.

[0082] S3.5. Based on the candidate entry gates, calculate the passage distance from the robot's current position to each candidate entry gate, the orientation adjustment amount from the candidate entry gate to the stopping point, the obstacle clearance in the neighborhood of the candidate entry gate, and the signal reliability in the neighborhood of the candidate entry gate, and sort them to obtain the optimal entry gate.

[0083] The optimal entry gate is determined by the following criteria: Passage distance is the cumulative path length obtained through A× algorithm search on the global occupied grid map using free grids as pathways; Orientation adjustment is represented by the absolute value of the angle between the direction from the geometric center of the candidate entry gate to the docking point and the base station's entry orientation; Obstacle clearance is represented by the minimum distance from the free grid in the candidate entry gate's neighborhood to the nearest obstacle distribution; Signal reliability is a gate-level evaluation index obtained based on the predicted RSSI average value of the effective grids in the candidate entry gate's neighborhood and the degree to which reliable communication conditions are met. Finally, the gates are sorted in the order of shortest passage distance, smallest orientation adjustment, largest obstacle clearance, and highest signal reliability to obtain the optimal entry gate.

[0084] S4. Based on the optimal entry gate, construct a multi-factor cost map by integrating obstacle risk, attitude adjustment cost, communication reliability and remaining power, and generate an adaptive home trajectory in segments.

[0085] S4.1. Based on the optimal entry gate and environmental control data, determine the starting grid, the first target grid, and the second target grid. Then, based on obstacle risk, attitude adjustment cost, communication reliability, and remaining power, assign a cost value to the passable grid in the global occupied grid map to obtain a multi-factor cost map.

[0086] Specifically, the robot's current position, heading-coupled home area, docking point position, obstacle distribution, real-time signal distribution, and body motion constraints are read from environmental control data, along with the remaining battery power. The grid mapped to the robot's current position in the global occupied grid map is designated as the starting grid, the target free grid corresponding to the optimal entry gate is designated as the first target grid, and the grid corresponding to the docking point position is designated as the second target grid.

[0087] Using the outer grid of the courtyard boundary, the grid corresponding to the obstacle distribution, and the restricted access grid in the body motion constraints as impassable grids, the single-step passage cost is calculated for all passable grids. The single-step passage cost includes four factors: obstacle risk, attitude adjustment cost, communication reliability, and remaining battery power. The calculation formula is as follows:

[0088] ;

[0089] in, Indicates from grid Move to adjacent grid The cost of one-step passage, Represents a grid With grid The distance between them This indicates the safe parking distance threshold. Represents a grid The minimum net distance to the nearest obstacle distribution. Indicates from grid Pointing grid The direction angle of movement, Indicates the starting grid to the grid. The cumulative path length, This indicates the maximum driving distance corresponding to the remaining battery power. Represents a grid The corresponding communication reliability, This represents a positive constant to prevent the denominator from being zero.

[0090] After assigning single-step passage costs to all passable grids, a multi-factor cost map is obtained.

[0091] S4.2 Based on the multi-factor cost map, the A× algorithm is executed to search from the current position to the optimal entry gate to obtain the first trajectory segment.

[0092] Based on the first trajectory segment, the second trajectory segment is generated grid by grid within the heading coupled home area, and it is determined whether the second trajectory segment meets the passage conditions corresponding to the environmental control data, until the second trajectory segment that meets the passage conditions is obtained.

[0093] Starting from the first target grid and ending at the docking point, and with the base station's orientation as the main direction, a second trajectory segment is generated grid by grid within the heading-coupled home area. The baseline of the second trajectory segment is represented by the projection line along the base station's orientation. After the second trajectory segment is generated, each grid on the second trajectory segment is checked to see if it meets the passage conditions corresponding to the environmental control data. If it does not meet the conditions, the first target grid is reselected within the optimal entry gate area in order of proximity to the geometric center, and the second trajectory segment is regenerated until a second trajectory segment that meets the passage conditions is obtained.

[0094] Based on the first trajectory segment and the second trajectory segment that meets the traffic conditions, the trajectory is spliced ​​together according to the driving sequence, and duplicate grids are removed to obtain the adaptive home trajectory.

[0095] S5. Based on the adaptive home trajectory, control the robot to return home, and detect posture deviation, dynamic obstacles and signal changes in real time until it completes a precise stop.

[0096] Based on the adaptive home trajectory, current location, optimal entry door and stopping point location, determine the trajectory tracking sequence, entry switching area and stopping endpoint grid;

[0097] Furthermore, a trajectory tracking sequence is established according to the order of the grids in the adaptive home trajectory. The grid corresponding to the robot's current position is determined as the tracking starting grid, the first unpassed trajectory grid in front of the current position is determined as the current target grid, the grid segment corresponding to the optimal entry gate is determined as the entry switching area, and the grid corresponding to the docking point position is determined as the docking ending grid.

[0098] Based on the trajectory tracking sequence, the robot is controlled to move along an adaptive home trajectory, and after entering the neighborhood of the current target grid at the current position, it switches to the next target grid to obtain the trajectory tracking result; based on the trajectory tracking result, pose deviation, dynamic obstacles and signal changes are continuously detected to obtain the trajectory validity determination result;

[0099] Specifically, based on the trajectory tracking sequence, the robot begins to move in the direction of the current target grid and continuously reads the SLAM localization results during the movement. It calculates the position deviation using the current position coordinates and the current target grid coordinates, and calculates the orientation deviation using the current orientation and the tangential direction of the current target grid. Then, it adjusts the robot's forward speed and turning speed according to the position deviation and orientation deviation, so that the robot always moves closer to the current target grid. When the current position enters the neighborhood of the current target grid, the next unpassed trajectory grid in the trajectory tracking sequence is replaced with the new current target grid, forming a continuous pose deviation detection result.

[0100] Based on the pose deviation detection results, while the robot moves along the adaptive home trajectory, it continuously reads LiDAR scan data, ultrasonic sensor ranging data, and the current position RSSI measured value. The current LiDAR scan data and ultrasonic sensor ranging data are projected onto the global occupancy grid map, and the current occupancy grid is compared with the obstacle distribution in the environmental control data grid by grid. When a new occupancy grid that is not in the environmental control data appears in the safe passage zone ahead of the adaptive home trajectory and remains in existence for multiple consecutive sampling cycles, the new occupancy grid is identified as a dynamic obstacle. At the same time, the current position RSSI measured value is compared with the reliable communication conditions of the grid corresponding to the real-time signal distribution. When the current position RSSI measured value is lower than the threshold corresponding to the reliable communication conditions, the current position is identified as a signal change position. Based on the detection results of pose deviation, dynamic obstacles, and signal changes, it is determined whether the trajectory segment between the current position and the current target grid remains valid. When the position deviation exceeds the position deviation threshold, the orientation deviation exceeds the orientation deviation threshold, there are dynamic obstacles in the safe passage zone ahead of the adaptive home trajectory, or the measured RSSI value of the current position does not meet the reliable communication conditions, the unexecuted trajectory segment after the current position is determined as a failed trajectory segment. When the position deviation, orientation deviation, dynamic obstacle detection results, and the measured RSSI value of the current position all meet the requirements, the unexecuted trajectory segment after the current position is determined as a valid trajectory segment, forming a trajectory validity determination result.

[0101] It should be noted that the position deviation threshold is set by first reading the positioning error radius corresponding to the SLAM positioning result, then reading the grid side length of the globally occupied grid map, and finally taking the larger value between the positioning error radius and a grid side length. The positioning error radius is used to reflect the measurement error of the current position, and the grid side length is used to reflect the map discretization error during trajectory tracking. The orientation deviation threshold is determined by reading the robot width, the docking safety distance threshold, the obstacle clearance in the optimal entry door neighborhood, the half-width of the heading coupling home area, and the neighborhood radius of the docking point. The first lateral margin is obtained by subtracting the docking safety distance threshold and half of the robot width from the obstacle clearance in the optimal entry door neighborhood. The second lateral margin is obtained by subtracting half of the robot width from the half-width of the heading coupling home area. The smaller value between the first and second lateral margins is taken as the allowable lateral offset. Finally, combined with the neighborhood radius of the docking point, the maximum orientation deviation that ensures the robot does not exceed the allowable lateral offset within the remaining approach distance is determined as the orientation deviation threshold.

[0102] Based on the trajectory validity determination result, the driving trajectory near the stop point is obtained or an adaptive home trajectory is regenerated until a precise stop is completed.

[0103] Furthermore, if the trajectory validity determination result is a valid trajectory segment, the robot continues to move along the trajectory tracking sequence and, after entering the entry switching area at the current position, slowly approaches the docking point along the second trajectory segment. If the trajectory validity determination result is a failed trajectory segment, the robot immediately stops moving, writes the current position, dynamic obstacles, and the measured RSSI value of the current position back to the global occupancy grid map, and then, using the current position as the new starting position, re-executes the candidate entry gate screening process in step S3 to obtain a new optimal entry gate. Then, it re-executes the multi-factor cost map construction and adaptive home trajectory generation process in step S4 to obtain a new adaptive home trajectory, and replaces the failed trajectory segment after the current position with the new adaptive home trajectory. This process is repeated until the current position enters the neighborhood of the docking point, and the orientation deviation between the current position orientation and the base station entry orientation is less than the orientation deviation threshold. At this point, the robot maintains a low-speed straight line along the base station entry orientation to the docking point, completing a precise docking.

[0104] In summary, this invention combines obstacle conflict to perform unilateral contraction and contralateral compensation to obtain a heading-coupled home area. This achieves directional correction of the home area boundary based on the courtyard boundary, obstacle distribution, real-time signal distribution, and aircraft motion constraints, ensuring that the area shape is consistent with the base station's entry orientation, thus providing stable spatial constraints for subsequent entry paths. By constructing a multi-factor cost map based on the optimal entry gate, integrating obstacle risk, attitude adjustment cost, communication reliability, and remaining battery power, and generating adaptive home trajectories in segments, this invention achieves integrated organization of the home path and the final docking process. This allows trajectory tracking, trajectory validity determination, and replanning processes to be executed around the same multi-factor cost constraints, thereby improving the adaptability of path planning and the stability of precise docking.

[0105] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for adaptively generating a robot's home trajectory based on intelligent control, characterized in that, include: Collect data on the robot's current position, docking point position, base station orientation, courtyard boundary, obstacle distribution, real-time signal distribution, and robot motion constraints to form environmental control data; Based on the environmental control data, a correspondence between the local coordinate system and the global occupied grid map coordinate system is established, an initial rectangular home area is generated, and unilateral contraction and contralateral compensation are performed in combination with obstacle conflict to obtain the heading coupled home area; Based on the heading-coupled home area, the overall offset correction of the heading-coupled home area is performed by incorporating wall obstruction, slope change and signal attenuation, and candidate entrance gates are screened on the corrected boundary to obtain the optimal entrance gate; Based on the optimal entry gate, a multi-factor cost map is constructed by integrating obstacle risk, attitude adjustment cost, communication reliability, and remaining power, and an adaptive home trajectory is generated in segments. Based on the adaptive home trajectory control system, the robot returns home by detecting posture deviations, dynamic obstacles, and signal changes in real time until it achieves a precise stop.

2. The adaptive generation method for robot home trajectory based on intelligent control as described in claim 1, characterized in that, The formation of environmental control data refers to writing the robot's current position, docking point position, base station orientation, courtyard boundary, obstacle distribution, real-time signal distribution, and robot motion constraints into a global occupancy grid map to form environmental control data.

3. The adaptive generation method for robot home trajectory based on intelligent control as described in claim 2, characterized in that, The process of establishing a correspondence between the local coordinate system and the global occupancy grid map coordinate system based on the environmental control data, and generating an initial rectangular home area, includes the following steps: Based on environmental control data, a local coordinate system is established, and a rotation matrix is ​​used to establish the correspondence between the local coordinate system and the global occupied raster map coordinate system; Based on the correspondence, as well as the courtyard boundary, obstacle distribution, real-time signal distribution, and body motion constraints, the main direction half-length constraint and the lateral half-width constraint are determined. Based on the main direction half-length constraint and the lateral half-width constraint, the half-length and half-width of the initial rectangular home region are determined, and the size parameters are obtained. Based on the size parameters, the four vertices in the local coordinate system are mapped to the global occupied raster map coordinate system using a rotation matrix, and then connected to form a closed rectangular boundary, thus obtaining the initial rectangular home area.

4. The adaptive generation method for robot home trajectory based on intelligent control as described in claim 1, characterized in that, The process of combining obstacle conflict to perform unilateral contraction and contralateral compensation to obtain the heading-coupled home region includes the following steps: The initial rectangular home area is discretized into multiple grid cells. Each grid cell is checked to see if it meets the passage conditions. Grid cells that do not meet the passage conditions are marked as conflict grids. Based on the conflict grid, the rectangular edges of the set of conflict grids are identified as conflict edges, and the conflict depth of each conflict edge is calculated. Based on the conflict edge and conflict depth, the conflict edge is gradually moved towards the stop point along a direction perpendicular to the conflict edge until the moved rectangular boundary no longer corresponds to the conflict grid, thus obtaining the middle home area; Calculate the size difference between the intermediate home region and the initial rectangular home region, and expand outwards along the rectangular side opposite the conflict side at the same grid step size until the compensated rectangular boundary reaches the expansion limit or restores the preset size requirement to obtain the candidate home region; Based on the candidate home region, the rectangular boundary after unilateral contraction and contralateral compensation is determined as the final boundary, thus obtaining the heading coupled home region.

5. The adaptive generation method for robot home trajectory based on intelligent control as described in claim 4, characterized in that, The process of performing an overall offset correction on the heading-coupled home area based on the heading-coupled home area, incorporating factors such as wall obstruction, slope changes, and signal attenuation, includes the following steps: Extract wall distribution, slope distribution and signal strength distribution around the heading-coupled home area to determine the degree of imbalance in the heading-coupled home area in the directions to the sides and front and back of the base station, and determine the restricted direction; Based on the restricted direction, the heading-coupled home area is gradually shifted away from the restricted direction until a corrected home area that meets the requirements of wall avoidance, slope adaptation, and signal reliability is obtained.

6. The adaptive generation method for a robot's home trajectory based on intelligent control as described in claim 5, characterized in that, The step of filtering candidate entry gates on the corrected boundary to obtain the optimal entry gate includes the following steps: Based on the corrected home area, the rectangular side of the base station's orientation and the two rectangular sides adjacent to the base station's orientation are used as candidate boundaries. Divide the candidate boundary into multiple continuous boundary segments of a preset length, and determine the passage conditions of the adjacent grids corresponding to each continuous boundary segment to obtain the candidate entry gate; Based on the candidate entry gates, the travel distance from the robot's current position to each candidate entry gate, the orientation adjustment amount from the candidate entry gate to the stopping point, the obstacle clearance in the neighborhood of the candidate entry gate, and the signal reliability in the neighborhood of the candidate entry gate are calculated and sorted to obtain the optimal entry gate.

7. The adaptive generation method for a robot's home trajectory based on intelligent control as described in claim 6, characterized in that, The construction of the multi-factor cost map refers to determining the starting grid, the first target grid, and the second target grid based on the optimal entry gate and environmental control data, and assigning a cost value to the passable grids in the global occupied grid map based on obstacle risk, attitude adjustment cost, communication reliability, and remaining power, thereby obtaining the multi-factor cost map.

8. The adaptive generation method for robot home trajectory based on intelligent control as described in claim 7, characterized in that, The segmented generation of the adaptive home trajectory includes the following steps: Based on the multi-factor cost map, the A× algorithm is used to search from the current position to the optimal entry gate to obtain the first trajectory segment. Based on the first trajectory segment, the second trajectory segment is generated grid by grid within the heading coupled home area, and it is determined whether the second trajectory segment meets the passage conditions corresponding to the environmental control data, until the second trajectory segment that meets the passage conditions is obtained. Based on the first trajectory segment and the second trajectory segment that meets the traffic conditions, the trajectory is spliced ​​together according to the driving sequence to obtain the adaptive home route.

9. The adaptive generation method for a robot's home trajectory based on intelligent control as described in claim 8, characterized in that, The process of controlling the robot to return home based on the adaptive home trajectory, detecting pose deviations, dynamic obstacles, and signal changes in real time until precise docking, includes the following steps: Based on the adaptive home trajectory, current location, optimal entry door and stopping point location, determine the trajectory tracking sequence, entry switching area and stopping endpoint grid; Based on the trajectory tracking sequence, the robot is controlled to move along an adaptive home trajectory, and after entering the neighborhood of the current target grid at the current position, it switches to the next target grid to obtain the trajectory tracking result; Based on the trajectory tracking results, pose deviation, dynamic obstacles and signal changes are continuously detected to obtain the trajectory validity determination results; Based on the trajectory validity determination result, the driving trajectory near the stop point is obtained or an adaptive home trajectory is regenerated until a precise stop is completed.

10. The adaptive generation method for a robot's home trajectory based on intelligent control as described in claim 9, characterized in that, The process of obtaining a driving trajectory close to the stop point or regenerating an adaptive home trajectory based on the trajectory validity determination result means that when the trajectory is valid, the robot continues to run along the adaptive home trajectory and enters the stop point along the subsequent trajectory of the optimal entry gate; when the trajectory fails, the robot stops moving, reselects the optimal entry gate, reconstructs the multi-factor cost map, and regenerates the adaptive home trajectory.

Images