CN-121995948-A - Dynamic obstacle pre-judging obstacle avoidance method and system for cleaning robot

CN121995948ACN 121995948 ACN121995948 ACN 121995948ACN-121995948-A

Abstract

The invention belongs to the technical field of robot motion control, and discloses a method and a system for pre-judging and avoiding a dynamic obstacle of a cleaning robot, wherein the method comprises the steps of acquiring spatial distribution state information of the obstacle in the operation environment of the cleaning robot and motion state information of the cleaning robot, and organizing the spatial distribution state information and the motion state information into a time sequence motion state set uniformly; based on a time sequence motion state set, deducing a motion expansion trend of an obstacle in a preset future time window to form a motion expansion area, performing spatial correlation analysis on the motion expansion area and a planning path of the cleaning robot to obtain a trend description result of the obstacle relative to the planning path, constructing an obstacle risk potential energy function according to the trend description result, representing the approach strength of the obstacle relative to the cleaning robot path in the preset future time window, acquiring the risk potential energy of the obstacle, ensuring the running safety and the path smoothness, and improving the cleaning efficiency and the environment adaptability of the cleaning robot.

Inventors

WANG ZHENYA
YAO YAQIANG
LIN SHUANGWU

Assignees

安徽橙犀科技有限公司

Dates

Publication Date: 20260508
Application Date: 20260130

Claims (10)

1. The dynamic obstacle pre-judging and avoiding method for the cleaning robot is characterized by comprising the following steps of: S1, acquiring spatial distribution state information of obstacles in an operation environment of a cleaning robot and motion state information of the cleaning robot, and organizing the spatial distribution state information and the motion state information into a time sequence motion state set in a unified way; s2, deducing a movement expansion trend of the obstacle in a preset future time window based on a time sequence movement state set to form a movement expansion area, and carrying out space correlation analysis on the movement expansion area and a planning path of the cleaning robot to obtain a trend description result of the obstacle relative to the planning path; S3, constructing an obstacle risk potential energy function according to a trend description result, representing the approach strength of the obstacle relative to the cleaning robot path in a preset future time window, and acquiring the risk potential energy of the obstacle; S4, when the risk potential of the obstacle exceeds a preset risk potential threshold, a dynamic safety corridor dynamically adjusted along with the obstacle is generated in front of the cleaning robot, and the width of the dynamic safety corridor is adjusted according to the risk potential of the obstacle; S5, re-planning a local path of the cleaning robot in a safe passing area limited by the dynamic safety corridor, outputting a corresponding motion control instruction, and driving the cleaning robot to finish the advanced avoidance of the dynamic obstacle.
2. The method for pre-judging obstacle avoidance of a dynamic obstacle for a cleaning robot according to claim 1, wherein the method for acquiring the spatial distribution state information and the motion state information of the cleaning robot comprises: In the running process of the cleaning robot, an environment sensing unit carried by the cleaning robot scans obstacles in the running environment, acquires spatial position information and occupied boundary information of the obstacles relative to the cleaning robot at each sampling moment, combines the spatial position information and the occupied boundary information, and uniformly represents the spatial position information and the occupied boundary information as spatial distribution state information of the obstacles; and synchronously acquiring the motion state information of the cleaning robot at sampling moments corresponding to the spatial distribution state information acquisition through a motion detection unit in the cleaning robot, wherein the motion detection unit comprises an odometer and an inertial measurement unit, and the motion state information comprises pose information and motion speed information of the cleaning robot.
3. The method for pre-judging obstacle avoidance of a dynamic obstacle for a cleaning robot according to claim 2, wherein the method for acquiring the time-series motion state set comprises: And sequentially arranging and caching different state description units according to the sampling time sequence to form a time sequence motion state set representing the change of the relative motion relation of the obstacle and the cleaning robot along with time.
4. A dynamic obstacle avoidance method for a cleaning robot as recited in claim 3, wherein the method of forming the motion extension comprises: Determining the relative movement speed of the obstacle relative to the cleaning robot by differentiating the spatial positions of the obstacle at adjacent sampling moments based on the spatial position change relation of the obstacle in the time sequence movement state set at different continuous sampling moments, and estimating the speed change rate of the obstacle by differentiating the relative movement speed of the cleaning robot at adjacent time intervals; In a preset future time window, the current movement speed and the speed change rate of the obstacle are synthesized, and the maximum displacement range possibly generated by the obstacle in the future time window is evaluated to obtain a movement expansion radius representing the future movement uncertainty of the obstacle; And constructing a spatial range which can be possibly reached by the obstacle in a future time window by taking the spatial position of the obstacle at the current moment as a center according to the motion expansion radius under a unified motion reference system adopted by the cleaning robot, and determining the spatial range as a motion expansion area of the obstacle.
5. The method for pre-judging obstacle avoidance of a dynamic obstacle for a cleaning robot of claim 4 wherein said method for deriving trend descriptive results for the obstacle relative to the planned path comprises: After forming a motion expansion area, generating a planning path based on a DWA algorithm in the process of executing a cleaning task by the cleaning robot, and selecting a planning path section positioned in a preset future time window under a unified motion reference system; Calculating the spatial relationship between a motion expansion area corresponding to the obstacle in a preset future time window and a planned path section, calculating Euclidean distance from each path point on the planned path section to each boundary position on the boundary of the motion expansion area, taking the Euclidean distance as the spatial distance from the path point to the boundary of the motion expansion area, and taking the minimum spatial distance from the path point to the boundary of the motion expansion area in the range of the planned path section; Judging the movement trend of the obstacle relative to the planned path in a preset future time window based on the minimum space distance, and judging that the obstacle presents an approaching trend relative to the planned path when the minimum space distance is smaller than a preset minimum space distance threshold value; and when the minimum space distance is greater than or equal to a preset minimum space distance threshold value, judging that the obstacle presents a parallel or distant trend relative to the planned path, and taking the judgment result as a trend description result of the obstacle relative to the planned path.
6. The method for pre-judging obstacle avoidance of a cleaning robot according to claim 5, wherein the method for acquiring risk potential of the obstacle comprises: According to the trend description result, constructing an obstacle risk potential energy function aiming at any perceived obstacle, and representing the approaching strength of the obstacle relative to the cleaning robot path in a preset future time window so as to acquire the risk potential energy of the obstacle.
7. The method for dynamically predicting obstacle avoidance of a cleaning robot of claim 6 wherein the method for creating a dynamic safety corridor dynamically adjusted with an obstacle comprises: When detecting that the corresponding risk potential of the perceived obstacle exceeds a preset risk potential threshold value in a preset future time window, judging that the obstacle forms a dynamic invasion risk to the current planning path of the cleaning robot, and triggering the generation of a dynamic safety corridor in front of the cleaning robot; And constructing a continuous dynamic safety corridor in front of the cleaning robot along the direction of the planned path under a unified motion reference system by taking the planned path generated at the current moment of the cleaning robot as a reference, wherein the dynamic safety corridor longitudinally covers a planned path section of the cleaning robot travelling in a preset future time window, and transversely reserves preset safety passing spaces for two sides of the planned path.
8. The method for pre-judging obstacle avoidance of a dynamic obstacle for a cleaning robot of claim 7, wherein the method for adjusting the width of the dynamic safety corridor comprises: According to the risk potential energy of the obstacle, the width of the dynamic safety corridor is adjusted, the preset basic safety width is used as the minimum width constraint of the dynamic safety corridor, the moving speed of the obstacle is introduced to carry out self-adaptive adjustment on the width of the dynamic safety corridor, and the transverse width of the dynamic safety corridor is dynamically updated along with the change of the moving state corresponding to the obstacle risk potential energy, so that the evolution trend of the obstacle risk potential energy is kept consistent in the time dimension.
9. The method for pre-judging obstacle avoidance of a dynamic obstacle for a cleaning robot according to claim 8, wherein the method for driving the cleaning robot to accomplish the advanced avoidance of the dynamic obstacle comprises: During the process that the cleaning robot executes the cleaning task and moves along the planned path, re-planning the local path of the cleaning robot in a future time window in a safe passing area defined by the dynamic safety corridor; And after the local path re-planning is completed, synchronously generating a motion control instruction of the cleaning robot, and enabling the cleaning robot to run along the local path according to the motion control instruction so as to realize the advanced avoidance of the dynamic obstacle.
10. A dynamic obstacle pre-judging obstacle avoidance system of a cleaning robot for implementing a dynamic obstacle pre-judging obstacle avoidance method of a cleaning robot as claimed in any one of claims 1 to 9, comprising: the motion state construction module is used for acquiring the spatial distribution state information of the obstacle in the operation environment of the cleaning robot and the motion state information of the cleaning robot, and organizing the spatial distribution state information and the motion state information into a time sequence motion state set in a unified way; The obstacle behavior analysis module is used for deducing the movement expansion trend of the obstacle in a preset future time window based on the time sequence movement state set to form a movement expansion area, and carrying out space association analysis on the movement expansion area and the planning path of the cleaning robot to obtain a trend description result of the obstacle relative to the planning path; the obstacle intention prejudging module is used for constructing an obstacle risk potential energy function according to the trend description result, representing the approach strength of the obstacle relative to the cleaning robot path in a preset future time window, and acquiring the risk potential energy of the obstacle; The system comprises a safety corridor generating module, a dynamic safety corridor generating module, a safety corridor generating module and a cleaning robot, wherein the safety corridor generating module is used for generating a dynamic safety corridor dynamically adjusted along with the obstacle in front of the cleaning robot when the risk potential of the obstacle exceeds a preset risk potential threshold value; The obstacle avoidance path execution module is used for re-planning a local path of the cleaning robot in a safe passing area limited by the dynamic safety corridor, outputting a corresponding motion control instruction and driving the cleaning robot to finish the advanced avoidance of the dynamic obstacle.

Description

Dynamic obstacle pre-judging obstacle avoidance method and system for cleaning robot Technical Field The invention relates to the technical field of robot motion control, in particular to a method and a system for pre-judging and avoiding dynamic obstacles of a cleaning robot. Background The existing method and system for pre-judging and avoiding the dynamic obstacle of the cleaning robot mainly have the following problems: Along with the wide application of intelligent cleaning robots in families, offices and public places, the improvement of the autonomous obstacle avoidance capability of the robots in dynamic environments becomes an important technical link for ensuring safe operation and improving cleaning efficiency. The existing obstacle avoidance method and system for the dynamic obstacle of the cleaning robot mainly carry out path planning and motion control by sensing the position and speed information of the obstacle, but a plurality of technical defects still exist in practical application. Most of the existing dynamic obstacle avoidance methods are only used for predicting the instantaneous position or the instantaneous speed of the obstacle at the current moment, and the future position of the obstacle is usually deduced by adopting a uniform speed assumption or a simple linear extrapolation mode, so that the trend of the movement state of the obstacle along with the time cannot be effectively described. When acceleration, deceleration or movement direction adjustment exists on the obstacle, the actual movement range of the obstacle in the future time is easily underestimated, so that obstacle avoidance decision hysteresis or insufficient safety margin are caused, and potential collision risks exist. Although the prior art introduces obstacle trajectory prediction, the prior art relies on complex machine learning models or behavior recognition algorithms, and the methods have large dependence on training data and computing resources. In a home or public environment, obstacles such as pedestrians, pets and the like have motion behaviors with strong randomness, so that the stability and the interpretability of a prediction result are insufficient, and the prediction result is difficult to reliably deploy on a cleaning robot platform with limited resources. The existing obstacle avoidance method usually takes a single predicted point or predicted track as a core to judge, lacks of spatial expression of future movement uncertainty of the obstacle, is difficult to form continuous and stable spatial constraint in the path planning process, and is easy to cause frequent re-planning or path jitter, so that cleaning efficiency and system reliability are affected. In the prior art, risks are judged based on the instantaneous space distance or whether an obstacle enters a preset safety area, quantitative description of the movement trend of the obstacle and the relative movement relation of the obstacle and the cleaning robot is lacking, and the actual threat degree of the obstacle to a planned path in future time is difficult to accurately reflect. In a dynamic scenario, the prior art generally only considers the current position or absolute speed of the obstacle, and does not model the obstacle speed uniformly with the motion state of the cleaning robot itself, resulting in hysteresis or distortion of the risk assessment results when the robot accelerates, decelerates or turns. For dynamic obstacles with sudden movement behaviors, such as pedestrians and pets with sudden acceleration, direction change or discontinuous track, the prior art lacks an explicit characterization mechanism for the behavior mutation characteristics, and risk underestimation is easy to occur, so that collision, frequent scram or path deviation is caused, and the safety and the continuity of a cleaning task are further affected. In view of the above, the present invention provides a dynamic obstacle pre-judging obstacle avoidance method for a cleaning robot to solve the above-mentioned problems. Disclosure of Invention In order to overcome the defects in the prior art, the invention provides the following technical scheme that the method for pre-judging and avoiding the obstacle of the dynamic obstacle of the cleaning robot comprises the following steps: S1, acquiring spatial distribution state information of obstacles in an operation environment of a cleaning robot and motion state information of the cleaning robot, and organizing the spatial distribution state information and the motion state information into a time sequence motion state set in a unified way; s2, deducing a movement expansion trend of the obstacle in a preset future time window based on a time sequence movement state set to form a movement expansion area, and carrying out space correlation analysis on the movement expansion area and a planning path of the cleaning robot to obtain a trend description result of the obstacle relative to the planning path; S3, cons