Patrol robot system with path planning and dynamic obstacle avoidance functions

Abstract

The invention relates to the field of patrol robots, in particular to a patrol robot system with path planning and dynamic obstacle avoidance functions, and aims to solve the problems that patrol paths and patrol periods of patrol robots cannot be intelligently adjusted and patrol resources are easily allocated unreasonably. According to the method, statistics is carried out on abnormal conditions obtained by robot patrol, credibility judgment is carried out on the abnormal conditions, the patrol alarm function is achieved, when the abnormal conditions are not confirmed to be alarm conditions, the abnormal conditions are reserved, the patrol weight of the corresponding area is modified through the abnormal conditions, and the patrol accuracy is improved. Therefore, the residence time of an area with an abnormal condition is automatically and dynamically adjusted, the continuous observation time of the abnormal condition is prolonged, the security of a key area is improved, meanwhile, interference planning can be carried out on the moving condition of an obstacle and the movement of the obstacle, the influence of the obstacle is dynamically evaluated, and an obstacle avoidance route is temporarily planned. And secondary dynamic adjustment is realized, and the configuration of patrol resources is optimized.

Description

Technical Field

[0001] The present invention relates to the field of patrol robots, in particular to a patrol robot system with path planning and dynamic obstacle avoidance functions. Background Art

[0002] With the steady growth of my country's social economy, more and more corporate factories, high-tech parks, and shopping malls are constantly appearing in people's lives. These places have put forward new and special demands on production safety. Usually, the safety tasks in these places are completed by safety inspectors. However, with the continuous expansion of inspection scope, mixed indoor and outdoor environments, high risks in special environments, and the continuous increase in employment costs, relying solely on safety inspectors can no longer meet the increasingly complex security needs.

[0003] Against the backdrop of rapid technological development, robotics has been widely used in various fields, such as exploration, reconnaissance, and search and rescue. The working environments in these fields are characterized by complex terrains, and the robot's motion performance has gradually improved with technological development. Therefore, some current technical solutions use robots for patrols to replace the shortcomings of manual patrols.

[0004] For example, the existing patent application CN2021104555663 discloses a technical solution, in which the robot is controlled to move according to a preset path and to judge the abnormality of the environment during the movement, thereby realizing patrol of the environment. At the same time, the robot's obstacle avoidance function is realized by radar positioning, which improves the intelligence of the robot. However, when judging the environment in this solution, the judgment results have two states: alarm and non-alarm. Therefore, when some environments do not meet the alarm standard but have a certain degree of minor abnormalities, the patrol robot is classified as a non-alarm state. Therefore, if the abnormality increases after the patrol, the patrol robot will not be able to quickly conduct a second inspection of the area with minor abnormalities, resulting in the inability to realize intelligent adjustment of the robot's patrol path and patrol cycle;

[0005] In response to the above technical problems, this application proposes a solution. Summary of the Invention

[0006] The present invention realizes the patrol alarm function by collecting statistics on abnormal situations obtained by robot patrols and making credibility judgments on the abnormal situations. When the abnormal situations are not confirmed as alarm situations, the abnormal situations are retained and the patrol weights of the corresponding areas are modified according to the abnormal situations, thereby autonomously and dynamically adjusting the stay time in areas where abnormal situations exist, increasing the continuous observation time of abnormal situations, improving the security of key areas, realizing the optimal allocation of patrol resources, and solving the problem that the patrol path and patrol cycle of patrol robots cannot be adjusted intelligently, and the patrol resource allocation is prone to unreasonable. A patrol robot system with path planning and dynamic obstacle avoidance functions is proposed.

[0007] The purpose of the present invention can be achieved through the following technical solutions:

[0008] A patrol robot system with path planning and dynamic obstacle avoidance functions includes a route interval management module, which is used to record the patrol path of the robot and obtain the patrolled path and the transit time of each path point in the patrolled path;

[0009] An abnormal area identification module, which counts the route passed by the patrol robot and counts the abnormal points determined on the route. The abnormal area identification unit divides the patrolled path into multiple sub-areas and obtains the sub-areas where the abnormal points are located;

[0010] A route planning module, which obtains data from the route interval management module and generates a route patrol repetition index. It obtains abnormal area related data from the abnormal area identification module and generates a route patrol enhancement index. The route planning module performs route priority planning based on the route patrol enhancement index and abnormal area related indicators.

[0011] A dynamic recognition module, which scans with sensors to obtain obstacle information and dynamically evaluates the obstacle information to obtain obstacle obstruction information;

[0012] An obstacle avoidance recognition module, after obtaining obstacle obstruction information, calculates the obstruction situation of the obstacle obstruction information on the patrol route, generates a temporary obstacle avoidance route, and sends the temporary obstacle avoidance route to the hybrid decision module;

[0013] A hybrid decision module obtains the route results of route priority planning, and uses the route results as a benchmark to synthesize the temporary obstacle avoidance route to obtain a dynamic route result. The hybrid decision module continuously generates temporary routes and uses the dynamic route results as the final patrol route.

[0014] As a preferred embodiment of the present invention, the route interval management module operates by obtaining a preset initial route when recording the patrolled route, and records the passed portion as the patrolled route;

[0015] After obtaining the preset initial route, the route interval management module divides the initial route into multiple sub-areas and records the time point when the robot leaves the route included in the sub-area as the passing time of the path point.

[0016] As a preferred embodiment of the present invention, the method for the abnormal area identification module to obtain abnormal points on the route is:

[0017] The abnormal area recognition module collects images through the 3D structured light camera, visible light camera and infrared camera every time the robot arrives at a sub-area, performs recognition and analysis based on the collected images, obtains the dynamic pixel area and the temperature-prominent pixel area in the image through a preset recognition algorithm, and recognizes the dynamic pixel area through the dynamic recognition algorithm. The recognition result is compared with the set intrusion sample database to determine the intrusion credibility. If the intrusion credibility reaches the set threshold, it is recorded as an intrusion warning. If it does not reach the set threshold, it is recorded as an abnormal point.

[0018] The temperature-prominent pixel areas are compared with the infrared recognition database to confirm the credibility of fire hazard areas and personnel activity areas. If the credibility reaches the set threshold, a danger warning is generated. If it does not reach the set threshold, it is recorded as an abnormal point.

[0019] As a preferred embodiment of the present invention, after the path planning module obtains the patrolled path and the elapsed time of the path point, the patrolled path is compared with the set initial path to obtain the non-patrolled path, and the path planning module assigns a non-repetition indicator to all non-patrolled paths, wherein the initial path obtained by the path planning module is a cyclic path, and after each patrol is completed, all paths are changed to non-patrolled paths;

[0020] After obtaining an anomaly point, the path planning module records the area to which the anomaly point belongs, numbers the area to which it belongs, and creates an empty set. For each anomaly point, a set of data is added to the set. The path planning module counts all areas with anomalies in the set initial path, and records the number of data in the set corresponding to the area as the patrol enhancement indicator.

[0021] As a preferred embodiment of the present invention, the path planning module selects each sub-area in the set initial path, and quantifies and superimposes the non-repetitive index and patrol enhancement index of each sub-area to obtain a final priority index;

[0022] The path planning module obtains the priority indicators of all sub-areas and the patrol speed of the set initial path, calculates the initial patrol time T based on the length of the initial path and the patrol speed, and allocates time through the final priority indicator to obtain the allocated time ti for each sub-area, where i is the number of the sub-area, ti is positively correlated with the final priority indicator, and ∑ti=T.

[0023] As a preferred embodiment of the present invention, the obstacle information obtained by the dynamic recognition module includes the obstacle volume and the obstacle movement vector. The method for the dynamic recognition module to obtain the obstacle volume is to create and identify three-dimensional coordinates through a 3D point cloud. The method for the dynamic recognition module to obtain the obstacle movement vector is to perform difference calculation on the three-dimensional coordinates of the 3D point cloud created twice in succession to obtain the movement distance and movement direction of the obstacle, and at the same time, calculate based on the time interval and distance between the two consecutive creations to obtain the movement speed of the obstacle.

[0024] As a preferred embodiment of the present invention, the obstacle avoidance recognition module predicts the route based on the obstacle volume, the obstacle movement vector and the robot movement speed to obtain the encounter risk index between the obstacle and the robot. The obstacle avoidance recognition module generates a temporary obstacle avoidance route based on the encounter risk index, wherein the temporary obstacle avoidance route includes a deceleration avoidance route and a detour avoidance route.

[0025] As a preferred embodiment of the present invention, the method for the obstacle avoidance recognition module to generate a temporary obstacle avoidance route is:

[0026] When the obstacle avoidance recognition module performs route prediction, if the obstacle and the robot come into contact, it is recorded as a complete encounter. If the obstacle and the robot do not come into contact, the closest distance between the obstacle and the robot is calculated and recorded as the danger distance. The set safety standard and the danger distance are proportionally calculated, and the obtained ratio is recorded as the encounter danger index. The obstacle avoidance recognition module compares the encounter danger index with the set threshold and determines whether to slow down and avoid or detour according to the comparison result.

[0027] As a preferred embodiment of the present invention, the hybrid decision module plans the stay time of each sub-area in the path according to the final priority index, and after generating a temporary obstacle avoidance route, recalculates the moving speed in the sub-area according to the temporary obstacle avoidance route.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] 1. In the present invention, during the patrol process of the patrol robot, the abnormal situations obtained by the patrol robot are counted and the credibility of the abnormal situations is judged. If it is confirmed as an alarm situation, an early warning output is directly performed, thereby realizing the patrol alarm function. When it is not confirmed as an alarm situation, the abnormal situation is retained, and the patrol weight of the corresponding area is modified according to the abnormal situation, thereby increasing the patrol robot's stay time in the area where the abnormal situation exists, improving the continuous observation time of the abnormal situation, and avoiding the omission of abnormal situations during the patrol process.

[0030] 2. In the present invention, during the patrol process, different weights are assigned to patrolled paths and unpatrolled paths, and the patrol time distribution planning of the entire patrol path is realized in combination with the weight of the occurrence of abnormal situations, so that the patrol speed when patrolling on the set path can be autonomously and dynamically adjusted, thereby improving the security intensity of key areas and realizing the optimal allocation of patrol resources.

[0031] 3. In the present invention, when performing obstacle avoidance operations, vector analysis is performed on obstacles to obtain the movement of obstacles, and interference planning is performed through the patrol robot's own movement path, so as to dynamically evaluate the impact of obstacles and temporarily plan obstacle avoidance routes. At the same time, secondary dynamic adjustment is performed based on the temporary obstacle avoidance route and patrol time distribution planning, so that the patrol robot can still maintain efficient and safe patrols in complex and changing environments, thereby improving the flexibility of the patrol path. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] To facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0033] Figure 1 is a system block diagram of the present invention;

[0034] Figure 2 It is a system flow chart of the present invention. DETAILED DESCRIPTION

[0035] The following is a clear and complete description of the technical solutions of the present invention in conjunction with the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0036] Example 1:

[0037] See also Figure 1 - Figure 2As shown, the patrol robot system with path planning and dynamic obstacle avoidance functions includes a route interval management module, an abnormal area recognition module, a path planning module, a dynamic recognition module, an obstacle avoidance recognition module and a hybrid decision module. The route interval management module is used to record the patrol path of the robot. When recording the patrolled path, the route interval management module operates by obtaining a preset initial route and records the passed part as the patrolled path;

[0038] After obtaining the preset initial route, the route interval management module divides the initial route into multiple sub-areas and records the time when the robot leaves the route contained in the sub-area as the passing time of the waypoint, thereby obtaining the patrolled path and the passing time of each waypoint in the patrolled path and sending them to the path planning module;

[0039] The abnormal area recognition module counts the routes passed by the patrol robot and counts the abnormal points determined on the routes;

[0040] The method for the abnormal area identification module to obtain route abnormal points is as follows:

[0041] The abnormal area recognition module uses 3D structured light cameras, visible light cameras, and infrared cameras to collect images every time the robot reaches a sub-area. It then performs recognition and analysis based on the collected images. It uses a preset recognition algorithm to obtain dynamic pixel areas and temperature-prominent pixel areas in the images. It then uses the dynamic recognition algorithm to identify the dynamic pixel areas and compares the recognition results with the set intrusion sample database to determine the intrusion credibility. If the intrusion credibility reaches the set threshold, it is recorded as an intrusion warning. If it does not reach the set threshold, it is recorded as an abnormal point.

[0042] The temperature-prominent pixel area is compared with the infrared recognition database to confirm the credibility of the fire hazard area and the human activity area. If the credibility reaches the set threshold, a danger warning is generated. If it does not reach the set threshold, it is recorded as an abnormal point.

[0043] The abnormal area recognition unit divides the patrolled path into multiple sub-areas and obtains the sub-area where the abnormal point is located;

[0044] The path planning module obtains data through the route interval management module. After obtaining the passing time of the patrolled path and the path point, the patrolled path is compared with the set initial path to obtain the non-patrolled path. The path planning module assigns non-repetition indicators to all non-patrolled paths. The initial path obtained by the path planning module is a circular path. After each round of patrol, all paths are changed to non-patrolled paths. The path planning module assigns indicators to all non-patrolled paths and patrolled paths, thereby completing the generation of route patrol repetition indicators. The indicator of the non-patrolled path is greater than the indicator of the patrolled path.

[0045] The abnormal area identification module obtains data related to abnormal areas and generates route patrol enhancement indicators. After obtaining abnormal points, the path planning module records the areas to which the abnormal points belong, numbers the areas to which they belong, and creates an empty set. For each abnormal point, a set of data is added to the set. The path planning module counts all areas with abnormal points in the set of the set initial path and records the number of data in the set corresponding to the area as the patrol enhancement indicator.

[0046] The route planning module prioritizes routes based on patrol enhancement indicators and abnormal area-related indicators. Specifically, the module selects each sub-area in the set initial route and quantitatively superimposes the non-repetition indicator and patrol enhancement indicator of each sub-area. Different weight values ​​are assigned to the non-repetition indicator and patrol enhancement indicator during the quantitative superposition, so that the patrol enhancement indicator has a greater weight than the non-repetition indicator, thereby obtaining the final priority indicator.

[0047] The path planning module obtains the priority indicators of all sub-areas and the patrol speed of the set initial path. It calculates the initial patrol time T based on the length of the initial path and the patrol speed. It allocates time according to the final priority indicator to obtain the allocated time ti for each sub-area, where i is the sub-area number, ti is positively correlated with the final priority indicator, and ∑ti = T;

[0048] The dynamic recognition module scans through sensors to obtain obstacle information. The obstacle information obtained by the dynamic recognition module includes obstacle volume and obstacle movement vector. The dynamic recognition module obtains obstacle volume by creating and identifying three-dimensional coordinates through 3D point clouds. The dynamic recognition module obtains obstacle movement vector by performing difference calculation on the three-dimensional coordinates of two consecutive 3D point clouds to obtain the movement distance and movement direction of the obstacle. At the same time, the obstacle movement speed is calculated based on the time interval and distance between the two consecutive creations. The movement speed, distance, direction and obstacle volume are recorded as obstacle obstruction information.

[0049] After obtaining the obstacle information, the obstacle avoidance recognition module calculates the obstacle information on the patrol route and generates a temporary obstacle avoidance route. The specific method is as follows:

[0050] S1: The obstacle avoidance module predicts the route based on the obstacle volume, obstacle movement vector, and robot movement speed;

[0051] S2: When the obstacle avoidance module predicts a route, if the robot makes contact with the obstacle, it records it as a complete encounter. If the robot does not make contact, it calculates the closest distance between the obstacle and the robot and records it as the danger distance. It then calculates the ratio of the set safety standard to the danger distance and records the resulting ratio as the encounter danger index.

[0052] S3: The obstacle avoidance recognition module compares the encounter risk index with a set threshold. If the encounter risk index is greater than the set threshold, a detour is performed. If the encounter risk index is not greater than the set threshold, a deceleration is performed. The deceleration avoidance route and the detour avoidance route are recorded as temporary obstacle avoidance routes, and the temporary obstacle avoidance routes are sent to the hybrid decision module.

[0053] The hybrid decision module obtains the route results of route priority planning, obtains the residence time of each sub-area in the path, and uses the route results as a benchmark to synthesize the temporary obstacle avoidance route. The moving speed in the sub-area is recalculated according to the temporary obstacle avoidance route to obtain the dynamic route result, so that the total time for the patrol robot to complete a complete path patrol is fixed. Every time the patrol robot encounters an obstacle and needs to slow down to avoid or detour to avoid it, the hybrid decision module will continuously generate a temporary route and use the dynamic route result as the final patrol route.

[0054] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to specific embodiments. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A patrol robot system with path planning and dynamic obstacle avoidance functions, characterized in that: It includes a route interval management module, which is used to record the patrol path of the robot and obtain the patrolled path and the passing time of each path point in the patrolled path; An abnormal area identification module, which counts the route passed by the patrol robot and counts the abnormal points determined on the route. The abnormal area identification unit divides the patrolled path into multiple sub-areas and obtains the sub-areas where the abnormal points are located; A route planning module, which obtains data from the route interval management module and generates a route patrol repetition index. It obtains abnormal area related data from the abnormal area identification module and generates a route patrol enhancement index. The route planning module performs route priority planning based on the route patrol enhancement index and abnormal area related indicators. A dynamic recognition module, which scans with sensors to obtain obstacle information and dynamically evaluates the obstacle information to obtain obstacle obstruction information; An obstacle avoidance recognition module, after obtaining obstacle obstruction information, calculates the obstruction situation of the obstacle obstruction information on the patrol route, generates a temporary obstacle avoidance route, and sends the temporary obstacle avoidance route to the hybrid decision module; A hybrid decision module obtains the route results of route priority planning, and uses the route results as a benchmark to synthesize the temporary obstacle avoidance route to obtain a dynamic route result. The hybrid decision module continuously generates temporary routes and uses the dynamic route results as the final patrol route.

2. The patrol robot system with path planning and dynamic obstacle avoidance functions according to claim 1, characterized in that: The route interval management module operates by obtaining a preset initial route when recording the patrolled route, and records the passed portion as the patrolled route; After obtaining the preset initial route, the route interval management module divides the initial route into multiple sub-areas and records the time point when the robot leaves the route included in the sub-area as the passing time of the path point.

3. The patrol robot system with path planning and dynamic obstacle avoidance functions according to claim 1, characterized in that: The method for the abnormal area identification module to obtain route abnormal points is: The abnormal area recognition module collects images through the 3D structured light camera, visible light camera and infrared camera every time the robot arrives at a sub-area, performs recognition and analysis based on the collected images, obtains the dynamic pixel area and the temperature-prominent pixel area in the image through a preset recognition algorithm, and recognizes the dynamic pixel area through the dynamic recognition algorithm. The recognition result is compared with the set intrusion sample database to determine the intrusion credibility. If the intrusion credibility reaches the set threshold, it is recorded as an intrusion warning. If it does not reach the set threshold, it is recorded as an abnormal point. The temperature-prominent pixel areas are compared with the infrared recognition database to confirm the credibility of fire hazard areas and personnel activity areas. If the credibility reaches the set threshold, a danger warning is generated. If it does not reach the set threshold, it is recorded as an abnormal point.

4. The patrol robot system with path planning and dynamic obstacle avoidance functions according to claim 1, characterized in that: After obtaining the patrolled paths and the elapsed time of the path points, the path planning module compares the patrolled paths with the set initial paths to obtain non-patrolled paths. The path planning module assigns a non-repetition indicator to all non-patrolled paths, wherein the initial paths obtained by the path planning module are cyclic paths. After each patrol, all paths are changed to non-patrolled paths. After obtaining an anomaly point, the path planning module records the area to which the anomaly point belongs, numbers the area to which it belongs, and creates an empty set. For each anomaly point, a set of data is added to the set. The path planning module counts all areas with anomalies in the set initial path, and records the number of data in the set corresponding to the area as the patrol enhancement indicator.

5. The patrol robot system with path planning and dynamic obstacle avoidance functions according to claim 4 is characterized in that: The path planning module selects each sub-area in the set initial path, and quantifies and superimposes the non-repetition index and patrol enhancement index of each sub-area to obtain the final priority index; The path planning module obtains the priority indicators of all sub-areas and the patrol speed of the set initial path, calculates the initial patrol time T based on the length of the initial path and the patrol speed, and allocates time through the final priority indicator to obtain the allocated time ti for each sub-area, where i is the number of the sub-area, ti is positively correlated with the final priority indicator, and Σti=T.

6. The patrol robot system with path planning and dynamic obstacle avoidance functions according to claim 1, characterized in that: The obstacle information obtained by the dynamic recognition module includes the obstacle volume and the obstacle movement vector. The method for the dynamic recognition module to obtain the obstacle volume is to create and identify three-dimensional coordinates through 3D point clouds. The method for the dynamic recognition module to obtain the obstacle movement vector is to calculate the difference between the three-dimensional coordinates of the 3D point clouds created twice in succession to obtain the movement distance and movement direction of the obstacle. At the same time, the obstacle movement speed is calculated based on the time interval and distance between the two consecutive creations.

7. The patrol robot system with path planning and dynamic obstacle avoidance functions according to claim 1, characterized in that: The obstacle avoidance recognition module predicts the route based on the obstacle volume, obstacle movement vector and robot movement speed to obtain the encounter risk index between the obstacle and the robot. The obstacle avoidance recognition module generates a temporary obstacle avoidance route based on the encounter risk index, where the temporary obstacle avoidance route includes a deceleration avoidance route and a detour avoidance route.

8. The patrol robot system with path planning and dynamic obstacle avoidance functions according to claim 1, characterized in that: The method for the obstacle avoidance recognition module to generate a temporary obstacle avoidance route is: When the obstacle avoidance recognition module performs route prediction, if the obstacle and the robot come into contact, it is recorded as a complete encounter. If the obstacle and the robot do not come into contact, the closest distance between the obstacle and the robot is calculated and recorded as the danger distance. The set safety standard and the danger distance are proportionally calculated, and the obtained ratio is recorded as the encounter danger index. The obstacle avoidance recognition module compares the encounter danger index with the set threshold and determines whether to slow down and avoid or detour according to the comparison result.

9. The patrol robot system with path planning and dynamic obstacle avoidance functions according to claim 1, characterized in that: The hybrid decision module plans the stay time of each sub-area in the path according to the final priority index, and after generating a temporary obstacle avoidance route, recalculates the moving speed in the sub-area according to the temporary obstacle avoidance route.

Images