Autonomous robot active obstacle avoidance method based on virtual formation

Abstract

The invention provides an autonomous robot active obstacle avoidance method based on virtual formation, and the method comprises the following steps: 1, enabling a robot to carry out the local movement along with a path point, and carrying out the active obstacle avoidance through employing a virtual formation method based on an artificial potential field during the period; 2, a virtual formation of a regular polyhedron structure is arranged around the robot to protect safe advancing of the robot; 3, the robot senses an unexpected obstacle in front, surrounding members of the virtual formation move towards the obstacle in advance and bypass the obstacle by using an artificial potential field method, and when artificial potential field guide / repulsive force between the members of the virtual formation and between the obstacles is balanced, the virtual formation stops moving; 4, the robot calculates the moving cost between each virtual formation member and the path point, wherein the moving cost depends on the distance and the steering angle; 5, the robot selects the position of the virtual member with the minimum moving cost as an obstacle avoidance direction, bypasses the obstacle and returns to the original path; and 6, after dynamic obstacle avoidance is completed, the robot rearranges the virtual formation to prevent the next dynamic obstacle. According to the method, a virtual formation structure and an artificial potential field method are combined, the obstacle avoidance orientation of a dynamic obstacle is rapidly judged and selected, and the global path is prevented from shifting by pre-planning a local path after obstacle avoidance.

Description

Technical Field

[0001] The invention relates to an active obstacle avoidance method for an autonomous robot, and belongs to the technical field of robot navigation. Background Art

[0002] Before the robot moves, it can obtain geographic data from the control center and follow the path points to avoid obstacles. However, when unexpected dynamic obstacles appear, the robot is generally required to use short-range perception technologies such as imaging to perform temporary active obstacle avoidance. However, sudden obstacle avoidance movement will cause the robot to deviate from the original set path. In a weak communication environment, the global travel cost will increase at best, and the device will be lost at worst.

[0003] In addition, robots are expensive, and dynamic obstacles such as large animal groups threaten their safety. Widely studied navigation methods are usually based on passive obstacle avoidance on static terrain. If a dynamic obstacle appears, the base station can sense it in advance, but it needs to update the robot's global environmental information through information exchange, which is difficult to implement under weak communication conditions. If the robot only uses short-range perception technology to take active obstacle avoidance, sudden obstacle avoidance behavior can easily lead to a deviation in the global path, and it is difficult to temporarily execute a high-complexity algorithm to adjust the path in the short gap between perception and obstacle avoidance. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present invention proposes an active obstacle avoidance method based on virtual formation, which quickly selects a reasonable obstacle avoidance orientation of the robot through the virtual potential field force structure, and combines the pre-planned local path after obstacle avoidance, thereby guiding the robot to flexibly avoid dynamic obstacles, ensuring its travel safety while maintaining the global path cost. This method combines the virtual formation structure with the artificial potential field method to quickly judge and select the obstacle avoidance orientation of dynamic obstacles, and prevents the deviation of the global path by pre-planning the local path after obstacle avoidance.

[0005] The main technical solutions adopted in the present invention are: Step 1-1. The robot follows the path point for local movement, and actively avoids obstacles using a virtual formation method based on artificial potential field; Step 1-2. The robot arranges a virtual formation of regular polyhedron structures around it to protect the robot's safe movement; Step 1-3. The robot senses an unexpected obstacle in front of it, and the members of the surrounding virtual formation move toward the obstacle in advance, using the artificial potential field method to bypass the obstacle. When the artificial potential field attraction / repulsion between the members of the virtual formation and the obstacle reaches a balance, the virtual formation stops moving; Step 1-4. The robot calculates the movement cost between each virtual member and the waypoint (depending on the distance and the turning angle); Step 1-5. The robot selects the virtual member position with the lowest moving cost as the obstacle avoidance direction, bypasses the obstacle, and returns to the original path; Step 1-6. After completing dynamic obstacle avoidance, the robot rearranges the virtual formation to prevent the next dynamic obstacle.

[0006] Preferably, in step 1-2, the virtual formation structure is used to quickly determine the obstacle avoidance direction, and the artificial potential field of the Platonic polyhedron structure is used to evenly deploy the obstacle avoidance direction around the irregularly shaped obstacle to guide the robot to avoid the obstacle reasonably.

[0007] Preferably, in steps 1-4, the obstacle avoidance behavior of the virtual formation represents that each virtual member moves to the boundary of the two-dimensional cross-section of the front of the obstacle. After each member of the virtual formation reaches the boundary, a point directly in front of it has a line of sight (LOS) link with the waypoint of the local path. Specific explanation of the virtual formation obstacle avoidance process using artificial potential field ( Figure 6 ), the obstacle has 6 faces: front, back, top, bottom, left, and right: 1) The target point (local path waypoint) applies virtual gravity to the virtual member, and the formation members move toward the target point in order. In the figure, member 1 moves first. After reaching the left side of the obstacle, the obstacle applies virtual repulsion to the virtual member, so that the potential field force of member 1 and the obstacle is balanced. Member 1 moves in the direction of the combined force of gravity and repulsion. After reaching the position that satisfies the line-of-sight link, the target point stops applying virtual gravity, and member 1 stops; 2) Member 2 moves, and after reaching the left side of the obstacle, it moves downward in the direction of the combined force. After reaching point A, the line-of-sight link condition is met, but member 1 applies virtual repulsion to it, so it moves straight ahead (the robot cannot obtain the shape of the back of the obstacle, so it only moves straight ahead after reaching the line-of-sight link). After the potential field forces of member 1, member 2 and the obstacle are balanced, member 2 stops. Similarly, member 3 moves behind the obstacle and reaches point B, meeting the line-of-sight link condition. Members 1 and 2 exert virtual gravity on it, so it moves forward until the potential field force reaches equilibrium; 3) Member 4 reaches the left side of the obstacle and moves upward in the direction of the resultant force. When it reaches the top of the obstacle, it does not meet the line-of-sight link condition. At this time, the potential field force it receives is the virtual attraction / repulsion of members 1-3, the repulsion of the obstacle, and the gravity of the target point. Among them, the gravity of the target point has the highest priority, but it stops after meeting the line-of-sight link condition. Member 4 moves in the direction of the resultant force and stops after the potential field force is balanced.

[0008] Beneficial effects: The present invention provides an autonomous robot active obstacle avoidance method based on virtual formation. The method combines the virtual formation structure with the artificial potential field method to quickly judge and select the obstacle avoidance direction of dynamic obstacles, and prevents the deviation of the global path by pre-planning the local path after obstacle avoidance. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 A virtual formation of a regular tetrahedron structure is arranged around the robot of the present invention; Figure 2 The robot of the present invention utilizes the artificial potential field method to bypass obstacles; Figure 3 Calculating the movement cost between each virtual member and the path point for the robot of the present invention; Figure 4 The robot of the present invention selects the virtual member position with the minimum moving cost as the obstacle avoidance orientation, bypasses the obstacle, and returns to the original path; Figure 5 Each virtual member of the present invention moves to the boundary of the two-dimensional cross section of the front side of the obstacle; Figure 6 This is a diagram of the virtual formation obstacle avoidance process using an artificial potential field in the present invention. DETAILED DESCRIPTION

[0010] In order to enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application are described below. Obviously, the described embodiments are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this application.

[0011] An active obstacle avoidance method based on virtual formation includes the following steps: Step 1-1. The robot follows the path point for local movement, and actively avoids obstacles using a virtual formation method based on artificial potential field; Step 1-2. The robot arranges a virtual formation of regular polyhedron structures around it to protect the robot's safe movement, such as Figure 1 As shown; Step 1-3. The robot senses an unexpected obstacle in front of it, and the members of the surrounding virtual formation move toward the obstacle in advance and use the artificial potential field method to bypass the obstacle. When the artificial potential field attraction / repulsion between the members of the virtual formation and the obstacle reaches a balance, the virtual formation stops moving. Figure 2 As shown; Step 1-4. The robot calculates the movement cost between each virtual member and the path point (depending on the distance and turning angle), such as Figure 3 As shown; Step 1-5. The robot selects the virtual member position with the smallest moving cost as the obstacle avoidance direction, bypasses the obstacle, and returns to the original path, such as Figure 4 As shown; Step 1-6. After completing dynamic obstacle avoidance, the robot rearranges the virtual formation to prevent the next dynamic obstacle.

[0012] Preferably, in step 1-2, the virtual formation structure is used to quickly determine the obstacle avoidance direction, and the artificial potential field of the Platonic polyhedron structure is used to evenly deploy the obstacle avoidance direction around the irregularly shaped obstacle to guide the robot to avoid the obstacle reasonably.

[0013] Preferably, in steps 1-4, the obstacle avoidance behavior of the virtual formation means that each virtual member moves to the boundary of the two-dimensional cross-section of the obstacle front (such as Figure 5 As shown in the figure, after the members of each virtual formation reach the boundary, a certain point directly in front of them has a line of sight (LOS) link with the waypoint of the local path.

[0014] Specifically explain the virtual formation obstacle avoidance process using artificial potential field (such as Figure 6 As shown in the figure, the obstacle has six faces: front, back, top, bottom, left, and right: 1) The target point (local path waypoint) applies virtual gravity to the virtual member, and the formation members move toward the target point in order. In the figure, member 1 moves first. After reaching the left side of the obstacle, the obstacle applies virtual repulsion to the virtual member, so that the potential field force of member 1 and the obstacle is balanced. Member 1 moves in the direction of the combined force of gravity and repulsion. After reaching the position that satisfies the line-of-sight link, the target point stops applying virtual gravity, and member 1 stops; 2) Member 2 moves, and after reaching the left side of the obstacle, it moves downward in the direction of the combined force. After reaching point A, the line-of-sight link condition is met, but member 1 applies virtual repulsion to it, so it moves straight ahead (the robot cannot obtain the shape of the back of the obstacle, so it only moves straight ahead after reaching the line-of-sight link). After the potential field forces of member 1, member 2 and the obstacle are balanced, member 2 stops. Similarly, member 3 moves behind the obstacle and reaches point B, meeting the line-of-sight link condition. Members 1 and 2 exert virtual gravity on it, so it moves forward until the potential field force reaches equilibrium; 3) Member 4 reaches the left side of the obstacle and moves upward in the direction of the resultant force. When it reaches the top of the obstacle, it does not meet the line-of-sight link condition. At this time, the potential field force it receives is the virtual attraction / repulsion of members 1-3, the repulsion of the obstacle, and the gravity of the target point. Among them, the gravity of the target point has the highest priority, but it stops after meeting the line-of-sight link condition. Member 4 moves in the direction of the resultant force and stops after the potential field force is balanced. Example

[0015] The following is an example of an inspection scenario based on an automated guided vehicle (AGV). The AGV moves along the path points and encounters moving obstacles during the process. The artificial potential field is used for virtual formation obstacle avoidance. The target point (local path waypoint) applies virtual gravity to the virtual members, and the formation members move toward the target point in sequence. After reaching the left side of the obstacle, the obstacle applies virtual repulsion to the virtual members, so that the potential field force of the first member and the obstacle is balanced. The first member moves in the direction of the combined force of gravity and repulsion. After reaching the position that satisfies the line-of-sight link, the target point stops applying virtual gravity and then stops; then the second formation member moves, and after reaching the left side of the obstacle, it moves downward in the direction of the combined force. After reaching the point that satisfies the line-of-sight link condition, the first member applies virtual repulsion to it, so it moves straight ahead (the robot cannot obtain the shape of the back of the obstacle, so it only moves straight ahead after reaching the line-of-sight link). After the potential field forces of the two members and the obstacle are balanced, the second member stops. Similarly, the third member moves behind the obstacle and reaches the point that satisfies the line-of-sight link condition. The first two members exert virtual gravitational force on it, so it moves forward until the potential field force reaches a balance. The fourth member reaches the left side of the obstacle and moves upward in the direction of the resultant force. After reaching the top of the obstacle, it does not meet the line-of-sight link condition. At this time, the potential field force it receives is the virtual attraction / repulsion of the first three members, the repulsion of the obstacle, and the gravitational force of the target point. Among them, the gravitational force of the target point has the highest priority, but it stops after the line-of-sight link condition is met. The fourth member moves in the direction of the resultant force and stops after the potential field force is balanced.

[0016] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. An autonomous robot active obstacle avoidance method based on virtual formation, characterized in that: The steps include: Step 1-1. The robot follows the path point for local movement, and actively avoids obstacles using a virtual formation method based on artificial potential field; Step 1-2. The robot arranges a virtual formation of regular polyhedron structures around it to protect the robot's safe movement; Step 1-3. The robot senses an unexpected obstacle in front of it, and the members of the surrounding virtual formation move toward the obstacle in advance and use the artificial potential field method to bypass the obstacle. When the artificial potential field attraction / repulsion between the members of the virtual formation and the obstacle reaches a balance, the virtual formation stops moving; Step 1-4. The robot calculates the movement cost between each virtual team member and the path point. The movement cost depends on the distance and the turning angle. Step 1-5. The robot selects the virtual member position with the lowest moving cost as the obstacle avoidance direction, bypasses the obstacle, and returns to the original path; Step 1-6. After completing dynamic obstacle avoidance, the robot rearranges the virtual formation to prevent the next dynamic obstacle.

2. The method for autonomous robot active obstacle avoidance based on virtual formation according to claim 1, characterized in that: In step 1-2, the virtual formation structure is used to quickly determine the obstacle avoidance direction, and the artificial potential field of the Platonic polyhedron structure is used to evenly deploy the obstacle avoidance direction around the irregularly shaped obstacles to guide the robot to avoid obstacles reasonably.

3. The method for autonomous robot active obstacle avoidance based on virtual formation according to claim 1, characterized in that: In steps 1-4, the obstacle avoidance behavior of the virtual formation means that each virtual member moves to the boundary of the two-dimensional cross-section of the front of the obstacle. After each member of the virtual formation reaches the boundary, a point directly in front of it has a line-of-sight link with the waypoint of the local path.

Images