WO2020220604A1

WO2020220604A1 - Real-time obstacle avoidance method and obstacle avoidance system for dynamic obstacles in multi-agv system

Info

Publication number: WO2020220604A1
Application number: PCT/CN2019/112408
Authority: WO
Inventors: 钱晓明; 楼佩煌; 张硕
Original assignee: 南京航空航天大学
Priority date: 2019-04-30
Filing date: 2019-10-22
Publication date: 2020-11-05
Also published as: CN110221601A

Abstract

A real-time obstacle avoidance method and an obstacle avoidance system for dynamic obstacles in a multi-AGV system. To address such problems as obstacles impacting the normal travel of AGVs or even causing collisions when a static or dynamic obstacle is present on a planned route of an AGV during an operating process of the AGV, a method is provided that is able to update the route in real time when an obstacle is present and avoid the obstacle, and thus travel normally. In said method, on the basis of a route planning algorithm, and by means of an obstacle matrix method, a sensor is used in the traveling process of an AGV to detect in real time whether an obstacle is present within a detection radius from a current position. If an obstacle is present, then the AGV moves to a position point closest to a target point, until the AGV moves to the target point.

Description

Multi-AGV system dynamic obstacle real-time obstacle avoidance method and obstacle avoidance system

Technical field

The invention relates to a multi-AGV system, in particular to a real-time obstacle avoidance method and an obstacle avoidance system for dynamic obstacles of the multi-AGV system.

Background technique

Nowadays, with the rapid development of modern industry, AGV systems applied to intelligent logistics have been widely used in flexible production lines and warehouse logistics. Since AGV can adjust transportation plans and increase system flexibility according to system instructions, using AGV as a transportation carrier can ensure transportation efficiency and reduce costs. In the digital workshop, the product types are diverse, the production cycle is strict, and the logistics requirements are complex. The AGV in a complex environment will inevitably encounter static obstacles and dynamic obstacles during operation, which seriously affects transportation efficiency. AGV is more stable, efficient and safe operation, and it is necessary to design efficient obstacle avoidance function for AGV.

Summary of the invention

In order to solve the technical problems raised by the above-mentioned background technology, the purpose of the present invention is to solve the problem of static or dynamic obstacles on the planned route of the AGV during operation, and the obstacles will affect the normal driving of the AGV or even collisions. When there are obstacles, the route is updated in real time to avoid the obstacles and then drive normally, and finally reach the target point. A method of real-time obstacle avoidance with multiple AGV systems dynamic obstacles.

The present invention provides a real-time obstacle avoidance method for dynamic obstacles in a multi-AGV system. The method is based on a path planning algorithm. During the driving of the AGV, through the obstacle matrix method, sensors are used to detect in real time whether there are obstacles in the current position within the detection radius. If there is an obstacle, the AGV will move to the position closest to the target point until the AGV moves to the target point.

Further, the method specifically includes the following steps:

Step 1. Determine the target point of the handling task and apply the path planning algorithm to generate the pre-driving path;

Step 2. The AGV moves according to the pre-driving path;

Step 3. Judge whether it has reached the end point, if yes, end the path planning, otherwise go to step 4;

Step 4. Using sensors, establish an obstacle detection matrix S based on the sensor detection radius, and determine the value of S. If the detection matrix value is 1, go to step 5, otherwise go to step 2;

Step 5: Calculate the distance between the edge of the sensor detection range corresponding to the value 0 in the detection matrix and the target point to obtain the minimum value, and move the AGV to this position;

Step 6, Judge the value of S, if one of the sectors takes the value 1, go to step 5, otherwise go to step 7;

Step 7: Judge whether the location point is reached, if yes, go to step 1, otherwise go to step 5.

Further, the step 4 is specifically:

Step 4.1, the detection area formed by a circle with the AGV as the center and the detection radius of the sensor as the radius, divide the detection area into several sectors;

Step 4.2, labeled sequentially as each sector S _i, which is representative of the

value

0 or 1, 0 is not detected within the current sector obstacle, represents an obstacle is detected, the distribution list of the obstacle S = [S ₁ , S ₂ ,..., S _M ], M is the number of sectors.

Further, the step 5 is specifically:

Step 5.1: Calculate the distance d _i from the location point p _i where the sector center line with a value of 0 intersects the edge of the sensor detection range to the target point in turn, and use it as a cost function for determining the best path;

Step 5.2, calculate d _i corresponding to the minimum position of the point p _{i min,} the p _{i min} point as a new target location, and then generates a path planning algorithm based on a pre-running path of travel.

Further, in step 4.1, the more the number of sectors, the more accurate the detection accuracy.

Furthermore, a sensor is placed on each of the four diagonals of the AGV, each sensor is divided into two independent detection areas, and the obstacle distribution in the surrounding environment of the AGV is sensed in real time. In step 4.1, 8 sectors are formed.

As a preference, the path planning algorithm is a local particle swarm path planning algorithm based on a static obstacle matrix.

The present invention also provides a multi-AGV system dynamic obstacle real-time obstacle avoidance system. The obstacle avoidance system includes a path planning algorithm module, a sensor detection module, and an obstacle avoidance algorithm module; wherein the path planning algorithm module is used to generate target-based The sensor detection module is used to detect the position of the obstacle in real time, and the obstacle avoidance algorithm module forms a driving path to avoid the obstacle based on the cost function when there is an obstacle in the detection area.

Further, the sensor module detection AGV of a circle, the radius of the detection area of the sensor to detect a radius of the circle formed by the detection area divided into a plurality of sectors; order mark each sector S _i, its

value

0 or 1, 0 means that no obstacle is detected in the current sector, 1 means that an obstacle is detected, there is an obstacle distribution list S=[S ₁ , S ₂ ,..., S _M ], M is the sector The number of.

Further, the sensor detection module establishes an obstacle detection matrix S based on the sensor detection radius, and the judged S takes a value. If the detection matrix value is 1, the obstacle avoidance algorithm module calculates the real-time obstacle avoidance driving path, If the detection matrix value is 0, the AGV generates a driving path based on the target point based on the path planning algorithm module.

Further, if there is a sector with a value of 1, the obstacle avoidance algorithm module calculates the distance between the edge of the sensor detection range corresponding to the value 0 in the detection matrix and the target point to obtain the minimum value, and moves the AGV to this Position; then judge whether to reach the location point, if yes, then plan the driving path based on the target point through the path planning algorithm module, the AGV continues to drive along the driving path, if not, plan the obstacle avoidance path through the obstacle avoidance algorithm module and then follow Its driving.

Further, the greater the number of sectors divided in the sensor detection module, the more accurate the detection accuracy.

As a preference, the sensor detection module contains 4 sensors, one sensor is placed on each of the 4 diagonals of the AGV, each sensor is divided into 2 independent detection areas, real-time perception of the distribution of obstacles around the AGV , Forming 8 sectors.

Further, the obstacle avoidance algorithm module calculates the distance between the edge of the sensor detection range corresponding to the value 0 in the detection matrix and the target point to obtain the minimum value, the position point corresponding to the minimum value, and use the position point as the new target The location points are driven on the pre-driving route generated according to the route planning algorithm.

As a preference, the path planning algorithm module uses a local particle swarm algorithm to calculate the pre-driving path from the current location point to the target point based on the global static obstacle matrix.

The present invention has the following beneficial effects:

In addition to static obstacles, it also fully considers the impact of dynamic obstacles on path planning. The obstacle avoidance strategy is simple and easy to implement, the algorithm complexity is low, and the deployment cost is low, and a good real-time obstacle avoidance effect for dynamic obstacles is achieved.

Description of the drawings

Figure 1 is an AGV operating environment of an embodiment of the present invention.

Fig. 2 is a schematic diagram of sensor distribution according to an embodiment of the present invention.

Figure 3 is a diagram of a detection model of an embodiment of the present invention.

Figure 4 is a collision-free obstacle avoidance path diagram of an AGV in a dynamic obstacle environment according to an embodiment of the present invention.

Detailed ways

The technical solution of the present invention will now be fully described in conjunction with the drawings. The following description is only a part of the implementation cases of the present invention, not all. Based on the implementation cases in the present invention, all other implementation cases obtained by those skilled in the art without creative work fall within the scope of protection of the rights of the present invention.

Example 1

A real-time obstacle avoidance method for dynamic obstacles in a multi-AGV system. The method is based on a path planning algorithm. During the driving of the AGV, through the obstacle matrix method, sensors are used to detect in real time whether there are obstacles in the current position within the detection radius. If there is an obstacle, the AGV moves to the position closest to the target point until the AGV moves to the target point.

Further, the method specifically includes the following steps:

Step 2. The AGV moves according to the pre-driving path;

Further, the step 4 is specifically:

Step 4.1, the detection area formed by a circle with the AGV as the center and the detection radius of the sensor as the radius, the detection area is equally divided into several sectors;

value

Further, the step 5 is specifically:

In the AGV operating environment as shown in Figure 1, there are static obstacles such as production equipment and construction facilities, and dynamic obstacles such as other AGVs in the workshop. The AGV is equipped with a sensor capable of detecting obstacles, as shown in Figure 2, where v _a is the current direction of movement of the AGV; L is the detection range radius of the sensor; s _i is the sensor detection sector, and its value is 0 or 1, 0 means that no obstacle is detected in the current sector, 1 means that an obstacle is detected; the total number of sectors is 8, the larger the value, the higher the resolution of the obstacle detection, and the AGV adopts obstacle avoidance measures to plan a new The more precise the path. Then there is a list of obstacle distribution: S=[s ₁ , s ₂ ,..., s ₈ ].

The obstacle avoidance algorithm process is as follows:

Step 1 Determine the target point of the handling task, and apply the path planning algorithm to generate the pre-driving path.

Step 2 The AGV moves according to the pre-driving path and detects obstacles in real time.

Step 3 determines whether it has reached the end point, if yes, end the path planning, otherwise go to step 4.

Step 4 judges the value of S, if one of the sectors has a value of 1, go to step 5, otherwise go to step 2.

Step 5: Calculate the distance d _i from the location point p _i where the sector center line with a value of 0 intersects the edge of the sensor detection range to the target point in turn, and use it as the cost function for determining the best path, which needs to satisfy s _i = 0 i ∈ [1, 8].

Step 6 The AGV moves to the position p _{i min} that is closest to the target point, di _min =min(d _i ), and detects obstacles in real time.

Step 7 judges the value of S, if one of the sectors has a value of 1, go to step 5, otherwise go to step 8.

Step 8: Judge whether the position point p _{i min} is reached, if yes, go to step 1, otherwise go to step 5.

In the example shown in Fig. 3, S=[1, 0, 0, 1, 0, 0, 1, 1]. Among them, s ₂ ＝s ₃ ＝s ₅ ＝s ₆ =0, s ₁ ＝s ₄ ＝s ₇ ＝s ₈ =1, p ₂ , p ₃ , p ₅ , and p ₆ are feasible candidate locations. The AGV needs to determine which of the points to be selected is the closest to the target point according to the obstacle avoidance strategy. In this example, p ₂ is the closest to the target point of the handling task, so select this point as the local target point and re-plan the path. In the process of the AGV moving from the current position to p ₂ , the value of S is determined at regular intervals, as described in step 7, through such a step-by-step iterative process, finally reaching the target point of the handling task. The collision-free path of the AGV in a dynamic obstacle environment is shown in Figure 4.

Example 2

value

In addition to static obstacles, the method and system provided by the present invention also fully consider the impact of dynamic obstacles on path planning. The obstacle avoidance strategy is simple and easy to implement, the algorithm complexity is low, and the deployment cost is low while achieving good dynamics. Real-time obstacle avoidance effect of obstacles.

The basic principles, main features and advantages of the present invention have been shown and described above. Those skilled in the industry should understand that the present invention is not limited by the above-mentioned embodiments. The above-mentioned embodiments and the description only illustrate the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will also have With various changes and improvements, the scope of protection claimed by the present invention is defined by the appended claims, specification and their equivalents.

Claims

A real-time obstacle avoidance method for dynamic obstacles in a multi-AGV system, characterized in that the method is based on a path planning algorithm. During the driving of the AGV, through the obstacle matrix method, a sensor is used to detect in real time whether the current position is within the detection radius Obstacles, if there are obstacles, the AGV moves to the position closest to the target point until the AGV moves to the target point.
A real-time obstacle avoidance method for dynamic obstacles in a multi-AGV system according to claim 1, wherein the method specifically includes the following steps:

Step 1. Determine the target point of the handling task and apply the path planning algorithm to generate the pre-driving path;

Step 2. The AGV moves according to the pre-driving path;

Step 3. Judge whether it has reached the end point, if yes, end the path planning, otherwise go to step 4;

Step 4. Using sensors, establish an obstacle detection matrix S based on the sensor detection radius, and determine the value of S. If the detection matrix value is 1, go to step 5, otherwise go to step 2;

Step 5: Calculate the distance between the edge of the sensor detection range corresponding to the value 0 in the detection matrix and the target point to obtain the minimum value, and move the AGV to this position;

Step 6, Judge the value of S, if one of the sectors takes the value 1, go to step 5, otherwise go to step 7;

Step 7: Judge whether the location point is reached, if yes, go to step 1, otherwise go to step 5.
The method for real-time obstacle avoidance of dynamic obstacles in multiple AGV systems according to claim 2, wherein the step 4 is specifically:

Step 4.1, the detection area formed by a circle with the AGV as the center and the detection radius of the sensor as the radius, divide the detection area into several sectors;

Step 4.2, labeled sequentially as each sector S i, which is representative of the value 0 or 1, 0 is not detected within the current sector obstacle, represents an obstacle is detected, the distribution list of the obstacle S = [S 1 , S 2 ,..., S M ], M is the number of sectors.
The method for real-time obstacle avoidance of dynamic obstacles in multiple AGV systems according to claim 3, wherein the step 5 is specifically:

Step 5.1: Calculate the distance d i from the location point p i where the sector center line with a value of 0 intersects the edge of the sensor detection range to the target point in turn, and use it as a cost function for determining the best path;

Step 5.2, calculate d i corresponding to the minimum position of the point p i min, the p i min point as a new target location, and then generates a path planning algorithm based on a pre-running path of travel.
A real-time obstacle avoidance method for dynamic obstacles in multiple AGV systems according to claim 3, wherein in step 4.1, the more the number of sectors, the more accurate the detection accuracy.
A real-time obstacle avoidance method for dynamic obstacles in a multi-AGV system according to claim 3, wherein a sensor is placed on each of the 4 diagonals of the AGV, each sensor is divided into 2 independent detection areas, and real-time sensing The distribution of obstacles in the surrounding environment of the AGV forms 8 sectors in step 4.1.
The real-time obstacle avoidance method for dynamic obstacles in a multi-AGV system according to claim 3, wherein the path planning algorithm is a local particle swarm path planning algorithm based on a static obstacle matrix.
A multi-AGV system dynamic obstacle real-time obstacle avoidance system is characterized in that the obstacle avoidance system includes a path planning algorithm module, a sensor detection module, and an obstacle avoidance algorithm module; wherein, the path planning algorithm module is used to generate target-based The sensor detection module is used to detect the position of the obstacle in real time, and the obstacle avoidance algorithm module forms a driving path to avoid the obstacle based on the cost function when there is an obstacle in the detection area.
The multi-AGV system dynamic obstacle real-time obstacle avoidance system according to claim 8, wherein the sensor detection module takes the AGV as the center of the circle, and the detection radius of the sensor is the detection area formed by the radius of the circle. divided into a number of sectors; each sector marked sequence S i, which is representative of the value 0 or 1, 0 is not an obstacle is detected within the current sector, represents an obstacle is detected, there is the obstacle distribution list S=[S 1 , S 2 ,..., S M ], M is the number of sectors.
The multi-AGV system dynamic obstacle real-time obstacle avoidance system according to claim 9, characterized in that the sensor detection module establishes an obstacle detection matrix S based on the sensor detection radius, and the judged S takes a value if all If the detection matrix value is 1, the obstacle avoidance algorithm module calculates a real-time obstacle avoidance driving path. If the detection matrix value is 0, the AGV generates a driving path based on the target point based on the path planning algorithm module.
The multi-AGV system dynamic obstacle real-time obstacle avoidance system according to claim 10, wherein if there is a sector with a value of 1, the obstacle avoidance algorithm module calculates the sensor corresponding to the value 0 in the detection matrix Detect the distance between the edge of the range and the target point, obtain the minimum value, and move the AGV to this position; then judge whether the position point is reached, and if so, plan the driving path based on the target point through the path planning algorithm module, the AGV Continue to drive along the driving path, if not, plan the obstacle avoidance path through the obstacle avoidance algorithm module and then drive along it.
The multi-AGV system dynamic obstacle real-time obstacle avoidance system according to claim 10, wherein the more the number of sectors divided in the sensor detection module, the more accurate the detection accuracy.
The multi-AGV system dynamic obstacle real-time obstacle avoidance system according to claim 10, wherein the sensor detection module includes 4 sensors, and one sensor is placed on each of the 4 diagonals of the AGV, each Each sensor is divided into 2 independent detection areas, real-time sensing the distribution of obstacles around the AGV, forming 8 sectors.
The multi-AGV system dynamic obstacle real-time obstacle avoidance system according to claim 10, wherein the obstacle avoidance algorithm module calculates the distance between the edge of the sensor detection range corresponding to the 0 value in the detection matrix and the target point. Distance, get the minimum value, the location point corresponding to the minimum value, use the location point as the new target location point, and generate the pre-driving route based on the path planning algorithm.
The multi-AGV system dynamic obstacle real-time obstacle avoidance system according to claim 10, wherein the path planning algorithm module is based on a global static obstacle matrix and uses a local particle swarm algorithm to calculate the current position point to the target point Pre-driving route.