WO2022094746A1

WO2022094746A1 - Multi-robot multi-task collaborative working method, and server

Info

Publication number: WO2022094746A1
Application number: PCT/CN2020/126092
Authority: WO
Inventors: 吴新开; 霍向; 马亚龙
Original assignee: 北京洛必德科技有限公司
Priority date: 2020-11-03
Filing date: 2020-11-03
Publication date: 2022-05-12

Abstract

A multi-robot multi-task collaborative working method, and a server. The multi-robot multi-task collaborative working method comprises: a robot central decision-making system establishing, by using a robot task matching optimization algorithm, the optimal matching relationship between available robots and tasks to be served; and the robot decision-making system planning, by using an intelligent robot path planning algorithm, a global path for task execution, and the robots executing respective tasks. The advantages of the method are that the method is simple and highly efficient, and the optimal task execution scheme can be quickly and efficiently planned for a plurality of robots in a complex scenario.

Description

A multi-robot multi-task cooperative working method and server

technical field

The invention relates to the technical field of multi-robot multi-task cooperative work, in particular to a multi-robot multi-task cooperative work method and a server.

Background technique

With the advancement of robot intelligence technology, intelligent service robots have been put into use in many fields. How to effectively carry out multi-robot and multi-task collaborative work planning to further improve the level of robot intelligence is an important research direction at present. However, the planning methods for multi-robots and multi-task cooperative work in the current traditional methods have problems such as complicated methods or the planning scheme makes the multi-robot cooperative work ineffective.

The method has certain innovation in the current research. Currently, patent application CN201910369515.1 provides a multi-task service robot and service system, and patent application CN202010190435.2 provides a robot multi-task control method, control device and terminal equipment. At present, there is a lack of systematic consideration in the planning of the multi-robot multi-task system work plan, the direct service matching between multi-robots and multi-tasks, and the planning of two aspects of robot-service multi-task path planning. It is necessary to provide a method for multi-robot multi-tasking. An intelligent planning approach that works together.

Therefore, in order to enrich the algorithm research in related fields and solve the problems of low work efficiency and poor service quality when multi-robots cooperate to perform multi-tasks in the market in a scientific way, an intelligent optimization method is designed.

SUMMARY OF THE INVENTION

In view of this, the present application provides a multi-robot multi-task cooperative working method and server.

This application is achieved through the following technical solutions:

A multi-robot multi-task cooperative working method, the method comprises the following steps:

Step 1: At the beginning of the working day, the central decision-making system of the robot determines multiple task service decision-making time points within 24 hours of the working day. According to the historical information of task requirements, more task services are set during the peak time period of task demand generation. Decision time points, set less task service decision time points in the off-peak time period of task demand generation;

Step 2: All robots upload the status information and remaining power information of the robot at the current time to the robot central decision-making system in real time;

Step 3: All robots use multiple positioning methods in real time to determine the current position in indoor or outdoor scenes and upload the position information to the robot central decision-making system;

Step 4: The central decision-making system of the robot obtains the task demand information in real time, and stores the obtained task demand information in the set of tasks to be served;

Step 5: The robot central decision-making system determines the corresponding task requirement weight and the task requirement fixed cost information from the set task database according to the task requirement type;

Step 6: When the robot central decision-making system determines in real time whether the current time has reached the task service decision time point;

Step 7: When the robot central decision-making system determines that the current time has reached the task service decision time point, the robot central decision-making system obtains the number of idle robots at the current time, the number of idle robots, the location information of idle robots, the remaining power of idle robots, and the working capacity of each idle robot. , the robot information about the motion speed of each idle robot, and save the information of the idle robot into the available robot set;

Step 8: The robot central decision-making system uses the robot task matching optimization algorithm to realize the matching relationship between the available robots and the tasks to be served according to the set of tasks to be served and the set of available robots at the task service decision time point;

Step 9: The robot central decision-making system sends the information of the tasks to be served corresponding to each available robot to the available robots. The decision-making system of the available robots uses the robot path planning intelligent algorithm to plan and obtain the global path for the available robots to perform tasks according to the tasks to be performed. ;

Step 10: The robot can change the state information from the idle state to the working state, and then execute the task according to the planned global path, and ensure the local path obstacle avoidance requirements in real time during the movement process;

Step 11: After the robot completes the task, the robot automatically checks whether the system is normal. If there is any fault, it will change from the working state to the maintenance state, and send the fault to the background. If there is no fault, it will return to the charging closest to the current location. stub and change the status information from working to idle.

Further, the state information in the above step 2 includes an idle state, a working state and a maintenance state.

Further, the positioning methods in the above-mentioned step 3 include but are not limited to: 1) use Beidou satellite navigation system, GPS navigation system, GLONASS navigation system or GALILEO navigation system to obtain position information; 2) use the environmental point cloud data detected by lidar The position information is obtained by analysis; 3) the position information is determined by visual analysis of the environmental image captured by the robot camera; 4) the position information is obtained by the trajectory measurement of the robot IMU and the odometer.

Further, the task requirement information in the above step 4 includes task requirement generation time, task requirement type, and task requirement location information.

Further, the robot task matching optimization algorithm in the above step 8 also includes:

The set of tasks to be served U={1,...,i,...,N _U }, the number of tasks to be served in the set of tasks to be served is N _U , the set of available robots R={1,...,j,..., N _R }, establish the following matching model, that is, by generating a robot task matching matrix Z, the elements z _j,i in the matrix Z indicate whether the jth available robot will serve the i-th task, z _j,i =1, determine The jth available robot serves the i-th task, z _{j, i} = 0, it is decided that the j-th available robot does not serve the i-th task, and the generated robot task matching matrix Z realizes the optimal matching model objective function that is The maximum objective function value is obtained, the matching relationship between the available robots and the tasks to be served is obtained,

Which matches the model objective function:

which matches the model constraints:

z _j,i ·(E _j,i +b _o )≤b _j

z _j,i ∈{0,1}

Among them, ri is the task requirement weight of the _i -th task, and this parameter is the reward obtained after completing the i-th task; z _j,i indicates whether the j-th available robot will serve the i-th task; λ _i is the i-th task. The loss cost of the i task, this parameter is the penalty obtained after failing to complete the i-th task; M is the number of available robots used; c ₁ is the fixed loss cost of the robot to perform the task; c ₂ is the motion of the robot to perform the task cost; ω _j is the path that the robot needs to move to perform the task; d _i is the path that the i-th task needs to move; E _j,i is the electricity consumed by the j-th available robot to serve the i-th task; b _o is power threshold; b _j is the remaining power of the jth available robot; g _i is the work capacity required by the i-th task, where the work capacity requirements include but are not limited to weight or volume; A _j is the work of the jth available robot ability.

Further, the robot path planning intelligent algorithm in the above step 9 also includes:

For the task to be performed by the jth available robot as determined in step 8

Determine the current position l _j of the jth available robot;

The system adds an intelligent virtual position point in the process of planning the robot path

Smart virtual location point

The movement cost to the current position l _j is a very small number, and the system can be initially set to a positive number approximately 0, and the current position l _j to the intelligent virtual position point

The movement cost of is a very large number, which is initially set in the system and can be set as the maximum value of the movement cost between the tasks in the current position l _j and U _j , the intelligent virtual position point

The movement cost between any task in U _j can be set to be the same as the movement cost between the current position l _j and any task in U _j ;

A task-related motion cost matrix D is established. The elements d _{h and k} in the matrix can be l _i .

Or any element in U _j , which represents the direct motion cost of h and k, and establishes the following path planning model, that is, by generating the robot path selection matrix Y, the elements y _{h, k} in the matrix Y indicate whether the jth available robot is determined. After passing position h or after the service completes task h, go to position k or go to service task k. y _h,k = 1, it is determined that the jth available robot will go to position k or go to service task k after the jth available robot passes through the position h or after the service completes the task h, y _h,k = 0, it is determined that the jth available robot passes through the position h. Or the service will not go to location k or go to service task k after completing task h. The generated robot path selection matrix Y realizes the optimal path planning model objective function, that is, the minimum objective function value, and then obtains the planned path of the robot,

The path planning model objective function is:

The path planning model constraints are:

y _h, k∈{0,1}

Among them, y _h,k indicates whether the jth available robot will go to position k or serve task k after passing through position h or after completing task h; d _h,k indicates the direct movement cost of h and k, h and k can be l _i ,

or any element in U _j ; Φ is the set

Any subset of , |Φ| is the number of elements in the set Φ, the set

The number of elements is N _lj .

Further, the present invention also provides a multi-robot multi-task cooperative work server, the intelligent optimization system includes: a processor and a memory for storing instructions, and when the above-mentioned instructions are executed by the processor, the processor realizes the process as claimed in claims 1-6. Any one of the multi-robot multi-task cooperative work methods

Compared with the prior art, the present invention has the advantages of simple method and high efficiency, and can quickly plan an optimal task execution scheme for multiple robots in a complex scene.

Description of drawings

Fig. 1 is the schematic flow chart of the optimization method of the present invention;

Detailed ways

Exemplary embodiments will be described in detail herein, examples of which are illustrated in the accompanying drawings. Where the following description refers to the drawings, the same numerals in different drawings refer to the same or similar elements unless otherwise indicated. The implementations described in the illustrative examples below are not intended to represent all implementations consistent with this application. Rather, they are merely examples of apparatus and methods consistent with some aspects of the present application as recited in the appended claims.

The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to limit the application. As used in this application and the appended claims, the singular forms "a," "the," and "the" are intended to include the plural forms as well, unless the context clearly dictates otherwise. It will also be understood that the term "and/or" as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

The present invention will be further described in detail below in conjunction with the accompanying drawings and examples.

FIG. 1 is a schematic flowchart of the optimization method of the present invention. The intelligent optimization method for multi-robot multi-task cooperative work provided by the present invention comprises the following steps:

Step 2: All robots upload the status information and remaining power information of the robot at the current time to the robot central decision-making system in real time, where the status information includes: idle state, working state and maintenance state;

Step 3: All robots use a variety of positioning methods to determine the current position in real time in indoor or outdoor scenes and upload the position information to the robot central decision-making system. The positioning methods include but are not limited to: 1) Using Beidou satellite navigation system, GPS navigation system, GLONASS navigation system or GALILEO navigation system to obtain position information; 2) use the environmental point cloud data detected by lidar to obtain position information; 3) use the environmental image captured by the robot camera to determine the position information through visual analysis; 4) use the robot IMU and odometer use trajectory measurement to obtain location information;

Step 4: The central decision-making system of the robot acquires task demand information in real time, and stores the acquired task demand information into the task set to be served, wherein the task demand information includes: task demand generation time, task demand type, task demand location and other information;

Step 5: The robot central decision-making system determines the corresponding task demand weight, task demand fixed cost and other information from the set task database according to the task demand type;

Step 7: When the robot central decision-making system determines that the current time has reached the task service decision time point, the robot central decision-making system obtains the number of idle robots at the current time, the number of idle robots, the location information of idle robots, the remaining power of idle robots, and the working capacity of each idle robot. , the robot information such as the motion speed of each idle robot, and save the information of the idle robot into the available robot set;

The robot task matching optimization algorithm includes: the set of tasks to be served U={1,...,i,...,N _U }, the number of tasks to be served in the set of tasks to be served is N _U , the set of available robots R ={1,...,j,...,N _R }, establish the following matching model, that is, by generating a robot task matching matrix Z, the elements z _j,i in the matrix Z indicate whether the j-th available robot will serve the i-th item Task, z _j,i = 1, decides the jth available robot to serve the i-th task, z _j,i = 0, decides that the j-th available robot does not serve the i-th task, the generated robot task matching matrix Z When the optimal matching model objective function is achieved, that is, the maximum objective function value, the matching relationship between the available robots and the tasks to be served is obtained. Which matches the model objective function:

which matches the model constraints:

z _j,i ·(E _j,i +b _o )≤b _j

z _j,i ∈{0,1}

Among them, ri is the task requirement weight of the _i -th task, and this parameter is the reward obtained after completing the i-th task; z _j,i indicates whether the j-th available robot will serve the i-th task; λ _i is the i-th task. The loss cost of the i task, this parameter is the penalty obtained after failing to complete the i-th task; M is the number of available robots used; c ₁ is the fixed loss cost of the robot to perform the task; c ₂ is the motion of the robot to perform the task cost; ω _j is the path that the robot needs to move to perform the task; d _i is the path that the i-th task needs to move; E _j,i is the electricity consumed by the j-th available robot to serve the i-th task; b _o is power threshold; b _j is the remaining power of the jth available robot; g _i is the working capacity required by the i-th task, which includes but is not limited to weight or volume; A _j is the jth available robot Ability to work;

Described robot path planning intelligent algorithm comprises: for the task to be performed determined in step 8 for the jth available robot

Determine the current position l _j of the jth available robot;

Smart virtual location point

The path planning model objective function is:

The path planning model constraints are:

y _h, k∈{0,1}

or any element in U _j ; Φ is the set

Any subset of , |Φ| is the number of elements in the set Φ, the set

The number of elements is N _lj ;

Those skilled in the art can understand that all or part of the steps in the above method can be completed by instructing relevant hardware through a program, and the program can be stored in a computer-readable storage medium, such as a read-only memory, a magnetic disk or an optical disk. Optionally, all or part of the steps in the above embodiments may also be implemented by using one or more integrated circuits. Correspondingly, each module/unit in the above embodiments may be implemented in the form of hardware, or may be implemented in the form of software function modules. form realization. The present invention is not limited to any particular form of combination of hardware and software.

It should be noted that the present invention can also have other various embodiments. Without departing from the spirit and essence of the present invention, those skilled in the art can make various corresponding changes and deformations according to the present invention, but these Corresponding changes and deformations should belong to the protection scope of the appended claims of the present invention.

Claims

A multi-robot multi-task cooperative working method, the method comprises the following steps:

Step 1: At the beginning of the working day, the central decision-making system of the robot determines multiple task service decision-making time points within 24 hours of the working day. According to the historical information of task requirements, more task services are set during the peak time period of task demand generation. Decision time points, set less task service decision time points in the off-peak time period of task demand generation;

Step 2: All robots upload the status information and remaining power information of the robot at the current time to the robot central decision-making system in real time;

Step 3: All robots use multiple positioning methods in real time to determine the current position in indoor or outdoor scenes and upload the position information to the robot central decision-making system;

Step 4: The central decision-making system of the robot obtains the task demand information in real time, and stores the obtained task demand information in the set of tasks to be served;

Step 5: The robot central decision-making system determines the corresponding task requirement weight and the task requirement fixed cost information from the set task database according to the task requirement type;

Step 6: When the robot central decision-making system determines in real time whether the current time has reached the task service decision time point;

Step 7: When the robot central decision-making system determines that the current time has reached the task service decision time point, the robot central decision-making system obtains the number of idle robots at the current time, the number of idle robots, the location information of idle robots, the remaining power of idle robots, and the working capacity of each idle robot. , the robot information about the motion speed of each idle robot, and save the information of the idle robot into the available robot set;

Step 8: The robot central decision-making system uses the robot task matching optimization algorithm to realize the matching relationship between the available robots and the tasks to be served according to the set of tasks to be served and the set of available robots at the task service decision time point;

Step 9: The robot central decision-making system sends the information of the tasks to be served corresponding to each available robot to the available robots. The decision-making system of the available robots uses the robot path planning intelligent algorithm to plan and obtain the global path for the available robots to perform tasks according to the tasks to be performed. ;

Step 10: The robot can change the state information from the idle state to the working state, and then execute the task according to the planned global path, and ensure the local path obstacle avoidance requirements in real time during the movement process;

Step 11: After the robot completes the task, the robot automatically checks whether the system is normal. If there is any fault, it will change from the working state to the maintenance state, and send the fault to the background. If there is no fault, it will return to the charging closest to the current location. stub and change the status information from working to idle.
The multi-robot multi-task cooperative working method according to claim 1, wherein the state information in the step 2 includes an idle state, a working state and a maintenance state.
A multi-robot multi-task cooperative working method according to claim 1, wherein the positioning method in the step 3 includes but is not limited to: 1) using Beidou satellite navigation system, GPS navigation system, GLONASS navigation system Or the GALILEO navigation system obtains the position information; 2) The position information is obtained by analyzing the environmental point cloud data detected by the lidar; 3) The position information is determined by the visual analysis of the environmental image captured by the robot camera; 4) The position information is determined by the robot IMU and odometer. Trajectory calculation to obtain position information.
The multi-robot multi-task cooperative working method according to claim 1, wherein the task requirement information in step 4 includes task requirement generation time, task requirement type, and task requirement location information.
A multi-robot multi-task cooperative working method according to claim 1, wherein the robot task matching optimization algorithm in step 8 further comprises:

The set of tasks to be served U={1,...,i,...,N U }, the number of tasks to be served in the set of tasks to be served is N U , the set of available robots R={1,...,j,..., N R }, establish the following matching model, that is, by generating a robot task matching matrix Z, the elements z j,i in the matrix Z indicate whether the jth available robot will serve the i-th task, z j,i =1, determine The jth available robot serves the i-th task, z j, i = 0, it is decided that the j-th available robot does not serve the i-th task, and the generated robot task matching matrix Z realizes the optimal matching model objective function that is The maximum objective function value is obtained, the matching relationship between the available robots and the tasks to be served is obtained,

Which matches the model objective function:

which matches the model constraints:

z j,i ·(E j,i +b o )≤b j

z j,i ∈{0,1}

Among them, ri is the task requirement weight of the i -th task, and this parameter is the reward obtained after completing the i-th task; z j,i indicates whether the j-th available robot will serve the i-th task; λ i is the i-th task. The loss cost of the i task, this parameter is the penalty obtained after failing to complete the i-th task; M is the number of available robots used; c 1 is the fixed loss cost of the robot to perform the task; c 2 is the motion of the robot to perform the task cost; ω j is the path that the robot needs to move to perform the task; d i is the path that the i-th task needs to move; E j,i is the electricity consumed by the j-th available robot to serve the i-th task; b o is power threshold; b j is the remaining power of the jth available robot; g i is the work capacity required by the i-th task, where the work capacity requirements include but are not limited to weight or volume; A j is the work of the jth available robot ability.
The multi-robot multi-task cooperative working method according to claim 1, wherein the robot path planning intelligent algorithm in step 9 further comprises:

For the task to be performed by the jth available robot as determined in step 8
Determine the current position l j of the jth available robot;

The system adds an intelligent virtual position point in the process of planning the robot path
Smart virtual location point
The movement cost to the current position l j is a very small number, and the system can be initially set to a positive number approximately 0, and the current position l j to the intelligent virtual position point
The movement cost of is a very large number, which is initially set in the system and can be set as the maximum value of the movement cost between the tasks in the current position l j and U j , the intelligent virtual position point
The movement cost between any task in U j can be set to be the same as the movement cost between the current position l j and any task in U j ;

A task-related motion cost matrix D is established. The elements d h and k in the matrix can be l i .
Or any element in U j , which represents the direct motion cost of h and k, and establishes the following path planning model, that is, by generating the robot path selection matrix Y, the elements y h, k in the matrix Y indicate whether the jth available robot is determined. After passing position h or after the service completes task h, go to position k or go to service task k. y h,k = 1, it is determined that the jth available robot will go to position k or go to service task k after the jth available robot passes through the position h or after the service completes the task h, y h,k = 0, it is determined that the jth available robot passes through the position h. Or the service will not go to location k or go to service task k after completing task h. The generated robot path selection matrix Y realizes the optimal path planning model objective function, that is, the minimum objective function value, and then obtains the planned path of the robot,

The path planning model objective function is:

The path planning model constraints are:

y h, k∈{0,1}

Among them, y h,k indicates whether the jth available robot will go to position k or serve task k after passing through position h or after completing task h; d h,k indicates the direct movement cost of h and k, h and k can be l i ,
or any element in U j ; Φ is the set
Any subset of , |Φ| is the number of elements in the set Φ, the set
The number of elements is N lj .
A multi-robot multi-task cooperative work server, characterized in that the intelligent optimization system comprises: a processor and a memory for storing instructions, and when the above-mentioned instructions are executed by the processor, the processor can implement any one of claims 1-6. The multi-robot multi-task cooperative working method.