CN108873894A - A kind of target following cooperative control system and method based on more unmanned boats - Google Patents

Abstract

The invention relates to a target tracking cooperative control system and method based on multiple unmanned boats. This system consists of a shore-based global positioning host connected to a single unmanned boat control system through a wireless communication module. The operation steps of this method are: 1) Formation generation process: use the auction algorithm to find the multi-unmanned boat-to-multi-target allocation scheme with the largest profit for the unmanned boat group; Path planning; 3) Use particle swarm-based prediction model to predict target trajectory, replace abnormal communication data, and track formation trajectory. By using the method proposed by the present invention, the calculation amount of the multi-round auction process is reduced, and the real-time requirement of multi-unmanned ship task distribution is achieved; the path planning method based on the geometric method and the trajectory tracking method based on the neural network are used to meet the requirements. The real-time and accuracy requirements of single-boat trajectory tracking control; using the particle swarm algorithm to predict the trajectory to compensate, improves the tracking ability of unmanned boats under limited communication conditions, and makes formation tracking highly reliable and stable .

Description

A multi-unmanned vehicle-based target tracking cooperative control system and method

technical field

The invention relates to multi-unmanned boat cooperative control technology, in particular to a multi-unmanned boat-based target tracking cooperative control system and method, belonging to the field of single unmanned boat target tracking and the field of multi-unmanned boat multi-target distribution cooperative control.

Background technique

As a miniaturized, intelligent and unmanned surface carrier platform that can navigate by remote control or autonomously, the unmanned vehicle can use different modules and equipment to complete the mission goals for specific mission types. Due to the increasingly complex marine environment, increasingly diverse operating tasks, and limited individual operating capabilities, the collaborative team has an overall performance that is superior to the sum of the individual performances of individual unmanned vehicles. This potential and considerable synergistic effect has become a research topic. The main power of boat cooperative control. The unmanned vehicle team can perform more diverse tasks, the system survivability is significantly improved, and the adaptability to complex environments is also stronger. In the military field, the coordination of multiple unmanned vehicles has demonstrated extremely important application value. Related applications include cooperative mine clearance, cooperative escort, cooperative situational awareness, cooperative tracking and cooperative encirclement and other combat tasks. In the civilian field, the coordination of multiple unmanned vehicles can greatly extend the scope of marine operations.

To form the comprehensive capability of the UAV team, it is first required that the UAV must have a high degree of autonomy and strong collaboration in the team. Therefore, the cooperative decision-making and control of unmanned vehicles has become a hot spot in the field of applied research on unmanned vehicles.

Cooperative target tracking is one of the effective methods to realize cooperative control of multiple unmanned vehicles. Researchers have achieved many research results on the problem of cooperative tracking and control of multiple unmanned vehicles, but there is still a gap between theory and practice. There is often more than one target for multi-boat cooperative tracking. The unmanned boat team must first find a target allocation plan that matches each single boat, and approximately generate the tracking formation of the required formation. Angle relationship, adjust your own trajectory in real time, and keep the team tracking formation. Considering that the collaboration between multiple boats inevitably requires communication, how to improve the reliability and stability of the tracking ability of the tracking team under limited communication conditions is also a problem of great practical value. Therefore, the research on multi-unmanned vehicle target tracking cooperative control system has important theoretical value and practical significance.

Contents of the invention

The object of the present invention is to provide a stable and operable multi-unmanned ship-based target tracking cooperative control system and method for the deficiencies of the prior art.

The technical solution adopted by the present invention to solve its technical problems is:

A cooperative control system based on multiple unmanned boats, including a shore-based global positioning host, a wireless communication module and a single unmanned boat control system. The shore-based global positioning host is used as a host computer to process the data collected by the sensor systems mounted on each unmanned boat, and obtain the target and the pose information of each unmanned boat. The wireless communication module is mainly responsible for uploading the sensing information of each boat to the shore-based host, and at the same time transmitting the pose information to the bottom of each unmanned boat.

The single unmanned boat control system includes a controller module, a sensor module and a motion module. The controller module is the decision-making center for controlling the motion of the unmanned vehicle, including the main controller and the co-controller. The main controller is used to receive the pose information sent by the host computer and the processed sensor information sent by the co-controller, and The control algorithm is embedded in it to control the autonomous movement of the unmanned boat. The co-controller is used to collect and process the information of the sensor module, and finally send it to the main controller. The sensor module detects external environmental information through various sensors mounted on it. The motion module is responsible for executing commands from the master controller to complete control tasks.

A method for target tracking cooperative control based on multiple unmanned boats, using the above-mentioned cooperative control system based on multiple unmanned boats, the specific implementation steps are as follows:

Step 1: In the formation formation process, for the target matching problem between multiple unmanned vehicles and multiple targets, use the auction-based allocation algorithm to construct a multi-stage task allocation bidding subsystem, and find the allocation scheme that maximizes the benefits of the unmanned vehicle group :

1.1 Preprocess the unmanned vehicles and targets in the area: delineate the unmanned vehicles and target tasks that meet the threshold conditions as a bidding sub-area;

1.2 Multiple rounds of bidding for each bidding sub-area:

(1-1) Assign roles to the unmanned boats participating in the auction, one of them will be the auctioneer, and the rest will be the bidders;

(1-2) The bidders calculate the cost required to complete each target task and start bidding;

(1-3) Calculate the overall income at the end of each round of bidding;

(1-4) Compare the overall income of each round of bidding, and take the income and the largest group as the final task allocation plan.

Step 2: Carry out geometric path planning for the motion of the unmanned boat from any initial state to the target point, and at the same time use the PID speed regulation mechanism based on the neural network to improve the motion control performance of the unmanned boat, and build a single unmanned boat motion control subsystem:

2.1 Carry out geometric path planning for the movement of the unmanned vehicle to the target point:

(2.1-1) According to the trajectory of the target, give the desired movement path and direction angle of the target;

(2.1-2) Establish the precise pose relationship between the unmanned vehicle and the target point at each moment during the movement process, and calculate the distance and angle deviation between the unmanned vehicle and the target;

(2.1-3) Deduce the linear velocity and angular velocity of the unmanned boat;

(2.1-4) According to the kinematics model of the unmanned vessel, the angular velocity of the linear velocity of the unmanned vessel is converted into the throttle and rudder angle.

2.2 Using the network-based PID speed regulation mechanism to improve the motion control performance of the unmanned vehicle:

(2.2-1) Let k=1, calculate the parameters in the path planning algorithm according to the initial position of the UAV and the target, and determine the initial value of the neural network structure;

(2.2-2) Calculate the control output throttle and rudder angle according to the speed and angular velocity of the unmanned boat;

(2.2-3) Use the speed feedback amount, calculate the error, use the online self-learning ability of the neural network to adjust the controller parameters, and calculate the control output;

(2.2-4) Update the parameter values in the path planning algorithm at the next moment;

(2.2-5) Judging whether the unmanned boat has reached the target point, the movement is over, otherwise, proceed to (2.2-6);

(2.2-6) Let k=k+1, return to step (2.2-);

Step 3: Use the particle swarm-based prediction model to predict the target trajectory, replace the abnormal communication data, track the formation trajectory, and build a multi-unmanned vehicle tracking control subsystem:

(3-1) Use the global positioning host to initialize the target and the position coordinates of the UAV;

(3-2) Construct the relative position and angle relationship between the target and the unmanned vessel according to the position coordinate information of each unmanned vessel, obtain the error dynamics equation, and then deduce the motion control law of the unmanned vessel;

(3-3) Establish an unmanned ship trajectory prediction model based on the nonlinear least squares method, use the particle swarm algorithm to solve the problem of trajectory prediction model parameter selection, update the velocity and position of each particle, and use m historical data as training samples, Calculate the fitness function, update the individual extremum of the particle with the value of a group of particles with the optimal fitness function, update the global extremum with the optimal value of the individual extremum, and take the best one with the best fitness function value for multiple iterations Preliminary establishment of a prediction model for group particles;

(3-4) Start formation movement: the unmanned boat continuously receives the coordinate information sent by the host computer, follows the target movement trajectory, and continuously uses the data of the next moment as a training sample to update the trajectory prediction model, and predicts the next moment backward The moving point of the target trajectory;

(3-5) According to the defined abnormal range of communication data, the unmanned boat makes a judgment on the obtained target position information: if the data is within the normal range, the unmanned boat tracks the target’s trajectory movement; if there is an abnormality in the communication data, Then enable the estimated point as the trajectory point of the unmanned boat at the next moment;

(3-6) Constantly judge whether the formation meets the formation requirements during the movement: if not, the UAV will make corresponding adjustments according to the formation requirements; if it meets the requirements, it will continue to move towards the target point until it reaches Target.

Compared with the prior art, the present invention has the following obvious outstanding substantive features and significant technological progress:

(1) The multi-unmanned vehicle-based target tracking cooperative control system has a reasonable design, strong system hierarchy, and is easy to implement in terms of hardware and software;

(2) In the task allocation stage, this method first preprocesses the unmanned boats and tasks into groups, thereby reducing the amount of calculation in the auction process and meeting the real-time requirements, and then adopts multiple rounds of auctions to improve the efficiency of target allocation ;

(3) The path planning based on the geometric method is simple and convenient, which is suitable for the needs of practical applications; the trajectory tracking method based on the neural network can meet the accuracy requirements;

(4) Use the optimization search ability of the particle swarm algorithm to predict the trajectory and improve the tracking of unmanned vehicles under limited communication conditions.

Tracking capability, thus making formation tracking highly reliable and stable.

Description of drawings

Fig. 1 is a block diagram of the overall structure of the collaborative control system based on multi-unmanned boats in the present invention.

Fig. 2 is a schematic flow chart of the overall method in the present invention.

Fig. 3 is a schematic diagram of a bidding framework for multi-stage task allocation in the present invention.

Fig. 4 is a flow chart of the auction party's unmanned boat algorithm in the method of the present invention.

Fig. 5 is a flow chart of the bidding party's unmanned boat algorithm in the method of the present invention.

Fig. 6 is a schematic diagram of the framework of the single unmanned boat motion control system in the present invention.

Fig. 7 is a schematic diagram of the framework of the multi-unmanned boat tracking control system in the present invention.

Detailed ways

Below in conjunction with accompanying drawing, preferred embodiment of the present invention is described in detail:

Embodiment one:

As shown in Figure 1: This target tracking cooperative control system based on multiple unmanned boats is composed of a shore-based global positioning host (1) connected to a single unmanned boat control system (3) through a wireless communication module (2). The base global positioning host (1) is used as the upper computer to process the data collected by the sensor system mounted on each unmanned boat, and obtain the pose information of the target and each unmanned boat; the wireless communication module (2) is mainly responsible for The perception information of each boat is uploaded to the shore-based global positioning host (1), and at the same time, the pose information can be transmitted to the bottom of each unmanned boat;

The single unmanned boat control system (3) includes a controller module (3-1), a sensor module (3-2) and a motion module (3-3), and the controller module (3-1) controls the unmanned boat The decision-making center of the movement, including the main controller (3-1-1) and co-controller (3-1-2), the main controller (3-1-1) is used to receive the pose information sent by the host computer, and The processed sensor information sent by the co-controller (3-1-2), and the control algorithm is implanted in it to control the autonomous movement of the unmanned boat; the co-controller (3-1-2) is used to collect the sensor module ( 3-2) and process the information, and finally send it to the main controller (3-1-1); the sensor module (3-2) detects the external environment information through various sensors equipped; the motion module (3-3) is responsible for Execute commands from the master controller (3-1-1) to complete control tasks.

Embodiment two:

As shown in Figure 2, this target tracking cooperative control method based on multi-unmanned vehicles uses the above-mentioned system to operate, and its operation steps are as follows:

Step 1: As shown in Figure 3, in the formation formation process, for the target matching problem between multiple unmanned vehicles and multiple targets, use the auction-based allocation algorithm to construct a multi-stage task allocation bidding subsystem, and find Largest allocation scheme:

1.1 Preprocess the unmanned vehicles and targets in the area: delineate the unmanned vehicles and target tasks that meet the threshold conditions as a bidding sub-area;

1.2 Multiple rounds of bidding for each bidding sub-area:

(1.2-1) Initialize n=1. Let M be the size of the unmanned boat set, and N be the set size of the target task;

(1.2-2) let i=n;

(1.2-3) For the target task c, calculate the cost of all unmanned boats to complete the target task, and record the unmanned boat with the smallest cost, and match it with c;

(1.2-4) i++, if i≤N+n-1, go to execute (1.2-3), otherwise execute down (1.2-5);

(1.2-5) Calculate the overall revenue of each round of bidding system, if n≤N, then n++, and then go to execute (1.2-2), otherwise, execute (1.2-6);

(1.2-6) Take the overall benefit and the largest group as the final allocation plan, and the task allocation ends.

1.3 As shown in Figure 4, the algorithm flow of the auctioneer is as follows:

(1.3-1) For the unmanned boats and target numbers in the bidding area, bid sequentially starting from the target task numbered 1.

(1.3-2) During each round of bidding, the auctioned unmanned boat first sends a task auction message. After receiving the auction message, each bidding unmanned boat calculates the cost of completing the target task and sends it to the auctioned unmanned boat. When all After bidding for the unmanned boat auction, compare and calculate the price of each bidding unmanned boat, find out the unmanned boat with the least cost and the highest income to complete the target task and notify the winning bidder, and then follow this step to complete the rest of the goals in turn The auction of the task, when all the target assignments are completed, calculate the sum of the income of this round of assignment results.

(1.3-3) Change the order of the auction, starting from the target task labeled 2, and the process is as described above, until the completion of the M-th target auction, compare and calculate a group of distribution schemes with the smallest sum of each round of auction revenue , notify all bidders for the unmanned boat, and the auction is over.

1.4 As shown in Figure 5, the algorithm flow of the bidder is as follows:

(1.4-1) After receiving the task auction message, the bidding unmanned boat calculates the income and cost of completing the task, and sends its own bid to the auction unmanned boat;

(1.4-2) Auction of unmanned boats collects the prices of bidding unmanned boats for comparison and notifies the winning bidder of unmanned boats;

(1.4-3) After multiple rounds of bidding, the distribution result of the group with the largest sum of revenue from the auction of the unmanned boat will notify the bidding of the unmanned boat, and the bidding of the unmanned boat will receive the final target distribution result, and the bidding process will end.

Step 2: As shown in Figure 6, carry out geometric path planning for the motion of the unmanned boat from any initial state to the target point, and at the same time use the PID speed regulation mechanism based on the neural network to improve the motion control performance of the unmanned boat, and build a single unmanned boat motion control Subsystem:

2.1 Carry out geometric path planning for the movement of the unmanned vehicle to the target point:

(2.1-1) According to the trajectory of the target, give the desired movement path and direction angle of the target;

(2.1-2) Establish the precise pose relationship between the unmanned vehicle and the target point at each moment during the movement process, and calculate the distance and angle deviation between the unmanned vehicle and the target;

(2.1-3) Deduce the linear velocity and angular velocity of the unmanned boat;

(2.1-4) According to the kinematics model of the unmanned vessel, the angular velocity of the linear velocity of the unmanned vessel is converted into the throttle and rudder angle.

2.2 Using the network-based PID speed regulation mechanism to improve the motion control performance of the unmanned vehicle:

(2.2-1) Let k=1, calculate the parameters in the path planning algorithm according to the initial position of the UAV and the target, and determine the initial value of the neural network structure;

(2.2-2) Calculate the control output throttle and rudder angle according to the speed and angular velocity of the unmanned boat;

(2.2-3) Use the speed feedback amount, calculate the error, use the online self-learning ability of the neural network to adjust the controller parameters, and calculate the control output;

(2.2-4) Update the parameter values in the path planning algorithm at the next moment;

(2.2-5) Judging whether the unmanned boat has reached the target point, the movement is over, otherwise, proceed to (2.2-6);

(2.2-6) Let k=k+1, return to step (2.2-2);

Step 3: As shown in Figure 7, use the particle swarm-based prediction model to predict the target trajectory, replace the abnormal communication data, track the formation trajectory, and build a multi-unmanned vehicle tracking control subsystem:

(3-1) Use the global positioning host to initialize the target and the position coordinates of the UAV;

(3-2) Construct the relative position and angle relationship between the target and the unmanned vessel according to the position coordinate information of each unmanned vessel, obtain the error dynamics equation, and then deduce the motion control law of the unmanned vessel;

(3-3) Establish an unmanned ship trajectory prediction model based on the nonlinear least squares method, use the particle swarm algorithm to solve the problem of trajectory prediction model parameter selection, update the velocity and position of each particle, and use m historical data as training samples, Calculate the fitness function, update the individual extremum of the particle with the value of a group of particles with the optimal fitness function, update the global extremum with the optimal value of the individual extremum, and take the best one with the best fitness function value for multiple iterations Preliminary establishment of a prediction model for group particles;

(3-4) Start formation movement: the unmanned boat continuously receives the coordinate information sent by the host computer, follows the target movement trajectory, and continuously uses the data of the next moment as a training sample to update the trajectory prediction model, and predicts the next moment backward The moving point of the target trajectory;

(3-5) According to the defined abnormal range of communication data, the unmanned boat makes a judgment on the obtained target position information: if the data is within the normal range, the unmanned boat tracks the target’s trajectory movement; if there is an abnormality in the communication data, Then enable the estimated point as the trajectory point of the unmanned boat at the next moment;

(3-6) Constantly judge whether the formation meets the formation requirements during the movement: if not, the UAV will make corresponding adjustments according to the formation requirements; if it meets the requirements, it will continue to move towards the target point until it reaches Target.

Claims (5)

1. a kind of target following cooperative control system based on more unmanned boats, which is characterized in that by bank base Global localization host（1） Module by wireless communication（2）Connect single unmanned boat control system（3）It constitutes, the bank base Global localization host（1）As upper Position machine is used to handle the data for the sensing system acquisition being mounted on each unmanned boat, obtains the pose letter of target and each unmanned boat Breath；The wireless communication module（2）It is mainly responsible for and the perception information of each ship is uploaded to the bank base Global localization host（1）, Posture information can be transferred to each unmanned boat bottom again simultaneously；

The list unmanned boat control system（3）Including controller module（3-1）, sensor module（3-2）And motion module（3- 3）, controller module（3-1）It is to control the decision center of unmanned boat movement, including master controller（3-1-1）With association's controller（3- 1-2）, master controller（3-1-1）For receiving the posture information of host computer transmission, and association's controller（3-1-2）It sends over Processed sensor information, and be implanted into control algolithm wherein, control unmanned boat autokinetic movement；Assist controller（3-1- 2）For acquiring sensor module（3-2）Information and process, be last transmitted to master controller（3-1-1）；Sensor module （3-2）External environmental information is detected by the various sensors of carrying；Motion module（3-3）It is responsible for executing and comes from master controller （3-1-1）Order complete control task.

2. a kind of target following cooperative control method based on more unmanned boats, using described in claim 1 based on more unmanned boats Target following cooperative control system operated, it is characterised in that include the following steps：

（1）In flight pattern generating process, for the object matching problem of more unmanned boats and multiple target, using based on auction Allocation algorithm constructs phased mission system and distributes auction subsystem, finds the allocation plan for making unmanned boat group Income Maximum；

（2）Unmanned boat is moved from arbitrary initial state to target point and carries out geometric path planning, while using based on nerve net The PID speed regulation mechanism of network promotes the motion control performance of unmanned boat, constructs single unmanned boat motion control subsystem；

（3）Target trajectory is predicted using the prediction model based on population, is replaced communication abnormality data, is carried out formation track Tracking, constructs more unmanned boat tracing control subsystems.

3. as claimed in claim 2 based on the target following cooperative control method of more unmanned boats, it is characterised in that the step (1) include：

1.1 unmanned boats being directed in region and target are pre-processed：The unmanned boat and goal task circle of threshold condition will be met It is set to an auction subregion；

1.2 carry out more wheel auctions to each auction subregion respectively：

（1.2-1）N=1 is initialized, if M is the size of unmanned boat set, N is the set sizes of goal task；

（1.2-2）Enable i=n；

（1.2-3）It for goal task c, calculates all unmanned boats and completes the cost of the goal task, and record least cost The unmanned boat of cost matches it with c；

（1.2-4）I++ turns to execute if i≤N+n-1（1.2-3）, otherwise execute downwards（1.2-5）；

（1.2-5）Every wheel auction bidding system integral benefit is calculated then to turn to execute if n≤N, n++（1.2-2）, otherwise, hold Row（1.2-6）；

（1.2-6）Take integral benefit and it is maximum one group as final allocation plan, task distribution terminates.

4. as claimed in claim 2 based on the target following cooperative control method of more unmanned boats, it is characterised in that the step (2) it specifically includes：

2.1 pairs of unmanned boats carry out geometric path planning to the movement of target point：

（2.1-1）According to the motion profile of target, the desired motion path of target and orientation angle are provided；

（2.1-2）Establish unmanned boat and target point accurate position orientation relation of each moment during the motion, calculate unmanned boat with The angular deviation of the distance between target；

（2.1-3）Derive linear velocity, the angular speed of unmanned boat；

（2.1-4）According to unmanned boat kinematics model, throttle and rudder angle are converted by unmanned boat linear velocity angular speed；

2.2 promote the motion control performance of unmanned boat using network-based PID speed regulation mechanism：

（2.2-1）K=1 is enabled, according to the initial position of unmanned boat and target, each parameter in path planning algorithm is calculated, determines mind Structure initial value through network；

（2.2-2）According to unmanned boat speed angular speed, control output throttle and rudder angle is calculated；

（2.2-3）Using velocity feedback quantity, error is calculated, is joined using the automatic measure on line capacity adjusting controller of neural network Control output is calculated in number；

（2.2-4）Update each parameter value in the path planning algorithm of subsequent time；

（2.2-5）Judge whether unmanned boat reaches target point, movement terminates, and otherwise, executes downwards（2.2-6）；

（2.2-6）Enable k=k+1, return continues to execute the（2.2-2）Step.

5. as claimed in claim 2 based on the target following cooperative control method of more unmanned boats, it is characterised in that the step (3) it specifically includes：

（3-1）Utilize Global localization host initialized target and the position coordinates of unmanned boat；

（3-2）The relative position angular relationship between target and unmanned boat is constructed according to the location coordinate information of each unmanned boat, is obtained To error dynamics equation, and then derive the motion control rule of unmanned boat；

（3-3）The unmanned boat trajectory predictions model based on nonlinear least square method is established, solves track using particle swarm algorithm Prediction model parameters On The Choice updates speed and the position of each particle, uses m historical data as training sample, meter Fitness function is calculated, with the individual extreme value of the value more new particle of one group of optimal particle of fitness function, using in individual extreme value Optimal value update global extremum, one group of particle that successive ignition takes fitness function value best tentatively establishes prediction model；

（3-4）Start movement of forming into columns：Unmanned boat constantly receives the coordinate information of host computer transmission, and target trajectory is followed to transport It is dynamic, while trajectory predictions model constantly is updated using the data of subsequent time as training sample, and predict subsequent time backward Target trajectory motor point；

（3-5）According to the communication data abnormal ranges of definition, unmanned boat judges the target position information of acquisition：If data Within normal range (NR), then unmanned boat tracking target trajectory movement, if there is communication data exception, enabling estimates a work For the motion profile point of unmanned boat subsequent time；

（3-6）Constantly judge whether flight pattern meets formation requirement during the motion：If not satisfied, then unmanned boat is according to volume Team requires to do corresponding adjustment, if meeting the requirements, continues to move to target point, until moving to target point.