Research Of Key Issues For Multi-robot Formation Based On Wirless Sensor Network

Integrating formation control of mobile robot with sensor network in complex working environment is a cutting-edge technology in formation control of mobile robot,which is of great significance to location,path planning and formation control of formation method of mobile robot.Traditional robot formation technology usually USES inertial navigation or odometer to track the robot when it is separated from the external navigation;However,with the working mode of inertial navigation,the accumulated navigation errors and odometer errors will limit the use of inertial navigation method in robot formation control over time.Formation control method based on wireless sensor network(WSN)fusion technology can make full use of the self-organization of sensor network,intelligent and multi-source features,such as the formation of obstacle positions around the mobile robot's own work position,have a more accurate recognition and measurement,and then improve the work efficiency of formation system,recognition rate,robustness,universality,anti-inter-ference performance,etc.Based on the traditional formation control theory of mobile robot,this paper makes an in-depth study on the navigation,path planning and formation control theory of formation control system of mobile robot at the present stage,systematically analyzes the existing problems of formation control of mobile robot at the present stage,proposes and studies a new method of combining wireless sensor network with formation control of mobile robot.This dissertation carries out the following three aspects of research on the key technologies of formation control of mobile robot based on wireless sensor network:(1)Aiming at the problems of update speed of GPS signal slowly,loss of GPS signal easily,error coordinate and error rate values generated by INS integral calculation in the absence of GPS assisted positioning in the existing GPS/inertial guidance system positioning method.This paper proposed a strapdown aided localization algorithm for wireless sensor networks in a loosely coupled manner.The algorithm takes the coordinates of WSN position solution result as input and establishes a linear equation with the inertial navigation system in the same coordinate system.The WSN data and INS data are fused by standard kalman filter,so as to realize the accurate positioning of inertial navigation mode.Experimental results show that WSN/INS navigation mode has higher positioning accuracy in wireless sensor network environment than INS navigation mode without WSN.(2)According to traditional wheeled robot environment perception,calculation ability is limited,the traditional path planning methods such as artificial potential field method,simulated annealing method in large range,dynamic environment is difficult to ensure accuracy and reliability of the fleet system input information,and thus bring path repetition and redundancy.This paper proposed a network environment of the whole connection neural network based on improved particle swarm optimization of two-dimensional mobile robot path planning algorithm.Firstly,the algorithm generates the global coordinate system based on the wireless sensor network environment and establishes the kinematics model and local coordinate system of the mobile robots;Secondly,with the fast classification ability of feed-forward neural network,a collision avoidance controller for mobile robots based on feed-forward neural network is proposed,which transforms the obstacle constraint into a unique collision penalty function;and then,the global path planning processes are represented by the energy function,and the constrained optimization problem is transformed into an unconstrained optimization problem for solving.The energy function outputs the offset Angle and running speed in real time.Combined with wireless sensor network,the migration direction of mobile robot can be adjusted online,so that mobile robots can move towards the target independently without collision.Finally,particle swarm optimization is used to optimize the energy function and select the optimal path for the robot to escape from the trap states.Experimental results show that compared with the other related path planning methods,for a wide range of work environment,in meet the accuracy requirement,under the condition of the same path search time,the proposed path planning method in the narrow path search space,improve the generated path smoothness and increase the convergence speed are improved significantly.(3)Aiming at the problems of single control method,poor anti-interference ability and slow formation convergence speed of traditional robot formation leaderfollower control method,etc.This paper proposed a moving kinematic model of a modified pilot-directed team algorithm that establishes physical connection between a robot and an anchor node in a wireless sensor network environment.This model redesigns and defines formation controller and new tracker for wireless sensor working environment,and introduces table updating method,which updates the current coordinate,angle,speed and distance between leader and all follower in the table in real time.The idea of the algorithm is based on the idea of distributed computing,is designed to improve the navigator based on the coordination of the mission.Algorithms both leader and follower has participated in the corresponding work,follower share the leading part of the task,the traditional algorithm to reduce the burden of the pilot and the amount of calculation,so that the formation control more coordination and flexible.The proposed algorithm overcomes the disadvantages of the traditional leader-follower algorithm,which only relies on the distance sensor to obtain leader-distance and angle data to form formation,which are verified by experiments with the real object and Lypaunov method.Compared with the formation control method in non-wireless sensor network environment,the proposed algorithm has higher stability,smaller error,shorter formation recovery time,and is feasible.