Research On Complete Coverage Path Planning Of Multi-Robot In Considering Of Human-Robot Harmony

Complete coverage path planning technology which is widely used in resource detection,chart drawing,smart cleaning and other industries,is a key problem in the field of mobile robots.The shortcomings of the existing methods are: focusing on the improvement of the area coverage,which may lead to the increase of repetitive coverage;taking pedestrians in the environment as a general moving obstacle,while ignoring the peculiar movement rules of pedestrians.In order to solve these problems,this thesis will study the complete coverage path planning problem of multi-robot in considering of human-robot harmony.The main result of this thesis can be summarized as follows:1.A complete coverage path planning algorithm with escape mechanism is proposed using cellular automata(CA)and rapaidly random tree(RRT)algorithm,to deal with the dead zone problem of biologically inspired neural network(BINN)coverage algorithm.Firstly,in order to obtain the best point to get out of dead zone,the CA system is established,and the evolution rules of the cell system are designed according to the requirements of searching for the point of trouble removal.A search algorithm for the escape point is proposed,which is not affected by the distribution of obstacles.Secondly,in order to realize the path planning for getting out of dead zone,The global sampling problem of the RRT algorithm is improved by using the neuron activity value of the BINN algorithm,and the BINN-RRT algorithm is proposed to shorten the planning time.Then,using the escape point search algorithm and BINN-RRT algorithm to form an active escape mechanism,by which the robot is prevented from getting confused in the dead zone and the repeated path is reduced.Finally,the feasibility of the proposed algorithm is verified by simulation experiments.2.A trajectory prediction method that considers pedestrian motion characteristics is proposed by merging social force model(SFM)and Kalman filter(KF),to improve the accuracy of pedestrian trajectory prediction.Firstly,the KF prediction model is established and the limitations of the model are analyzed.Secondly,the motion model of pedestrians is modeled based on SFM,and some model parameters are identified by particle swarm optimization algorithm.Then,SFM-KF pedestrian trajectory prediction method is proposed by merging pedestrian motion model and KF prediction model.Finally,the prediction error of SFM-KF method is shown to be smaller than KF through the Gazebo software designing simulation experiment.3.A multi-robot complete coverage path planning method is proposed,for the coverage problem in multi-regional human-machine coexistence environment.Firstly,the task space is decomposed based on boustrophedon decomposition method,the influence of the initial motion direction of the robot into the sub-area is analyzed,and the multi-robot task assignment method based on the ant colony algorithm is proposed.Secondly,the prediction method realizes the complete coverage of the region based on the SFM-KF pedestrian trajectory considering of human-robot harmony.Then,the multi-robot collaboration covers large-scale human-machine coexistence environment combined with the task assignment algorithm and the coverage algorithm considering of human-robot harmony.Finally,the simulation environment designed by Gazebo software is used to verify the effectiveness of the method.