| Multi-robot search has a wide range of applications when robots perform complex tasks such as reconnaissance,search and rescue,and detection in dangerous or harsh environments.At present,for complex and changeable unknown environments or unstructured environments such as battlefields and disaster site,due to the lack of prior knowledge or the difficulty of identifying environmental elements and the lack of global map information,multi-robot cooperation is required to conduct area coverage search for unknown environments.The multi-robot area search algorithms based on the biologically inspired neural network model have no requirements for the obstacle distribution or target information in the unknown areas.The environmental information obtained during the search process is reflected in the neural network model,and the robot automatically plans the search path according to the gradient decreasing principle.However,there are some problems in the research of intelligent search algorithm based on biological inspiration,such as insufficient consideration of environmental information,easy to fall into local optimization in the later stage of search,and weak coordination of multiple robots in the covering search.These problems bring difficulties and challenges to the application of multiple robots in the coverage search.Therefore,this thesis conducts in-depth research on the problems existing in multi-robot cooperate area coverage search in unknown environments,and innovates from the aspects of robot search efficiency in the later stage,multi-robot cooperativity,and adaptability to complex environments.Four multi-robot search algorithms are designed:1)Aiming at the problem of low late-stage search efficiency in the bio-inspired based multi-robot coverage search algorithms in unknown environments,on the basis of the biologically inspired neural network search algorithm,a multi-robot area search algorithm based on rolling optimization decision is designed in this thesis(rolling optimization decision based cooperative search algorithm).For unknown environments,we first build an environmental information representation model that combines the grid map with the Glasius bio-inspired neural network(GBNN).The environmental information detected by the robot during its movement will be presented in the form of the neuron activity value landscape of GBNN.In order to improve the search efficiency in late-stage,the distributed model predictive control method is introduced as the robot search path decision mechanism,and the differential evolution algorithm is used to optimize the solution to obtain the next search path.Finally,simulation experiments under different task area areas are designed,and the simulation results show that the proposed algorithm can effectively improve the late-stage search efficiency of the robot compared with the biologically inspired neural network search algorithm.2)Aiming at the problems that the area near obstacles cannot be fully searched and the cooperativity among multiple robots is not strong in the rolling optimization decision based cooperative search algorithm,we further propose a multi-robot area search algorithm based on iteratively cooperative decision-making(iteratively cooperative decision based search algorithm).Based on the cooperative search algorithm of rolling optimization decision,the search efficiency function is improved to enhance the search of the area near the obstacle,and an iteratively cooperative decision mechanism is designed to make each robot fully consider the decision information of other robots when making the search path decision.Finally,the search simulation experiments in different environments are designed,and the simulation results show that the proposed algorithm can effectively improve multi-robot cooperative performance compared with the rolling optimization decision based cooperative search algorithm.3)When the communication distance between robots is limited and the number of multi-robot systems is large,how to ensure that the multi-robot system can perform area search tasks stably and effectively in unknown environments is an urgent problem to be solved.To this end,we propose a multi-robot area search algorithm based on grouping cooperation mechanism.Based on the iteratively cooperative decision search algorithm,the algorithm further considers local communication and designs a grouping cooperation mechanism,which divides all robots into multiple search groups and uses the iteratively cooperative decision mechanism in each group to plan the next search path.Finally,simulation experiments are designed under different numbers of robots,and the simulation results show that the proposed algorithm can effectively perform area coverage search for multiple robots with different scales compared with existing algorithms.4)In order to solve the problem that the search efficiency of the existing methods may decrease in the complex obstacle environment,we propose a multi-robot area search algorithm based on the improved neural dynamics model.Firstly,the algorithm improves the characteristics of the GBNN model to eliminate the influence of too many obstacles on the robot in the task area.Secondly,an auxiliary search mechanism based on the improved GBNN model is proposed to help the robot find the undetected area quickly when it falls into the local optimum.Finally,the search simulation experiments under different obstacle environments are designed,and the simulation results show that the proposed algorithm can effectively improve the adaptability of multi-robot to complex obstacle environments compared with existing algorithms. |
