Research On Multi-object Search Method Of Swarm Robots In Complex Environment

Swarm robotic system is a recent hot topic,which is inspired by complex natural systems such as social insects(ants,bees,etc.)or groups of collaborative animal groups.The global behavior of swarm robot systems emerges from local rules implemented at the level of individual robots.The robot multi-objective search system is a typical task platform of the swarm robot system research,which has an important research value,and is beneficial to the search and rescue,group confrontation,cooperative handling,target protection and leadership guard,and also can be widely used in the anti-terrorism,military,security and vigilance,etc.However,the current theory of multi-target search for group robots is still incomplete,the search environment involved is relatively simple,the scalability of the search system is not strong,and the multi-object search algorithm is complex.Aiming at the above problems,this paper intends to study the multi-object search system of group robots in complex environment based on artificial intelligence method.The main research work and achievements are as follows.(1)A group robot cooperative control strategy(IEPSO-SVF)based on simplified virtual stress model is proposed.Task division strategy is a key problem to be solved first in the multi-objective search of group robots.Aiming at this problem,a novel dynamic closed-loop self-organizing task division model considering probabilistic obstacle(PO-TRT)is proposed on basis of the existing division of tasks based on the probability principle and the target response threshold in this paper,and the resource allocation level of the group robot is further optimized.As for the collision avoidance problem of robot in unknown complex environment,a simplified virtual force analysis model of individual robot is abstracted,and the model is introduced at the level of the subgroup alliance interior and roaming individual,and the input strategy of robot motion control is designed accordingly.In addition,aiming at the robot’s roaming strategy,a roaming strategy based on the principle of maximum search area and different search directions from the two neighborsis designed in this paper.The simulation results show that the proposed cooperative control strategy for group robots(IEPSO-SVF)based on the simplified virtual force model can reasonably allocate the robot resource level,effectively solve the problems of collision between the internal members of the system and between the system and the environment,and improve the search efficiency of the system.(2)A multi-objective cooperative search algorithm(MSRCPC)for swarm robots with consider the actual communication constraints is proposed.A multi-objective cooperative search algorithm(MSRCPC)for swarm robots with practical constraints is proposed.Aiming at the static multi-objective search problem for group robots in unknown complex environments,most of the existing research results are applied to specific target search scenarios,and lack a standard multi-target search algorithm framework,and do not simultaneously consider the algorithm performance,obstacle avoidance and group communication of group robots in actual search scenarios.In order to handle these problems,an ideal multi-object search framework for swarm robots is proposed based on finite state machine,then,in order to solves the obstacle avoidance problem of swarm robots in the process of multi-object search,the simplified virtual force model is utilized based on the this framework,finally,considering the communication interaction problem in the coordination behavior of group robots and the random line-of-sight problem of individual robots in the actual communication process,a distributed neighborhood interaction model(RS-TVCS)based on the time-varying feature group of restricted random line-of-sight is constructed.By embedding this sub-algorithm into the whole algorithm framework,a new multi-object collaborative search algorithm(MSRCPC)considering practical constraints is proposed.The proposed MSRCPC algorithm can greatly improve the search performance of swarm robot system,and makes the whole system with a better scalability and practicality.Simulation results show the effectiveness of the proposed method.(3)A multi-object search algorithm for swarm robot in an unknown 3D mountain environment is proposed.Most of the existing 3D environment obstacle avoidance algorithms are potential field method,which needs to consider the location information of all obstacles around the robot,and is easy to fall into local optimal and require complicated calculation,and cannot well meet the requirements of real-time obstacle avoidance characteristics of multi-target search in the swarm robot.This paper first focuses on solving the obstacle avoidance problem of swarm robot in the mountain environment.A new 3D curved obstacle tracking-algorithm is designed by discretizing the mountains within the detection range of robot obstacles.Then,the task assignment model and virtual force model in 2D space are extended to 3D,and a particle swarm search model with kinematic constraints is constructed,which considers the kinematic constraints and the limitations of communication ability of robot.Finally,a new multi-target search algorithm for the swarm robot in unknown3 D mountain environment is proposed by means of the designed 3D surface obstacle-tracking algorithm.Numerical simulation results demonstrate the effectiveness of the proposed algorithm.(4)In order to verify the validity of the above theoretical results,An integrated experimental system for cooperative control of microsmall group robots was constructed,including single robot,local autonomous wireless communication network of microsmall group robot control system,and outdoor combined positioning system based on GPS and RSSI.Finally,based on the above integrated experimental system,the static multi-objective search algorithm in 2D space proposed in the third and fourth chapters is verified.Experimental results demonstrate the validity of the results of this paper.