Research On Cooperative Method Of Swarm Robots Inspired By Bacterial Intelligence

The swarm robotic system is a typical representative of artificial intelligence systems,inspired by complex natural social beings,such as bird flocks,fish schools and ant colony,etc.,showing collective behavior through local interactions between individuals or between individuals and the environment.With the wide application of swarm robot systems in various fields,future robotic systems,such as ground mobile robots,water(underwater)robots or unmanned aerial vehicles(UAVs),are bound to develop in the direction of intelligence and swarming.Compared with single robot system,swarm robotic system has parallel processing ability and high robustness.In view of the characteristics of simple individual structure,local communication,small storage unit memory and limited computing power in the swarm robot system,how to design effective and efficient cooperative control methods under these constraints to achieve complex collective behavior to complete more tasks is a key problem in the study of swarm robots.This paper focuses on the robot motion control and swarm robot cooperation inspired by bacterial intelligence,and carries out a series of studies on multi-source search,the deployment and aggregation,and pattern generation of swarm robots.The specific contents are as follows:(1)At present,single target source search has been widely studied.However,in reality,the situation of multi-target source search is ubiquitous.When the number of sources is unknown,how to effectively and efficiently search for multi-source is the difficulty of the robot multi-source search problem.In the biological world,source search is the basic behavior of some organisms.The mechanisms behind these biological behaviors have facilitated the development of robotic source search techniques.Inspired by the foraging behavior of Physarum polycephali in nutrient-poor petri dishes,a multi-source search method based on gradient foraging of Physarum polycephalum is proposed,in which the robot can imitate the gradient climbing behavior of Physarum polycephali pseudopodia to traverse all sources in the environment.At the same time,local trap escape strategy based on noise introduction and obstacle avoidance strategy based on comprehensive gradient are proposed,so that it can carry out local trap escape and real-time obstacle avoidance.Simulation results show the effectiveness of the proposed algorithm.(2)There is intelligence behind bacteria,which is mainly reflected in the ability to adapt to the environment.In this paper,a bacterial chemotaxis-inspired cooperative strategy for swarm robots is proposed to solve the deployment and aggregation problems of swarm robots from the perspective of low-level motion control of individuals.For the deployment task,in the proposed method,the robot performs "Swimming" and "Tumbling" similar to bacteria to achieve swarm robot deployment.Meanwhile,the algorithm is extended to solve swarm robot aggregation.In addition,the proposed method introduces a search factor to prevent the generation of multiple subgroups.During the deployment and aggregation process of swarm robots,the system is completely distributed and there is no network structure connecting some robots.The simulation results can verify that the algorithm is superior to the comparison algorithms in terms of iterative consumption and success rate.In particular,the performance advantage of the proposed algorithm is more obvious in aggregation task with a large number of robots.(3)Furthermore,the robot is considered as a particle,and the robot moves continuously.A cooperative strategy of swarm robots with adaptive velocity adjustment is proposed.First,a fitness function based on the distance between robots is constructed to evaluate the robot’s current score.Second,the introduction of "evolution speed" and "robot aggregation degree"strengthens the influence of inertia weight,and provides an effective mechanism for adaptive velocity regulation.Robots perform behaviors similar to bacterial chemotaxis based on the the fitness function values changes at adjacent moments,enabling the deployment or aggregation of swarm robots.Finally,simulations verify the effectiveness of the proposed algorithm.In particular,the proposed method can effectively balance the success rate and the number of iterations compared with its variants.(4)Swarm robot pattern generation can be viewed as a special form of deployment,that is,deploying robot swarms uniformly within a characteristic pattern or on pattern target points.Aiming at the self-organized pattern generation of swarm robots,a hierarchical swarm robot pattern generation method inspired by E.coli controller is proposed,which decomposes the pattern generation into two stages:the aggregation stage and the pattern generation stage.In the pattern generation stage,the robot can flexibly switch between running and rotating,and adjust its position to self-organize to generate predefined pattern.Among them,a decision factor is introduced to effectively evaluate the direction and degree of the neighbor’s forces on it,and determine the rotating angle.Simulation results verify the effectiveness of the proposed cooperative control strategy on pattern generation and pattern retention.Aiming at the existence of target points in predefined patterns,this paper proposes a swarm robot pattern generation method inspired by bacterial gene expression control.In this method,each robot is regarded as a bacterium.By studying the mechanism of gene expression control on bacterial movement,a new robot movement model is constructed to help the robot autonomously search for target points on the pattern and quickly generate predefined patterns.Simulation experiments verify the effectiveness of the method.