Research On Pigeon-Inspired Optimization With Cooperation-Competition Mechanism And Its Cooperative Application For Multiple Robots

In recent years,the research achievements,scientific topics,and academic conferences on swarm intelligence are booming.Currently,the hotspots of swarm intelligence are the optimization algorithm based on swarm mechanisms and multiple robots cooperative applications based on swarm behaviors.The goal is to complete a task through distributed autonomous decision-making and information interaction for multiple agents.According to the cooperation-competition relationships shown in multiple robots cooperative applications,a coevolution pigeon-inspired optimization with cooperation-competition mechanism is proposed and applied to cooperative region search and cooperative path planning tasks.The main contents and innovations of this paper can be summarized as follows:First,a new path planning planner for a single robot,the connection point method,is proposed to find the shortest path that the existing path planners are difficult to find the optimal path in trap-like environment and the computation time increases sharply with zooming in the grid map.We prove that the path with the sub-convex corner point as the search node is the optimal path via plane geometry.As the grid map zooms in,the computation time and path length of the connection point method will be better than that of other planners.Second,inspired by the cooperation-competition relationships of subgroups in the natural world while resisting enemies and fighting for food,we introduce the concept of cooperation-competition mechanisms into the intelligent algorithm.A distributed intelligent algorithm,coevolution pigeon-inspired optimization with cooperationcompetition mechanisms is presented.The pigeon flock is divided into several subgroups,and every sub-group represents a robot.The cooperation mechanism ensures robots cooperate to complete the specified task,and the competition mechanism is used to cope with the potential conflicts between robots.Third,we also consider the range constraint of the robot for the first time,and a dynamic two-stage scheme is designed for multiple robots cooperative region search.In the region search stage,a coevolution pigeon-inspired optimization based on the cooperation-competition mechanism is used to maximize the search reward on the premise of avoiding collisions between robots.Robots will return to their respective bases under orientation constraints and range restrictions in the return stage.Fourth,this paper design a search tracking approach in view of the case that there is no grid shortest path algorithm with orientation constraints.This approach derives from the knowledge of region search,the cells contained in the grid path are modeled as key cells,and other blank cells are regarded as known regions.Every robot will search for maximum reward and track key cells under orientation constraints.The application object of search tracking approach is not only suitable for a single robot,but also can be extended to multiple robots.Lastly,we add the orientation constraints of the robot,the unexpected local environment changes,and failures for some robots to the multiple robots cooperative path planning problem.A cooperative path searching approach based on the decoupling idea is proposed,which consists of two coupling processes and can minimize the total travel distance for the robots.In the first phase,the connection point method is extended to multiple robots regardless of any restrictions.The cooperative search tracking approach is applied to a large number of robots,and the above constraints are considered in the second phase.