An Event-Triggered Distributed Optimization Approach And Its Application To Signal Source Localization

At present,the research and application of multiple robot cooperative control have developed rapidly in the fields of national defense,military,and daily life.The intelligent robot cooperative control algorithm makes the multi-robot system achieve robust,intelligent and scalable.However,due to the expansion of control tasks,the limitation of communication bandwidth resources and the restriction of computer calculation power,the application of multi-robot systems with continuous time communication control algorithm has limitations.How to reduce the update frequency of the controller of the mobile robot with cost function and relieve the communication pressure have attracted more and more attention.In this paper,two-wheeled differential mobile robot was taken as the research object.An event-triggered distributed optimization algorithm was designed to address the convex optimization problem with equality constraints in a multi-robot system.The finite-time and fixed-time consensus algorithms were also considered.The main work is as follows:Firstly,the dynamic model of mobile multi-robot system was analyzed and the control input was transformed into the speed.Secondly,the event-triggered time sequence was analyzed based on the undirected graph theory.Furthermore,an event-triggered finite-time consensus distributed optimization algorithm was designed.This method not only saved the communication resources among mobile robots,but also dealt with the convex optimization problem with cost function.Then,based on the correlation theorems and lemmas,the stability condition of system in finite-time was calculated.In the simulation of signal source location,the event-triggered finite-time distributed optimization algorithm combined with a particle filter algorithm was applied to evaluate the feasibility of the event-triggered rate and the accuracy of the signal source localization.Then,an event-triggered fixed-time consensus distributed optimization algorithm was designed.The consistency of a multi-robot system based on this algorithm was completed within a fixed time with high practicability,and was not affected by the initial position.Finally,the event-triggered finite-time consensus algorithm was applied to the real signal source localization experiment with multiple wheeled differential mobile robot.This article independently designed the corresponding hardware and software implementation schemes.What's more,the event-triggered rate and the accuracy of signal source location of the multi robot cooperative control algorithm were analyzed through comparative experiments.Results show that the event-triggered finite-time consensus distributed optimization has a higher control accuracy and a better utilization of communication resources.