Distributed control for complex missions: Quasi‐linear parameter varying approach

Abstract

A distributed flocking control scheme is presented for a swarm of non-holonomic vehicles modeled as linear parameter varying (LPV) systems.The vehicles are assumed to perform tasks that arise when exploring scalar fields on R 3 , which may represent, for example, concentration levels of a toxic substance in air or water. Here we consider in particular level surface tracking and level surface monitoring missions. It is assumed that agents can sense gradient and Hessian information of the scalar field at their current position. The control architecture of each vehicle is structured into two modules: a flocking filter which represents a virtual layer of the network, and local tracking controllers for vehicles on the physical layer. The flocking filter receives data from neighboring vehicles and generates a reference signal. An LPV tracking controller for the vehicles, and a network flocking filter that stabilizes the whole network, are designed separately. Simulated mission scenarios with swarm of autonomous underwater vehicles demonstrate the practicality of the proposed approach.