Distributed Control of Multi-Agent Systems via Linear Quadratic Differential Games with Partial Information

Abstract

A multi-agent system consisting of linear heterogeneous agents is considered in this paper. Distributed control laws for each agent are designed through the formulation of linear quadratic differential games with partial information. Exact and approximate solutions for the differential games are provided before the problem of formation control with limited communication is considered. A numerical example is provided to illustrate the theory.

A Game Theoretic Framework for Distributed Control of Multi-Agent Systems with Acyclic Communication Topologies

“…This setting is similar to the one considered in [19] and is such that, in the full information case, the "classical" differential game, i.e. the differential game defined by cost functionals (6), subject to the dynamics of the global system (2), is recovered.…”

“…[12]- [14] and references therein). Some of these approaches are based on differential game theory [15]- [19], which studies the strategic interactions of individual decision-making players. These players affect, through the selection of their strategies, the behaviour of a common dynamical system.…”

A Hybrid Controller for Multi-Agent Collision Avoidance via a Differential Game Formulation

“…exploration [1] and autonomous monitoring [2], [3]. Several approaches to solve control problems related to multi-agent systems involve concepts borrowed from game theory, for instance, in the context of motion planning, pursuit-evasion problems and formation control (see, e.g., [4], [5], [6], [7], [8]). However, differential games are, in general, difficult to solve.…”