Distributed formation control with multiple systems' constraints: from geometry to distribution

In his thesis, Jin Chen addresses several problems of formation control with multiple system' constraints. Firstly, Chen consider the formation control problem where agents have heterogeneous sensing mechanism, distance-based or bearing-only. Leveraging heterogeneous sensing rigidity framework to define the desired shape and using the persistence notion for heterogeneous sensing formation, he explores gradient-based control laws for achieving the desired formation with heterogeneous sensing information. The design and analysis of distributed heterogeneous control begin with a core group of agents capable of maintaining the formation, and the system is scaled up by adding agents one at a time to the network. 

Chen also proposes a distance-based distributed formation-motion control law for unicycle agents. By applying recent advances in transversal uniform exponential stability, he proves the local exponential convergence of the unicycles to the desired rigid formation, where all agents move together at a constant reference speed. Chen also developed a range-only distributed formation control law for a multi-agent system. To obtain accurate relative bearing estimates for each agent, a continuous-time bearing estimator has been designed. Using contraction theory, Chen establishes the exponential convergence of the proposed observer. 

Finally, Chen introduces a novel concept called cloud formation control, which extends traditional formation control from precise geometry to flexible distribution, allowing for the representation of non-strict collective shapes. To achieve the desired dispersion formation, he proposes two control strategies: centralized and decentralized. For both approaches, Chen demonstrates exponential convergence of the multi-agent system.

Dissertation: https://hdl.handle.net/11370/d52c7760-4e4a-4880-91a9-d0b9d989d567