Decentralized flocking of multi-robot systems with connectivity maintenance

Connectivity maintenance of flocking is vital for stability and fast state convergence of the entire multirobot system. To this end, the problem of flocking of multiple nonholonomic wheeled mobile robots with connectivity maintenance on directed communication graphs is investigated, provided that the initial network is strongly connected but not balanced. First, a new class of smooth and bounded artificial potential fields is developed to satisfy the requirements of connectivity maintenance, collision avoidance and distance stabilization, simultaneously. Furthermore, by combining potentialbased gradient control techniques and the consensus seeking control schemes, we propose a set of smooth and bounded control protocols for leaderless flocking or flocking with local leader-follower with connectivity maintenance. Those control protocols not only avoids the chattering and saturation caused by the discontinuous/nonsmooth and unbounded flocking algorithms, but also extend the traditional connectivity-preserving flocking algorithms for both the individual motion model and the network topology types. Finally, nontrivial simulations and experiments are performed to validate the effectiveness of the proposed smooth and bounded control strategies.