Deadlock resolution for navigation of wheeled robots in continuous state-space

In this paper a method for preventing deadlocks between two vehicles is proposed, which avoids partitioning the state space into a discrete system. We propose there are advantages in some situations for solving the deadlock problem from a continuous state space perspective. A hierarchy between the vehicles is enforced, and by always ensuring the secondary vehicle allows the primary vehicle to have a convergent trajectory, deadlocks can be avoided in most situations. By introducing a recovery scheme which may be invoked in extraordinary conditions, we are able to prove the vehicles converge to their respective targets. Simulations confirm the validity of the approach.