Collision-Free Multi-Agent Coverage Control for Non-Cooperating Swarms: Preliminary Results

TL;DR

The Optimal Reciprocal Collision Avoidance method used for safe navigation in multi-agent scenarios is adapted to suit the needs of Voronoi-based coverage control with more than one swarm.

Abstract

The main contribution of this paper is a methodology for multiple non-cooperating swarms of unmanned aerial vehicles to independently cover a common area. In contrast to previous research on coverage control involving more than one swarm, this paper does not assume cooperation between distinct groups but considers them as entirely independent units following their own objectives. Using Voronoi tesselation, collision-free motion of agents within the same swarm has been proved before. However, as is shown in Example 1 of this paper, in the case of multiple swarms with inter-swarm but without intra-swarm collaboration, these guarantees do not hold. We address this issue by proposing an algorithm to achieve maximum coverage with multiple swarms while avoiding collisions between agents. Thus, the Optimal Reciprocal Collision Avoidance method used for safe navigation in multi-agent scenarios is adapted to suit the needs of Voronoi-based coverage control with more than one swarm. The functionality of the proposed technique is validated through Monte Carlo simulations.

Figures (5)

• Figure 1: Voronoi tesselation of an area $Q$ containing four generators
• Figure 2: Simulation of coverage control with two non-cooperating swarms without collision avoidance showing the trajectories and final positions of the agents of swarm 1 (S1) and swarm 2 (S2) (a) at the first iteration (b) after 50 iterations (c) after convergence to their final positions. Notice that the red agents are hidden behind the blue ones.
• Figure 3: Geometrical interpretation of $\mathcal{VO}_{AB}^{\tau}$ and half-plane $ORCA_{AB}^{\tau}$
• Figure 4: Initial positions of agents tested in Monte Carlo simulations
• Figure 5: Simulation of coverage control with two non-cooperating swarms with collision avoidance showing the trajectories and final positions of the agents of swarm 1 (S1) and swarm 2 (S2) (a) at the first iteration (b) after 50 iterations (c) after convergence to their final positions.

Theorems & Definitions (2)

• Example 1
• Example 2