Negotiation of drivable areas of cooperative vehicles for conflict resolution

Abstract

We address the problem of cooperative conflict resolution for multi-vehicle motion planning in mixed-traffic scenarios, where automated and manually-driven vehicles co-exist. We propose a novel solution based on reachability analysis, which provides the drivable area of each collaborative traffic participant. Overlapping drivable areas are redistributed so that each traffic participant receives an individual area for motion planning. We do not stipulate a specific method for predicting the future motion of non-communicating traffic participants. Furthermore, uncertainties in the initial states of the cooperative vehicles, e.g. due to sensor noise, can be easily integrated. A byproduct of our approach is that collaborative groups can be automatically found by identifying conflicting drivable areas; if no conflict exists, collaboration becomes unnecessary. We demonstrate the redistribution of drivable areas with two numerical examples.