Scalability in urban autonomy through overlapping decompositions of hybrid-state systems

This study investigates the issue of scalability for autonomous vehicles and systems in urban environments. As the number of agents such as infrastructure elements and communication-capable vehicles increases, the systems that are designed to sense/track/control these agents get burdened under the increased number of tracks/measurements. By employing the concept of overlapping decompositions, it is possible to divide and parallelize the workload, with limited degradation of overall system performance. The paper demonstrates this methodology on simple urban scenarios, involving autonomy of infrastructure elements and autonomous vehicle interactions.