Controlling traffic flow by making federated decisions with Autonomous Vehicles

As Autonomous Vehicles (AV) increase and get connected with each other, they will be overwhelmed with information, much of which will be duplicate or redundant, and time and effort must be spent to compress it into manageable and usable bits. We propose to setup Intelligent Road Side Units (IRSU) to communicate with each AV and other suitably equipped vehicles in the traffic stream within their range. This releases the AVs from handling large amounts of data and allows them to allocate their resources to autonomous functions. The IRSUs enable the interaction of AVs with the infrastructure as well as enabling connectivity between each other. The architecture is completed by connecting these IRSUs to a Central System (CS) which can coordinate the decisions of each to achieve a common goal like throughput of traffic, avoidance of shocks etc. This framework allows the federation of decision making and learning and control of traffic with AVs. The CS controls the actions of the IRSUs each of which makes and communicates decisions to the individual AVs. The decisions made by each IRSU are independent of the decisions of other units, and are coordinated by the CS, achieving the full benefits of federation.

People

Project ID: 1005