Improvements in the sensor technology and the emergence of connected vehicles and wireless communication have provided a base for the development of a new generation of vehicles with the ability to navigate on roads without human intervention. Autonomous vehicles use different sensor technologies such as LIDAR, ultrasonic, etc. to mimic the inception of human drivers of the surrounding environment in order to navigate on streets. Many applications have been envisioned for autonomous vehicles. Among those, the Autonomous Intersection Management (AIM) has received special attention.
AIM is a type of intersection control and management in which conflicting traffic (autonomous vehicles) are rerouted through an intersection with minimum delay and stops, using gaps between conflicting movements, without the requirement of traditional traffic signals.
In AIM, as developed by the University of Texas, the intersection can be managed at the local level by an algorithm that acts as a person responsible for managing conflicting movements of autonomous vehicles within the shared area of the intersection. Approaching autonomous vehicles to the intersection are able to send requests to the intersection manager asking for possibility of traveling through the intersection. The intersection manager receives all the requests from all the approaching autonomous vehicles from different conflicting movements and tries to finds non-conflicting trajectories for safe traverse of vehicles through the intersection. If the intersection manager finds a non-conflicting trajectory for a request made by a vehicle, the request is accepted and specific areas (cells) of the intersection is reserved at specific time for the requesting vehicles to travel through the intersection and the proper speed is advised to the vehicle to make to its reservation in time. The method takes into account the movements of non-autonomous vehicles by routing them through an intersection using the conventional traffic signals. In this case, autonomous vehicles are able to travel through the intersection at any time (even when the signal is red) while permitted by the intersection manager; however, non-autonomous vehicles can only travel through the intersection when the signal is green for their intended travel path.
There are several challenges facing the implementation of such a system. Examples include the penetration rate of autonomous vehicles and reliability of communication. As visible in the graph below, to achieve operational benefits over traditional intersections operated by traffic signals and conventional vehicles, a penetration rate of more than 90% of autonomous vehicles is required. The reliability of the communication platform is also important in such application as failure in wireless communication can result in collisions.