An unmanned aerial vehicle (UAV) must be able to safely reach its destination even, when it can only gather limited information about its environment. When an obstacle is detected, the UAV must be able to choose a path that will avoid collision with the obstacle. For the collision avoidance scheme, we apply the velocity obstacle approach since it is applicable even with the UAV's limited sensing capability. To be able to apply the velocity obstacle approach, we need to know the parameter values of the obstacle such as its size, current velocity and current position. However, due to the UAV's limited sensing capability, such information about the obstacle is not available. Thus, by evaluating sensor readings, we get the changes in the possible positions of the obstacle in order to generate the velocity obstacle and make the UAV choose a collision-free trajectory towards the destination. We performed simulation on different obstacle movements and the collision-free trajectory of the UAV is shown in the simulation results.

No abstract

A scheduling scheme for autonomous intersection crossing is proposed and evaluated. The scheduling scheme determines the flow of autonomous vehicles into an intersection without traffic signals. The objective of the scheduling scheme is to schedule the vehicles’ entrances into the intersection such that there will be no collision among vehicles and the intersection is efficiently utilized. Our scheduling scheme uses reservation-based scheduling approach, and the scheduling is formulated as an optimization problem to find the best sequence of vehicles’ entrances into the intersection. Our scheme is made more practical by allowing the vehicles to move at any speed within a speed range, and it is shown that it is fast enough to be used in real time. Through simulation, it is also shown that our scheduling scheme significantly outperforms the first-come, first-served approach, which is the approach used by other reservation-based scheduling schemes.

No abstract

No abstract