Haptic support for avoiding static and dynamic obstacles in UAV tele-operation

Abstract

Impoverished sensory input makes tele-operation of Unmanned Autonomous Vehicles a difficult task. Automation support can provide assistance to the operator, but may also produce automation surprise and a risk of loss of situation awareness, when the human operator fails to notice the actions of automation in high-workload situations. Previous work applied haptic feedback based on an artificial risk field to assist in avoidance of static obstacles with a small helicopter UAV. An off-line analysis of that solution shows that it would not be sufficiently effective for the avoidance of dynamic, moving obstacles. A new haptic assist algorithm, based on velocity obstacle theory was developed. This algorithm was first tested in off-line analyses, in scenarios with only static obstacles and in scenarios with a dynamic obstacle added. The off-line analysis showed the feasibility of the algorithm. The algorithm was then ported to a real-time environment and evaluated in pilot-in-the loop simulations, to verify the solutions and investigate the acceptability of the haptic feedback. Results indicate that the velocity obstacle approach works for both stationary and moving obstacles, in most, but not all scenarios.