Modelling Safe Landing Zone Detection Options to Assist in Safety Critical UAV Decision Making

Abstract

Unmanned Aerial Vehicles (UAVs) have the potential to significantly enhance current practice for a range of applications including emergency response in the aftermath of natural disasters and crowd surveillance during major sporting events. However, before this potential can be realized there are a number of safety related technical challenges which must be addressed including provision of a Safe Landing Zone (SLZ) detection algorithm which would be executed in the event of a UAV emergency. In the event of such an emergency a key consideration of any safety related algorithm is remaining flight time which can be influenced by battery life and weather conditions. Therefore within this paper we present preliminary work in modelling the execution time of three SLZ detection options, one of which incorporates a human-in-the-loop. While it may be desirable to always involve a human-in-the-loop in decision making concerning the optimal SLZ, this, and alternative options involving collaboration with other UAVs may not be feasible given the constraint of remaining flight time. The models discussed are subsequently used in conjunction with an estimate of the UAV's remaining flight time to assist in autonomous decision making upon occurrence of a safety critical event.