Computational architecture for autonomous decision-making in unmanned aerial vehicles

Shankar Sankararaman; Kai Goebel

doi:10.1117/12.2304506

Translator Disclaimer

10 May 2018 Computational architecture for autonomous decision-making in unmanned aerial vehicles

Shankar Sankararaman; Kai Goebel

Author Affiliations +

Proceedings Volume 10639, Micro- and Nanotechnology Sensors, Systems, and Applications X; 106391Y (2018) https://doi.org/10.1117/12.2304506
Event: SPIE Defense + Security, 2018, Orlando, Florida, United States

Abstract

This paper presents a computational architecture to facilitate autonomous decision-making under uncertainty for safe operation of drone-like vehicles. The proposed framework is based on identifying and predicting the occurrence of various risk-factors that affect the safe operation of such vehicles and estimating the likelihood of occurrence of these risk-factors. This analysis is then used to select trajectories for the operation of the vehicle. Feasible trajectories are classified into four different categories: nominal and safe, off-nominal but safe, unsafe and abort the mission, and unsafe and ditch the vehicle. An important challenge in the operation of drones is that there are several sources of uncertainty that affect their operation; these sources of uncertainty arise from wind conditions, imprecise future power-demands, inexact future trajectories, etc. Therefore, it is important to develop a decision-making framework that can incorporate all these sources of uncertainty and make decisions that are robust to the presence of such uncertainty. Potential risk-factors such as dynamic obstacles, battery drain, etc. are identified and the likelihood of occurrence of these risk-factors are predicted preemptively and proactively in order to facilitate risk-informed safety-assured decision-making.

Conference Presentation

Citation Download Citation

Shankar Sankararaman and Kai Goebel "Computational architecture for autonomous decision-making in unmanned aerial vehicles", Proc. SPIE 10639, Micro- and Nanotechnology Sensors, Systems, and Applications X, 106391Y (10 May 2018); https://doi.org/10.1117/12.2304506

ACCESS THE FULL ARTICLE