A constraint-based flight control system architecture for UAVs using the iTaSC framework

Abstract

(Semi-) autonomous complex UAV missions, such as inspection or search-and-rescue in uncertain dynamic environments, require obstacle avoidance and operator shared control. Combining humans' cognitive abilities with fast automation is the key for such missions. This paper presents a flight control system architecture based on the instantaneous Task Specification using Constraints (iTaSC) methodology and software framework. iTaSC is a flexible constraint-based programming approach that generates a robot motion at runtime which automatically derives the input for a low-level controller taking into account constraints and intentions from the operator, obstacles and mission constraints. This setup is experimentally validated by navigating a multirotor UAV safely through a GPS-denied corridor using (intuitive) shared control with a pilot. In addition to the pilot's commands, automatic obstacle avoidance and object tracking are performed real-time through various onboard sensors and with limited onboard computational power.