Model Predictive Control based Collision Avoidance Controller for Octocopter

Octocopter are MIMO systems and have on-board flight controller to maintain flight stability. These flight controllers are usually based on cascaded PID controllers which may not offer satisfactory performance while carrying out complex tasks like collision avoidance. Also the external disturbances and constraints severely effects the 6-DOF states and increases the controller design quite challenging. In proposed work, Linear model predictive control (LMPC) based collision avoidance controller addressing the challenges of controller design for MIMO plants. The design is based on collision cone approach and have potential of increasing the efficiency of the flight controller by reducing computation efforts caused due to cascaded PID loops. The strategy is suitable to deal with real time challenges such as external disturbances and constraints in collision avoidance controller. Simulation results reveal the efficacy of LMPC based collision avoidance controller for slow and fast varying disturbances. Formulation of LMPC also suggests that, it has a advantage over conventional PID control strategy for MIMO system.