Aggressive Motion Planning for a Quadrotor System with Slung Load Based on RRT

Recent advances in technologies related to Un-manned Aerial Vehicles have led to applications in challenging tasks such as carrying a payload through a small window. Common approaches to solve this task are based on Reinforcement Learning, Quadratic Programming, and Model Predictive Control. Although they provide optimal motion planning, they rely on the necessity of a quadratic cost function or linear constraints for the obstacles. We propose to use a variant of the Rapidly-exploring Random Tree algorithm to solve the task without relying on convex constraints, tuning weights for cost functions or linear constraints. We show that our approach was able to solve the task of crossing a small window with a slung load, and we suggest future modifications to accelerate the convergence rate of the algorithm. The main advantage of the proposed approach is the capability finding a collision-free trajectory for the whole system that is not restricted to linear constraints, and does not depend on cost functions.