Obstacle Avoidance for 2D Quadrotor with Hanging Load

Ayush Agrawal, Nathan Bucki, Prasanth Kotaru, David Meister, and Martin Xu
MEC231A Experiential Advanced Control Design I, Fall 2017

In this project, we use tools from optimal control to generate and stabilize feasible and safe trajectories for a two-dimensional model of a quadrotor with a cable suspended payload in cluttered environments. Due to the high degree of underactuation and nonlinear dynamics of the system, optimization-based techniques, for specifying control objectives and complex desired behaviors, are better suited over traditional control design methods for these systems. In particular, we develop our trajectory generation scheme using recent methods developed within the optimization-based collision avoidance framework, where obstacles are modeled as polyhedra and nondifferentiable collision avoidance constraints are reformulated into smooth nonlinear constraints. We then use a controller in the Model Predictive Control (MPC) framework to track that trajectory and compare it with an infinite-time horizon LQR controller.