Embedded Files
Olov Andersson, Oskar Ljungqvist, Mattias Tiger, Daniel Axehill, Fredrik Heintz
Linköping University
We proposed a general optimization-based receding-horizon lattice-based motion planning framework with collision avoidance functionality for both complex 3D environments and moving obstacles. This includes planning with both dynamics and time, such that the quadcopter can plan trajectories around, and move out of the way of, other agents. We also demonstrated real-time performance on a difficult warehouse scenario with multiple moving obstacles, both people and other UAVs flying at varying altitudes. To the best of our knowledge, no other optimization-based and dynamically-feasible approach has demonstrated this capability in real-time.
We proposed a general optimization-based receding-horizon lattice-based motion planning framework with collision avoidance functionality for both complex 3D environments and moving obstacles. This includes planning with both dynamics and time, such that the quadcopter can plan trajectories around, and move out of the way of, other agents. We also demonstrated real-time performance on a difficult warehouse scenario with multiple moving obstacles, both people and other UAVs flying at varying altitudes. To the best of our knowledge, no other optimization-based and dynamically-feasible approach has demonstrated this capability in real-time.
Accepted to the 57th IEEE Conference on Decision and Control, 2018.
Accepted to the 57th IEEE Conference on Decision and Control, 2018.
Illustration of the motion planning and control architecture that is used for the quadcopter platform in the simulation experiments. The modules that are colored in blue are considered in this work.
Supplementary Material
Supplementary Material
Simulation results. Several different scenarios are show-cased.
Google Sites
Report abuse