Skip to content. Skip to navigation
McGill Home SOCS Home
Personal tools
You are here: Home Announcements and Events Seminars profile

Seminars
Fall 2013 Schedule
Winter 2014 Schedule
Archive


         
     
 
2013/07/30, MC103, 12 - 12:30

Kinodynamic Non-Holonomic Motion Planning for UAVs: A Minimum Energy Approach
Nir Rikovitch , Master Student, Department of Mechanical Engineering, McGill

Abstract:

In recent years unmanned aerial vehicles (UAVs) have gained popularity for various applications, both civilian and military. As technology advanced, UAVs became cheaper, lighter and more readily available than their manned versions. Due to limited aircraft size and load carrying capabilities, battery capacity substantially reduces flight endurance. Therefore, generating flight paths for a UAV which take into account the dynamics of the vehicle, while at the same time minimizing the energy requirements is a worthwhile avenue to pursue. This seminar presents the minimum energy affine quadratic regulator (MEAQR) RRT* - an asymptotically optimal sampling-based motion planning algorithm for non-linear, non-holonomic, under-powered and under-actuated systems in a-priori known and static environments. The generated trajectory connects the initial and goal positions with a collision-free, dynamically feasible trajectory that adheres to the magnitude and bandwidth constraints of actuation, while minimizing energy consumption.

The algorithm is demonstrated on two example systems (1) a simple 2D pendulum with actuation constraints and (2) a quadrotor described by a 13-dimensional state-space model. In addition, an environment modeling procedure is designed, implemented and tested for a-priori map building required in order to test the algorithm on a real life environment.