Combined Task and Motion Planning for Mobile Robots


Addressing the combined task and motion-planning problem is becoming increasingly important as a growing number of diverse robotics applications in navigation, search-and-rescue, manipulation, and surgical procedures involve reasoning with both discrete actions and continuous motions. Towards this end I perform research that brings together aspects from both the AI and Robotics communities into a framework that couples sampling-based motion planning with discrete solvers to generate dynamically-feasible collision free trajectories while considering high level task specifications (via. Regular Languages, Linear Temporal Logic, etc.).

Project Aims

  1. Reduce the cognitive workload of human operators by adapting high-level planning formalisms to form the basis of a mission specification language
  2. Enhance the autonomy by constructing a hybrid planning framework that combines high-level planning with sampling based motion planning
  3. Develop a means for the robot to recover and adapt to changing contextual conditions by developing robust replanning capabilities

Key Publications

[pdf] James McMahon and Erion Plaku. Robot motion planning with task specifications via regular languages. Robotica, 35(1), 26-49, 2015. [ bib ]

[pdf] James McMahon and Erion Plaku. Sampling-based tree search with discrete abstractions for motion planning with dynamics and temporal logic. 2014 IEEE/RSJ International Conference on Intelligent Robots and Systems, Chicago, IL. pp. 3726-3733, 2014. [ bib ]