Integrated Task and Motion Planning for Robotic Systems


To develop a robotic system that will be able to accomplish high-level tasks specifications considering high-dimensional and nonlinear dynamics in the face of uncertain, unanticipated, and dynamically changing situations. The goal is to design autonomous robots for manipulation tasks which are able to work cooperatively in warehouses, housekeeping, and other robot assistant applications. In order to achieve this, we leverage modern Satisfiability Modulo Theories (SMT) solvers and trajectory synthesis techniques. Our basic idea is a counterexample-guided synthesis which combines logic inference with optimization to easy the computation due non-convexities from logic constraints.

Issues Involved or Addressed

Develop algorithms to synthesize robust trajectories and sequencing them for known environment to ensure a high-level task cooperatively, and to adapt the trajectories during run-time to deal with unknown obstacles or other agents in the environment. Implement these algorithms in robot such as Pioneer 3AT/3DX and Baxter.

Methods and Technologies

  • Cyberphysical Systems
  • Formal Methods
  • Linear Quadratic Regulator (LQR) Trees
  • Lyapunov Theory
  • Optimal Control
  • Signal Temporal Logic
  • Satisfiability Modulo Theories

Academic Majors of Interest

  • Electrical Engineering
  • Computer Science and Engineering
  • Aerospace and Mechanical Engineering

Preferred Interests and Preparation