Driving MPC

trajectory optimization and model predictive control for autonomous driving

This is my Fall 2022 project at the Robotic Exploration Lab at CMU. Some key points:

  • Employed ALTRO, an augmented Lagrange iLQR algorithm, to solve constrained trajectory optimization problems for autonomous driving applications, using bicycle models.
  • Investigated local planning and control frameworks, such as time-varying LQR and model predictive control (MPC) using OSQP, to ensure safe and efficient operation, while handling control limits and obstacles.
  • Algorithms and 3D visualization are implemented in Julia.
Replan to avoid obstacles and reach the goal.