RandUP-RRT

Accompanying repository of Robust-RRT: Probabilistically-Complete Motion Planning for Uncertain Nonlinear Systems

Installation

Minimum requirement

1. Python 3.7 or above

Setting up a virtual environment (optional)

1. Run python3 -m venv $ENV_PATH
2. Activate the virtual environment with source $ENV_PATH/bin/activate

Setting up randUP_RRT

1. Clone this repository from https://github.com/StanfordASL/randUP_RRT.git
2. From randUP_RRT, run pip install .

Running Provided Examples

If randUP_RRT was set up in a virtual environment, the virtual environment must be active in the terminal used to run the examples.

For all planning examples, use the optional argument --num_particles=1 to switch to RRT planning. Use --num_particles=i for planning with i RandUP particles. By default, planning is done with RandUP-RRT with the default number of particles.

Nonlinear Quadrotor

• Run python3 planning/quadrotor_examples/plan_quadrotor.py to plan and produce Figure 3.

• Run python3 planning/quadrotor_examples/generate_quadrotor_statistics.py to produce Table 1. This script does not take in num_particles.

Planar Pusher

• Run python3 planning/planar_pusher_examples/feedback_pusher.py to plan and visualize the planar pusher with PyBullet.

Hybrid Jumping Robot

• Run python3 planning/hybrid_integrator_examples/plan_hybrid_integrator.py to plan and produce Figure 5.

Repository Structure

• hybrid_integrator, planar_pusher, quadrotor_planner define the respective physical systems
• planning contains all planning-related code
  • *_examples contain the scripts for running examples
  • randup_rrt.py implements RandUP-RRT