Robust Predictive Motion Planning by Learning Obstacle Uncertainty

This is the code for the article

@article{zhou2025robust,
  title={Robust predictive motion planning by learning obstacle uncertainty},
  author={Zhou, Jian and Gao, Yulong and Johansson, Ola and Olofsson, Bj{\"o}rn and Frisk, Erik},
  journal={IEEE Transactions on Control Systems Technology},
  year={2025},
  pages={},
  doi={10.1109/TCST.2025.3533378},
  publisher={IEEE}
}

Jian Zhou and Erik Frisk are with the Department of Electrical Engineering, Linköping University, Sweden.

Yulong Gao is with the Department of Electrical and Electronic Engineering, Imperial College London, United Kingdom.

Björn Olofsson is with the Department of Automatic Control, Lund University, Sweden.

Contact: zjzb1212@qq.com

Packages for running the code

The programming is by Python. To run the code you need to install the following key packages:

CasADi: https://web.casadi.org/

HSL Solver: https://licences.stfc.ac.uk/product/coin-hsl

pytope: https://pypi.org/project/pytope/

Note: To install the HSL package can be a bit comprehensive, but the solvers just speed up the solutions. You can comment out the places where the HSL solver is used, i.e., "ipopt.linear_solver": "ma57", and just use the default linear solver (mumps) of CasADi.

Introduction to the files

I. In 1_Reach_Avoid folder:
  (1) main.ipynb is the main file for simulation.
  (2) ModelingSVTrue.py models the SV.
  (3) Planner_P.py is the Proposed method.
  (4) Planner_D.py is the Deterministic method.
  (5) Planner_P.py is the Robust method.
  (6) The data is saved in the folder plot for reproduction of the results in the paper.

II. In 2_Active_Evasion folder:
  (1) main.ipynb is the main file for simulation.
  (2) ModelingSVTrue.py models the SV.
  (3) Planner_P.py is the Proposed method for motion planning of the EV.

III. In 3_Highspeed_Overtaking folder:
  (1) main.ipynb is the main file for simulation.
  (2) ModelingSVTrue.py models the SV.
  (3) Planner_P.py is the Proposed method for motion planning of the EV.   (4) cdf_f.mat and cdf_x.mat save the distribution information of the SV trained from a real-world dataset. This information is used to model the SV in the overtaking scenario.

IV. In 4_Encounter_Scenario_With_Same_Control folder:
  (1) main.ipynb is the main file for simulation.
  (2) ModelingSVTrue.py models the SV.
  (3) Planner_EV.py is the Proposed method for motion planning of the EV.   (3) Planner_SV.py is the Proposed method for motion planning of the SV.   (4) In this case both the EV and SV use the same strategy.

V. In 5_rounD_Scenario folder:
  (1)main.ipynb is the main file for simulation.
  (2)Planner_P.py is the Proposed method for motion planning of the EV.
  (3) EV_Data.npy, SV0_Data.npy, and SV1_Data.npy save the data of the involved vehicles in the rounD dataset scenario.

Remarks

I. The code for generating the animations has been removed as a result of version compatibility of Python.

II. The code for hardware experiments can be easily designed based on the published code here.