DWPP: Dynamic Window Pure Pursuit for Robot Path Tracking Considering Velocity and Acceleration Constraints

DWPP is a pure pursuit variant that computes velocity commands in the velocity space (the v–ω plane) and selects the command within the dynamic window closest to the path-tracking condition ω = κv, so that the commands always respect the velocity and acceleration constraints.

🟢 Nav2 implementations

• Official Nav2 (ROS 2 Lyrical and later): DWPP is integrated into the official Nav2 repository. 👉 https://github.com/ros-navigation/navigation2
• ROS 2 Humble / Jazzy: a standalone Nav2 plugin is available. 👉 https://github.com/Decwest/nav2_dynamic_window_pure_pursuit_controller

Simulator quickstart

The kinematic simulator used for the simulation studies in the paper lives in simulator/. It is a self-contained Python package (no ROS required) that reproduces the paper's tables and figures deterministically.

# run all studies (outputs under results/simulation/)
PYTHONPATH=. uv run python simulator/run_studies.py --study all

# run a single study
PYTHONPATH=. uv run python simulator/run_studies.py --study smoke  # smoke|app|rpp|noise|lookahead|isotime

# run the minimal assertions
PYTHONPATH=. uv run python simulator/test_simulator.py

See simulator/README.md for details.

Folder Structure

• simulator/ Kinematic simulator and simulation studies used in the paper (see above).

• scripts/ Simulation scripts for various pure pursuit methods from the conference-paper experiments. You can refer to these to understand how DWPP can be implemented.

• results/ Result figures and tables used in the paper.

Citation

If you use this code, please cite the following paper:

Fumiya Ohnishi, Masaki Takahashi, “Dynamic Window Pure Pursuit for Robot Path Tracking Considering Velocity and Acceleration Constraints”, Proceedings of the 19th International Conference on Intelligent Autonomous Systems, Genoa, Italy, 2025.