Skip to content

Latest commit

 

History

History
88 lines (61 loc) · 5.33 KB

README.md

File metadata and controls

88 lines (61 loc) · 5.33 KB
Raw

Reachability-based Trajectory Design with Neural Implicit Safety Constraints

Project Page | Paper | Dataset

Introduction

We present Reachability-based Signed Distance Functions (RDFs) as a neural implicit representation for robot safety. RDF, which can be constructed using supervised learning in a tractable fashion, accurately predicts the distance between the swept volume of a robot arm and an obstacle. RDF’s inference and gradient computations are fast and scale linearly with the dimension of the system; these features enable its use within a novel real-time trajectory planning framework as a continuous-time collision-avoidance constraint. The planning method using RDF is compared to a variety of state-of-the-art techniques and is demonstrated to successfully solve challenging motion planning tasks for high-dimensional systems faster and more reliably than all tested methods.

Paper: Reachability-based Trajectory Design with Neural Implicit Safety Constraints. [Arxiv]

Dependency

  • Run pip install -r requirements.txt to collect all python dependencies.
  • IPOPT is required to run the planning experiments.
  • Gurobi is optionally required if the users would like to generate datasets. Generated datasets are also available at Google Drive.
  • WANDB is optionally required if the users would like to train the RDF models. Pretrained RDF models are available in trained_models/.
  • CORA 2021 is optionally required for users to compute JRS in reachability/joint_reachable_set/gen_jrs_trig with MATLAB script.

Reproducing Results

Building Datasets

Datasets are available at Google Drive. After downloading the datasets, put them under dataset/.

Users who would like to run building dataset on their own can run bash scripts/generate_datasets.sh in the repo home directory (i.e., rdf/).

Note that building the datasets generally takes days of time. To collect the datasets more efficiently, the paper chose to launch dataset collecting program (*.py in generate_dataset/) with multiple processes using different random seeds, then combine these sub-datasets together.

Users can follow similar practice by modifying the provided script and launching dataset collecteing programs in parallel or across different machines.

Training RDF Models

Pretrained models are available in trained_models/.

If the users would like to run training on their own, an example script to train a 3D7Links Manipulator is provided in scripts/train_rdf_model.sh. Note that WANDB is used to monitor training and is thus a dependency to run training.

Users can follow the guide from WANDB to install it, and then run bash scripts/train_rdf_model.sh to launch training with the provided hyperparameters in the training script. Users are free to tune the hyperparameters for model performance. The hyperparameters we used to obtain the pre-trained models are present in the same directory as the models.

Run Planning Experiments

IPOPT is required to run the planning experiments as the framework to solve non-linear programming optimization problems.

To reproduce the 2D planning experiments, run bash scripts/run_2d_planning.sh.

To reproduce the 3D planning experiments, run bash scripts/run_3d_planning.sh.

The results will be in planning_results/ as generated by the planning program.

Other Experiments

Here are the procedures to reproduce the other experiments in the paper.

  1. Compare RDF with SDF
cd experiments
bash run_compare_rdf_and_sdf.sh
  1. Compare RDF with QP
cd experiments
bash run_compare_time_with_QP.sh
  1. Evaluate RDF models on testset

Download the testsets from Google Drive and create a directory named test_dataset/ under the repo home directory (i.e., rdf/). Put the datasets in rdf/test_dataset/.

Then run

cd experiments
bash run_evaluate_model_on_testset.sh

Credits

  • reachability/ referred some part of CORA.
  • environments/robots/urdf_parser_py is extracted from urdf_parser_py and modified to our end.

Citation

The paper with an overview of the theoretical and implementation details is published in Robotics: Science and Systems (RSS 2023). If you use RDF in an academic work, please cite using the following BibTex entry:

@misc{michaux2023reachabilitybased,
      title={Reachability-based Trajectory Design with Neural Implicit Safety Constraints}, 
      author={Jonathan Michaux and Qingyi Chen and Yongseok Kwon and Ram Vasudevan},
      year={2023},
      eprint={2302.07352},
      archivePrefix={arXiv},
      primaryClass={cs.RO}
}