| # CaDeLaC Datasets | |
| Datasets used to train the models for the **Context-Aware Deep Lagrangian Networks for Model Predictive Control (CaDeLaC)**. | |
| ## Panda Datasets | |
| - `panda_mj_101_rand_envs_20_runs_50Hz_lqr`: 100 environments with randomly sampled loads at the end-effector, along with one environment with only the robot. Each environment contains 20 runs with random initialization and joint references, for a total of 2020 runs. | |
| - `panda_mj_nominal_env_20_runs_50Hz_lqr`: 20 runs of a single environment only with the robot without any load. | |
| If you find this dataset useful, please consider citing: | |
| ``` | |
| @misc{schulze2025_cadelac, | |
| title={Context-Aware Deep Lagrangian Networks for Model Predictive Control}, | |
| author={Lucas Schulze and Jan Peters and Oleg Arenz}, | |
| year={2025}, | |
| eprint={2506.15249}, | |
| archivePrefix={arXiv}, | |
| primaryClass={cs.RO}, | |
| url={https://arxiv.org/abs/2506.15249}, | |
| } | |
| ``` | |
| For more information, please refer to the code repository: [https://github.com/Schulze18/cadelac](https://github.com/Schulze18/cadelac). | |
| ## License | |
| MIT |