Evaluation of tracking performance and robustness for a hybrid locomotion controller
Data files
Dec 11, 2023 version files 178.67 MB
Abstract
Legged locomotion is a complex control problem that requires both accuracy and robustness to cope with real-world challenges. Legged systems have traditionally been controlled using trajectory optimization with inverse dynamics. Such hierarchical model-based methods are appealing due to intuitive cost function tuning, accurate planning, generalization, and most importantly, the insightful understanding gained from more than one decade of extensive research. However, model mismatch and violation of assumptions are common sources of faulty operation. Simulation-based reinforcement learning, on the other hand, results in locomotion policies with unprecedented robustness and recovery skills.
Yet, all learning algorithms struggle with sparse rewards emerging from environments where valid footholds are rare, such as gaps or stepping stones. In this work, we propose a hybrid control architecture that combines the advantages of both worlds to simultaneously achieve greater robustness, foot-placement accuracy, and terrain generalization. Our approach utilizes a model-based planner to roll out a reference motion during training. A deep neural network policy is trained in simulation, aiming to track the optimized footholds. We evaluate the accuracy of our locomotion pipeline on sparse terrains, where pure data-driven methods are prone to fail. Furthermore, we demonstrate superior robustness in the presence of slippery or deformable ground when compared to model-based counterparts. Finally, we show that our proposed tracking controller generalizes across different trajectory optimization methods not seen during training. In conclusion, our work unites the predictive capabilities and optimality guarantees of online planning with the inherent robustness attributed to offline learning.
README: Usage Notes
Datasets and code necessary to replicate all figures used in the article "DTC: Deep Tracking Control". The zip folder contains the following directories:
- foothold_tracking: Contains data for evaluating tracking performance. Used to generate Fig. 3 B.
- foothold_tracking_MPC: Contains data for evaluating tracking performance when deployed with a different MPC method. Used to generate Fig. 3 C.
- optimizer_frequency: Contains data for evaluating success rate as a function of different update frequencies of the planner. Used to generate Fig. 7 A.
- simulation_study: Same data as above, but separated into different terrain types. Used to generate Fig. 7 B and Fig. 7 C.
- tensor_board: Contains training curves. Used to generate Fig 7.D.
Each of the above folders contains a main.m file. To generate the associated figures, simply run that file with MATLAB.
Methods
This dataset was generated using the algorithm described in the article titled "DTC: Deep Tracking Control". The data was collected either on the real robot (stored in rosbags), or in simulation (stored directly in Matlab files).