Deep Learning Methods for Soft Robot Control: Data-Driven Modeling, Policy Learning, and Dynamic Prediction

Soft robots, relying on compliant materials, exhibit inherent safety and environmental adaptability in complex environments, but their strong nonlinear characteristics make modeling and control particularly challenging. This paper employs a literature review and comparative analysis approach to summarize research on deep learning-based control of soft robots, focusing on data-driven modeling, policy learning, and dynamic prediction. The results show that deep networks can serve as surrogates for numerical models, reinforcement learning facilitates the learning of complex control strategies, and time-series models improve dynamic prediction capabilities. The integration of physical priors with deep learning, as well as learning mechanisms focused on safety and few-shot learning, will be important directions for promoting the practical application of soft robots.