A Comprehensive Investigation of Control Strategies for Soft Robots: Challenges, Methods, and Future Directions

Abstract

Soft robots have shown broad application prospects in medical and detection fields by virtue of their intrinsic compliance and human- computer interaction safety, but the nonlinear deformation and multi-source disturbances brought about by their hyperelastic materials have led to precise control becoming the core bottleneck restricting their engineering. This paper conducts a systematic study on the control technology of soft robots, firstly sorts out the core control difficulties of three mainstream drive methods, such as fluid drive, intelligent material drive, and cable drive, and comprehensively summarizes the technical paths, applicable scenarios, and limitations of existing control strategies such as open-loop feedforward, improved PID, model predictive control, and data-driven learning. secondly, it deeply analyzes the breakthrough potential of emerging technologies such as perception-control integration and physical-data fusion; Finally, the core bottlenecks and future development directions of current research are clarified. The results show that the existing control strategies still have obvious deficiencies in modeling accuracy, robustness and real-time, and the fusion of physical priori and data-driven hybrid control is the key direction to solve the high-precision control problem of soft robots. This study provides a systematic theoretical reference and technical context for the engineering application of soft robot control technology.