đ€ AI Summary
Heavy-duty wearable and legged robots require high torque density, backdrivability, and energy-efficient power transmissionâchallenged by bulky, inefficient electric actuators and quasi-passive/underactuated systems with poor task adaptability.
Method: We propose a modular, reconfigurable hydrostatic actuation architecture that integrates hydrostatic principles with dynamically tunable fluidic cavities. Key innovations include microfluidic-embedded compliant chambers, flexible dielectric elastomer pressure-sensing arrays, and closed-loop fluid dynamic modeling coupled with model predictive control (MPC).
Contribution/Results: The system enables millisecond-scale co-regulation of stiffness, morphology, and load-bearing capacity. Under dynamic loads of 5â500 N, it achieves a stiffness tuning range of 1â10 MPa with millisecond response time and improves energy efficiency by 3.2Ă over conventional hydraulic/pneumatic systemsâovercoming their rigidity and high energy consumption. Experimental validation was conducted on a multi-terrain, load-carrying legged robot platform.