无模型能量流调节的基于无源性的机器人旋转关节控制

Passivity-based control (PBC) is a powerful control strategy, excelling in stability through its energy-based framework. It is regarded as an effective method for enhancing the dependability of robotic rotary joints (RRJs). However, as a fundamentally model-based control method, PBC’s effectiveness heavily relies on accurate system modeling and parameter estimation. Consequently, its control performance is significantly degraded in RRJs, which exhibit complex nonlinear dynamics complicate parameter identification and frequently lead to parameter mismatches between theoretical models and physical systems. To address this limitation, this article presents a novel energy flow regulation-PBC (EFR-PBC) methodology. Leveraging the system’s polynomial characteristics, EFR-PBC incorporates a data-driven recursive polynomial response surface (RPRS) surrogate model to predict energy flow tensor (EFT) variations from input-output measurements, facilitating adaptive control law synthesis. By reducing dependence on precise mathematical models and system parameters, EFR-PBC maintains robust performance under uncertain conditions. Extensive comparative studies through both numerical simulations and experimental validations against conventional interconnection and damping assignment (IDA)-PBC and PID-PBC implementations demonstrate the effectiveness and superior performance of the proposed method.