Adaptive Model Predictive Control for Hybrid Energy Storage Systems in Mobile Robots under Coupled Disturbances

Abstract

High-voltage transmission line inspection robots frequently encounter complex obstacle-crossing conditions, resulting in high-frequency and drastic transient power demands. Traditional hybrid energy storage system (HESS) energy management strategies primarily focus on mechanical loads, significantly ignoring the high-frequency additional electromagnetic (EM) dissipation induced by strong spatial magnetic fields. This oversight often leads to severe bus voltage drops and rapid battery polarization aging under extreme conditions. To address this issue, an Adaptive Model Predictive Control (A-MPC) strategy fusing electromagnetic perception is proposed in this paper. First, a multi-physical coupled power demand model is established by integrating the mechanical dynamics with the spatial EM dissipation. Second, a dynamic A-MPC architecture is designed, utilizing the Sequential Quadratic Programming (SQP) algorithm to reconstruct the penalty weights of the cost function online based on real-time EM disturbance intensity. Simulation results under multi-modal composite conditions demonstrate that the proposed strategy exhibits superior transient robustness. It effectively truncates the peak transient current of the battery to 9.52 A and suppresses the bus voltage drop within 0.35 V. Furthermore, the equivalent cycle aging capacity of the battery is significantly reduced by 44.0% over the full life cycle, achieving a deep unification of transient dynamic anti-disturbance and long-term economic operation.