Energy-saving control framework for lunar rover driving on low-gravity and rough roads

The lunar rover (LR) is essential equipment for lunar surface exploration; however, energy efficiency remains a critical factor restricting its driving range on rugged lunar terrain. To address this issue, an energy-saving control strategy integrating offline optimization with clustering statistics is proposed, aiming to satisfy the requirements of low computational complexity and real-time operation inherent to the LR system, while closely approximating an optimal energy-efficient operational sequence determined offline. Specifically, dynamic programming (DP) is employed to derive the optimal energy-saving trajectory for LR velocity and the corresponding energy distribution within the powertrain system along a predetermined route. Subsequently, an energy-efficient coupling relationship among LR velocity, battery utilization, and motor operation is established as a reference for expert formulation of fuzzy logic control rules. Preliminary extraction of fuzzy logic rules from the DP-derived optimal energy-saving sequence is accomplished through fuzzy c-means clustering (FCM), after which the rules are further refined by applying an adaptive neuro-fuzzy inference system (ANFIS). Simulation results demonstrate that the proposed strategy achieves over 73 % of DP-level performance and outperforms deterministic rule-based methods by 13.12 %, with reduced motor power fluctuations and improved efficiency distribution.