High-Speed Wheel Performance Linear Regression Modeling on Loose Terrain Based on Single-Wheel Experiment

Expanding the range of exploration of lunar rovers is a critical challenge, and improving their traversal speed plays a key role in addressing it. To support this, accurate measurement of wheel performance under high-speed conditions is essential, ultimately allowing the development of predictive models that account for velocity and wheel slip. Our research group has developed a high-speed single-wheel testbed capable of evaluating wheel performance at speeds up to $10 \mathrm{~km} / \mathrm{h}$. In this study, we conducted experimental measurements using the testbed by varying three input conditions: vertical load, slip ratio, and speed. The corresponding outputs-drawbar pull and sinkage-were measured using force sensors. We then applied a data-driven regression model to the collected dataset, deriving regression expressions to predict output performance based on the input conditions. The results show that sinkage can be accurately represented as a linear function of load and slip ratio. In contrast, the traction coefficient exhibited comparable influence from all three input factors, but its behavior was not well captured by a linear regression model. These findings demonstrate the potential of our regression approach to predict rover mobility based on control parameters and contribute to both mission planning and the formulation of mechanics-based mobility models.