Estimation of terramechanics parameters of wheel-soil interaction model using particle filtering

Accurate estimation of the parameters affecting the wheel-soil interaction terramechanics of an extraterrestrial rover is key to the success of its mission. Traditional approaches to estimating the relevant parameters based on laboratory tests lead to predictions that show significant deviation from experimental observations. The objective of this article is to apply dynamic Bayesian estimation techniques on the measurements from simple single wheel tests to estimate the terramechanics parameters. This ensures that the parameter estimation takes into account the scatter that invariably exists in physical measurements. A mathematical model for a rigid wheel driven on a dry (0% moisture content) granular soil medium is considered to model the planetary regolith. It is demonstrated that adopting Bayesian techniques for terramechanics parameter estimation leads to good predictions for the drawbar pull, torque and the wheel sinkage. This bypasses the need for using more complex models which in turn require additional parameters to be estimated.