Soil parameter modification used for boosting predictive fidelity of planetary rover's slippage

In a planetary exploration mission, a rover's slippage badly limits its control performance. This paper presents an online method to modify the predictive errors for the slippage caused by the variability of the soil parameters and limitation of the predictive models. To modify the slippage appropriately, all of the soil parameters related to the predictive models are analyzed to estimate their sensitivities to slippage, and tests indicate that phi and K have the greatest influence on the slippage. Based on the differential principle, phi and K are decoupled and solved using the predictive model errors. To eliminate the errors in the modification process caused by the models' nonlinearity, an effective coefficient is proposed, and its definition is given based on Taylor's formula. To implement a more stable modification, some necessary constraints are presented. The validity of this method is verified using simulations and experiments. The results show that approximately three steps are needed to reach the convergence points; it has a good response with a fluctuation in the input slippage; a reasonable initial value will be beneficial for the modification. In summary, this method is effective at eliminating predictive slippage errors; the proposed effective coefficient and method are helpful at boosting the modification's convergence speed and maintaining its stability. (C) 2014 ISTVS. Published by Elsevier Ltd. All rights reserved.