Estimating the Behaviour of a Prismatic-Revolute Robot using a Robust Filtering Strategy

Abstract

Estimating the states of a manipulator is a challenging task as it consists of sinusoidal functions that cannot be represented by a simple, linear model. In such cases, the well-known extended Kalman filter (EKF) may not yield reliable estimates. The calculation of the Jacobian matrix, as part of the EKF, may not be straightforward and introduces errors in the nonlinear approximations. A relatively new estimation strategy called the sliding innovation filter (SIF) offers an alternative solution to the EKF. The SIF forces the estimates to be within a region of the measurements, with some differences due to system modeling. In this paper, the SIF is used to estimate the states of a robotic arm of type prismatic-revolute (PR). The results are compared with the well-known EKF. A faulty scenario is considered when the robot parameters are poorly defined. The results demonstrate the effectiveness and robustness of the SIF, and offers an alternative estimation strategy for estimating different robot types.