Sliding Mode Based Virtual Sensor for Feedback Linearization in Delta Robots with Unknown Dynamic Model

Abstract

Delta robots are of nonlinear dynamics, therefore applying linear controllers to satisfy the required trajectory tracking is a challenge which becomes more significant when the robot model is unknown. This paper proposes a stable virtual sensor to be used for feedback linearization for delta robots with unknown dynamic model. The sensor design is based on the second order sliding mode observer. This sensor considers that only the inertia of the robot upper links with the motor inertia are known. As a measurement, only the active joints position angles are measured. The stability is proven using Lyapunov theory and explicit relations for the design parameters are obtained. To overcome the effect of the transient response of this sensor, an adaptive tuning gain with maximum value of unity is introduced in the feedback to improve the compensation performance. The results validate this sensor and prove its dramatic performance improvement in trajectory tracking when combined with the PD controller.