Robust Trajectory Control of Robot Manipulators Using Time Delay Estimation and Internal Model Concept

Endpoint perfect tracking control of robots — A robust non inversion-based approach

“…22), which is not negligible. Thus in short term, although, IMC provides some perfect tracking control scheme theoretically, due to lack of perfect data for modeling, it can not work as the perfect controller in practice [21]. In the proposed control approach, in spite of assumption of no prior knowledge about system's dynamic and parameters for the control law design, we have still good response.…”

“…Differentiating both (17) and (19) with respect to time and using (16) and (20) with this assumption that γ can be modeled by a p th order Ordinary Differential Equation (ODE) [24] shown in (21), we have the following results…”

“…The uncertainty is modeled by a p th order differential equation usually. Then, set the uncertainty equation in (21) to zero to calculate b j . By choosing p =1 for the uncertainty and seting the uncertainty equation to zero, we have considered γ in the form of an arbitrary constant.…”

“…17, it can be seen that the tracking performance of the IMC system is excellent and the tracking error is so small that it can be neglected in reality. However, the key point is that achieving this trajectory tracking ability is affordable subject to knowing the system's accurate model, which is hard to be obtained in a complex nonlinear plant such as the robot manipulator under study [21]. The aforementioned weakness can partly be eliminated by using a neural network [26], but still there are a few problems.…”

“…Third, in many cases, the controllers are realized on the computer systems, therfore time delay in feedback control loop is unavoidable. Thus, IMC can't work as the perfect controller [21]. In this paper, we will present a new approach called Robust Non Inversion-Based Perfect Tracking Control (RNIBPTC) scheme overcoming almost all the above mentioned weakness.…”