Learning and Adaptive Control Systems on Robotics

The need for effective control methods to deal with poorly-modeled dynamical systems in the presence of parametric uncertainties is an important area in the research field of robotics control. Adaptive control for such systems is very attractive because it adjusts controller parameters in order to provide adaptation mechanisms. In addition, the presence of nonsmooth nonlinearities such as dead-zone, hysteresis, backlash and saturation in robotic systems can cause major system performance deterioration. Adaptive methods are able to accommodate these undesirable nonlinearities.
On the other hand, artificial neural networks have been studied for years and their properties of learning and adaptation have made them of extreme use in robot control systems. Recently, there are works that have succeeded in designing neurocontrollers in closed-loop applications and at the same time provide guaranteed stability and performance. It is well known that the linearity in the parameters and availability of a known regression matrix are required. Although neurocontrollers are adaptive learning systems, they do not need this assumption.

The aim of this Research Topic is to invite original latest research achievements in learning and adaptive control for robotic systems in order to showcase significant research results in this area. Both theoretical and practical works focusing on this theme would be of great interest.

Potential topics include, but are not limited to the following:
- Neural network based adaptive control of robotic;
- Learning based adaptive robot control;
- Learning based adaptive control for nonlinear with nonsmooth nonlinearities for robotic applications;
- Deep learning based adaptive for robot control;
- Iterative learning based adaptive robot control;
- Fuzzy logic based adaptive control;
- Robust adaptive learning control;
- Fractional-order neural networks controllers;
- Other industrial applications of learning based adaptive control.