Neural network control has been a research hotspot in academic fields due to the strong ability of computation. One of its wildly applied fields is robotics. In recent years, plenty of researchers have devised different types of neural networks, such as zeroing neural network (ZNN), recurrent neural network (RNN) and gradient neural network (GNN), to address complex control issues in robotics in real life setting. Specifically, different kinds of robotics systems, such as redundant manipulators, multi-agent robotics systems, dual-arm robots, and mobile robots, are investigated to achieve different kinds of control features, such as quadratic-programming-based optimal control of redundant manipulators, the remote center of motion control (RCM) of medical robots, and admittance control of robot-environment interaction, etc. However, some knotty issues still exist, such as uncertain structure parameters, position and orientation control, performance index optimization, and obstacle avoidance.
The potential research directions can be divided into two parts: theoretical improvement and application innovation. Firstly, aiming at the current neural network (NN)-based solvers, we expect to see more technologies to improve the performance from several perspectives, such as solving speed, control performance, learning efficiency, and robustness. Different types of optimal algorithms, such as quasi-newton method, gradient descent method, genetic algorithm, etc, can be exploited to tackle the problem. In addition, investigations on neural networks to control robots are the main part of the Research Topic, including zeroing neural network (ZNN), recurrent neural network (RNN), gradient neural network (GNN), dynamic neural network (DNN), etc. In general, innovations in neural networks are of great significance in this topic.
Secondly, control schemes are willing to be proposed to tackle knotty realistic issues, including but not limited to uncertainties of robot structure information, human-robot interactions, and motion-force control of robots. Apart from that, applications on different types of robots are encouraged, including but not limited to: redundant manipulators, dual-arm robots, mobile robots, continuum robots, and soft robots.
Topics of contributing papers include, but are not limited to
? Theory of neural networks for convergence, robustness, and other characteristics
? Innovative design of neural networks (for improving performance, continuous-time/discrete-time systems, software/hardware implementation, etc.)
? Neural network approaches to optimization and adaptive dynamic programming
? Distributed neural networks or distributed optimization for mathematics, control, etc.
? Neural networks related models or learning systems
Neural network control has been a research hotspot in academic fields due to the strong ability of computation. One of its wildly applied fields is robotics. In recent years, plenty of researchers have devised different types of neural networks, such as zeroing neural network (ZNN), recurrent neural network (RNN) and gradient neural network (GNN), to address complex control issues in robotics in real life setting. Specifically, different kinds of robotics systems, such as redundant manipulators, multi-agent robotics systems, dual-arm robots, and mobile robots, are investigated to achieve different kinds of control features, such as quadratic-programming-based optimal control of redundant manipulators, the remote center of motion control (RCM) of medical robots, and admittance control of robot-environment interaction, etc. However, some knotty issues still exist, such as uncertain structure parameters, position and orientation control, performance index optimization, and obstacle avoidance.
The potential research directions can be divided into two parts: theoretical improvement and application innovation. Firstly, aiming at the current neural network (NN)-based solvers, we expect to see more technologies to improve the performance from several perspectives, such as solving speed, control performance, learning efficiency, and robustness. Different types of optimal algorithms, such as quasi-newton method, gradient descent method, genetic algorithm, etc, can be exploited to tackle the problem. In addition, investigations on neural networks to control robots are the main part of the Research Topic, including zeroing neural network (ZNN), recurrent neural network (RNN), gradient neural network (GNN), dynamic neural network (DNN), etc. In general, innovations in neural networks are of great significance in this topic.
Secondly, control schemes are willing to be proposed to tackle knotty realistic issues, including but not limited to uncertainties of robot structure information, human-robot interactions, and motion-force control of robots. Apart from that, applications on different types of robots are encouraged, including but not limited to: redundant manipulators, dual-arm robots, mobile robots, continuum robots, and soft robots.
Topics of contributing papers include, but are not limited to
? Theory of neural networks for convergence, robustness, and other characteristics
? Innovative design of neural networks (for improving performance, continuous-time/discrete-time systems, software/hardware implementation, etc.)
? Neural network approaches to optimization and adaptive dynamic programming
? Distributed neural networks or distributed optimization for mathematics, control, etc.
? Neural networks related models or learning systems