Xiaogang Ruan

Bionic Self-Learning of Two-Wheeled Robot Based on Skinner's Operant Conditioning

Aiming at the problem about the posture balance control of two-wheeled self-balancing mobile robot, a learning algorithm that it is made up of BP neural network and eligibility traces based on the operant conditioning theory is put forward as a learning mechanism of the two-wheeled robot. The algorithm utilizes the characters of eligibility traces about quicker learning speed, higher reliability, ability in resolving effect about delay, and combines with BP neural network, so that the two-wheeled robot can obtain the movement balance skills of controlling like a human or animal by interacting, studying and training with unknown environmental, and realize the posture balance control of the two-wheeled robot. Finally, some simulation experiments have been done in the undisturbed and disturbed environment respectively. The simulation results show that a learning mechanism of the complex learning algorithm consisted by BP neural network and eligibility traces based on Skinner’s operant conditioning theory can realize the posture balance control to the two-wheeled robot, and reveal the better dynamic character and robustness. And it also has the higher research significance in theory and the application value in project.

Dynamic Model and Analysis of the Flexible Two-Wheeled Mobile Robot

Aiming to the flexible structure feature of waist, the flexible two-wheeled mobile robot was designed based on dynamics of flexible multi-body system. Adopting the assumed mode method, the robot can be abstracted to a three-link robot system, and the flexible robot dynamics model is established using D the Alembert-Lagrange, then the dynamics equation of this systems is obtained using the oblique angles of three poles and the turn corners of two wheels as the parameters, and the vibration equation of connecting rod with flexible waist is also gained. Listing some material steps at establishing the mathematics model, and from this some simulation analysis was carried out. The simulation results confirm that the model is efficient, and it has provided the theory basis for the movement and control research of the flexible robot system.

On-line NNAC for Two-Wheeled Self-Balancing Robot Based on Feedback-Error-Learning

A method based on cerebellar model with feedback-error-learning was proposed to control a two-wheeled self-balancing robot. The forward controller which is an on-line neural network adaptive controller is corresponding with cerebellar and expresses the function of cerebellum in balancing control problem. The robot is a kind of MIMO system, so the cerebellar cortex part has two nerve cell output. From the simulation, we can see that the robot can be balanced in fixed position well by this method, and can improve the feedback control to some extent. It is a kind of bionic control method which imitates the cerebellum part of brain.

A 3-D simulation of unicycle robot based on virtual prototyping technology

Co-simulation of ADAMS and MATLAB for designing and developing mechatronics and control system is presented in this paper. The cosimulation technology provides an avenue for realizing mechatronics system design loop and control system design synchronously without the need for building prototypes. The virtual prototype developed in ADAMS can be used to investigate dynamic behavior in a 3-dimensional environment. Though ADAMS has the capability of implementing a closed loop control of the virtual prototype, its capability is quite restricted. On the other hand, MATLAB is well known for designing control systems. Co-simulation allows us to get the benefit of both. The co-simulation platform was used to verify different features of a unicycle robot. Mechanical drawings of the robot are first created using CAD software, e.g., SOLIDWORKS and imported into the ADAMS. Dynamics of the robot in autonomous mode are simulated using the co-simulation platform in which the controller is designed and implemented using MATLAB. Simulation results can then be used to modify mechanical design and improve the control method. Although the virtual platform is tested for one specific system, it can be easily used for design of other mechatronics systems and control methods e.g., model free control, learning control and developmental robotics.

Self-Balance Control of Two-Wheeled Robot Based on Skinner's Operant Conditioning

Aiming at the movement balance problem of the two- wheeled robot, the operant conditioning theory of artificial cerebellar sensorimotor systems is used, and the theory adopts a learning mechanisms of Skinners operation conditioned based on the learning algorithm of recurrent neural network, through learning and training, the two-wheeled robot can obtain the skills of movement balance control like a robot or animal in the process of gradually forming, developing and improving by self- organization. The simulation results show that the Skinners operation conditioning has such a virtue of stronger self-learning ability and higher robustness.

Dynamic Modeling and Simulation of a Flexible Two-Wheeled Balancing Robot

This paper focuses on deriving the dynamic model of a novel flexible mobile robot, which has elastic joint, so that the robot is able to be simulated with a computer and analyzed dynamically. First, the dynamic equations are derived from Routh formulation with the nonholonomic constraints of the robot, and then the nonholonomic constraint forces are removed so that the dynamic equations are transformed to an input-affine form at last. Three illustrative simulation results are given which show that the dynamic model presented in this paper is reasonable.

Research of Dynamic Model and Control Ling of Flexible Two-Wheel Upright Self-balance Humanoid Robot

The dynamic modeling problems of flexible two-wheel upright self-balance humanoid robot are researched. A dynamic modeling is obtained by the Lagrange equation and dynamics mechanics theory. Using springs imitate human lumbar spine, considering the lumbar spine of the flexible robot, which has an actual length without being treated as an approximate point. All these are different from previous robots. The quality of upper and lower base is cancelled. The dynamic model is linearized and its spatial equations of state are established. The simplified dynamic model is obtained by designing its structure in simple ways, in this way. It is convenient to control the robot. The state feedback controller(LQR) with good robustness is designed on Matlab. The stability of systems is proved by the experimental results. Validity and rationality of the system modeling and the controller designing are verified through the performance experiments of the prototype. The research also supplies theoretical instructions for developing the dynamic control system in the flexible two-wheel upright self-balance humanoid robot. It has great significance in design and research for the humanoid robots.