Back Propagation Neural Network Control Research Articles

SEMG Based ANN for Shoulder Angle Prediction

Many kinds of upper limb rehabilitation systems have been developing for physically weak and/or injured patients to assist their daily life activities and promote their quality of life. Among those systems, EMG controlled rehabilitation systems provide the most effective and fastest ways to restore the lost functions due to such weakness and injuries. This paper presents the prediction of shoulder angle based on acquisition of surface electromyogram (sEMG) signals. Backpropagation neural network (BPNN) controller is developed to predict the angle of shoulder flexion/extension and abduction/adduction movements. Virtual Reality (VR) human model is developed to simulate the predicted shoulder angle which results from BPNN controller. Four sEMG signals are collected from user arm muscles and processed to extract their feature with root mean square (RMS) method. Then, the signals features directed to the neural network as input and the network predicts the angle of the shoulder joint as an output. The angles from BPNN drive the shoulder joint of the VR human model in virtual environment. Experiments were carried out to evaluate the effectiveness of the developed system and it was found that the constructed BPNN model and VR model can well represent the relationship between sEMG and shoulder joint angles and rotation. These positive results elaborate a move to design and develop the EMG controlled upper limb rehabilitation robot system to rehabilitate the physically weak person and paralysed patients.

A FUZZY NEURAL NETWORK FOR THE ACTIVE VIBRATION CONTROL OF A CENTRIFUGAL PENDULUM VIBRATION ABSORBER

In this study, we develop a fuzzy back-propagation (BP) neural network controller for active vibration control of a centrifugal pendulum vibration absorber (CPVA). The fuzzy BP neural network controller systems can be viewed as a conventional fuzzy algorithm for coarse tuning. The BP algorithm can also be applied for fine tuning, in this case to regulate the anti-resonance frequency in an active pendulum vibration absorber (APVA), by suppressing vibration of the carrier. The dynamic model of the APVA was developed and simulated using MATLAB. In the simulation results, when the frequency of the disturbance changes, the outputs of the fuzzy BP neural network controller are used to determine an appropriate value for the torque of the active pendulum such that the vibration amplitude of the carrier is minimized. A comparison of the carrier vibration results for the CPVA, the fuzzy algorithm and the fuzzy BP algorithm is performed. The simulation results demonstrate the effectiveness of the proposed fuzzy BP neural network APVA for reducing the carrier vibrations.

A neural network controller for congestion control in ATM multiplexers

This paper presents and adaptive approach to the problem of congestion control arising at the User-to Network Interface (UNI) of an ATM multiplexer. We view the ATM multiplexer as a non-linear stochastic system whose dynamics are ill-defined. Real-time measurements of the arrival rate process and the queueing process, are used to identify, and minimize congestion episodes. The performance of the system is evaluated using a performance-index function which is a quantative measure of “how well” the system is performing. A three-layers backpropagation neural network controller generates a signal that attempts to minimize congestion without degrading the quality of the traffic. During periods of buffer over-load the control signal, adaptively, modulates the arrival process such that its peak-rate is throttled-down. As soon as congestion is terminated, the control signal is adjusted such that the coding rates are restored back to their original values. Adaptability is achieved by continuously adjusting the weights of the neural network controller such that the performance of the system, measured by its performance index function, is maximized over a certain optimization period. The performance index function is defined in terms of two main objectives: (1) to minimize the cell loss rate (CLR), i.e., minimize congestion episodes, and (2) to maintain the quality of the video/audio traffic by maintaining its original source coding rate. The neural network learning process can be viewed as a specialized form of reinforcement learning in the sense that the control signal is reinforced if it tends to maximize the performance index function. Performance evaluation results prove that this approach is effective in controlling congestion while maintaining the quality of the traffic.