Adaptive signal processing
The appeal of parallel distributed processing
Parallel distributed processing: explorations in the microstructure of cognition, vol. 1
Adaptive Volterra filters for active control of nonlinear noiseprocesses
IEEE Transactions on Signal Processing
The particle swarm - explosion, stability, and convergence in amultidimensional complex space
IEEE Transactions on Evolutionary Computation
Expert Systems with Applications: An International Journal
Hi-index | 0.00 |
In this paper, we address the Active Noise Cancellation (ANC) as a non-linear control problem. The controller of system is designed, using a multi layer perceptron neural network. The neural weights are adapted based on minimization of the measured noise at silence region. We propose a new method based on Particle Swarm Optimization (PSO) to determine the network weights in an adaptive manner. The modification of PSO algorithm was conducted to the noise cancellation system, in order to handle sudden change of the input noise characteristics. In contrast to the conventional gradient descent type algorithms, the proposed method does not require the estimation of the secondary path parameters. This not only reduces the computational complexity of system, it also improves the stability of ANC system, especially where the secondary path requires a non-linear model. Another advantage of the proposed system is that the adaptation algorithm needs no change when the structure of controller is modified.