Accelerating the convergence of adaptive filters for active noise control using particle swarm optimization

The excessive exposure to certain kinds of noise can lead to health problems. In order to avoid such a situation, the use of noise attenuation devices is a common solution. Among those devices, the Active Noise Control (ANC) systems have gained prominence over the years due to the technological development and costs reduction. Despite of the ANC good performance in terms of low frequency sounds attenuation, the convergence speed for this kind of system is an important issue when dealing with real time applications in dynamic environments. This paper presents an alternative solution to accelerate the system response based on the use of coefficients obtained via a training process with Particle Swarm Optimization (PSO). Those coefficients are used during the filter initialization and provide a response time reduction of up to 97.06%, followed by an attenuation improvement of about 12.23% respect to FxLMS-based adaptive filters initialized with null coefficients.