This simple simulation shows the implementation of FbLMS algorithm for a single channel feedback active noise control system. Here, the controller generates an "anti-noise" signal to result a destructive interference at the sensor position. The objective is to minimize the noise residue.
FbLMS does not employ any reference sensor. Instead, it uses linear predictor for generating reference signal. Thus, FbLMS is suitable for reducing narrow-band noise.
Similar to feed-forward system, the FbLMS algorithm comprises of two tasks: the "off-line" identification of the secondary propagation path that stays between the actuator and sensor; and the "on-line" control, in which the parameters of the controller are adjusted.
As usual, I use only simple programming technique and basic commands. Function "filter()" can be substitute with convolution routine, i.e., y=h(k)*s(k). A brief explanation on the system and line by line comments are provided. |
