Frequency-domain periodic active noise control and equalization systems
Ji, Li, (Graduate student of engineering)
MetadataShow full item record
The time-domain waveform synthesis method for periodic active noise control (ANC) can be implemented using the impulse train as the reference signal. The time- domain periodic ANC was first implemented and analyzed in this research. This time- domain periodic ANC system was then realized in the frequency-domain. The system transfer functions in steady state for both cases were derived theoretically and verified by the computer simulations. The adaptive filter used in the periodic ANC system has N weights, where N is equal to the normalized period of the primary signal. The adaptive filter functions as a comb filter; thus, the ANC system forms notches at the harmonic frequencies to attenuate all harmonic-related components in the primary noise. The design of the frequency-domain periodic ANC system pursues maximal attenuation of the primary noise. Since each complex weight of the adaptive filter represents the individual harmonic component, controlling the residual noise spectrum can be accomplished by adjusting the adaptive filter coefficient at each frequency bin. The frequency-domain periodic active noise equalization (ANE) system was designed by splitting the adaptive filter output into two branches and the corresponding gain factors were inserted to satisfy the requirement for algorithm convergence and adjustability of residual noise. According to the value of the gain parameter, the ANE system functions at cancellation, attenuation, neutral, and enhancement modes. If N is equal to a power of 2, //-point FFT can be used to reduce computational requirements for both periodic ANC and ANE systems. Comparing to the time-domain periodic ANC using the impulse train as the reference signal, complexity of the frequency- domain periodic ANC is significantly lower. Real-time experiments for testing both time- domain and frequency-domain periodic ANC and the frequency-domain ANE systems were conducted on the TMS320C30-based system.