| ID | 369090 |
| Call Number | 621.38224 S531N |
| Title Proper | Novel adaptive neuro fuzzy inference system ANFIS for active noise control |
| Language | ENG |
| Author | Sharma, Manoj Kumar ; Vig,Renu,Supervisor |
| Publication | Chandigarh, University institute of engineering & Technology, 2020. |
| Description | xvii:127p.+ 28cm. |
| Summary / Abstract (Note) | To control noise, active noise control (ANC) system is being implemented in many applications such as headphones, MRI, cars, yachts, etc. In the present work, the ANC system is implemented to decrease the ambulance siren noise for the comfort of the patient lying in the ambulance. The ANC systems based on conventional algorithms i.e. LMS and FxLMS are implemented for ambulance siren noise reduction and their performance comparison is done. The experiments are performed in the laboratory to obtain the secondary path coefficients for different filter lengths thereby achieving the optimum performance of the system. Further, it is seen that in most situations the placing of an error microphone at the desired spot i.e. around the patient s ear is not feasible. Keeping this in view, the virtual sensing technique is incorporated in the ANC system. In the present research, an FxLMS virtual sensing technique (F-VST) based algorithm is proposed to overcome this problem. It is observed that better attenuation is achieved at the desired (virtual) location with the proposed algorithm. The conventional algorithm, being the linear algorithm, tends to degraded performance in the non-linear scenario. To overcome this limitation, the soft computing-based algorithms such as ANFIS, which are non-linear, are applied for ANC. However, the existing ANFIS model generally applies a backpropagation algorithm that has limitations of slow convergence and a tendency to get stuck in local minima resulting in compromised performance. In this thesis, a novel adaptive neuro-fuzzy inference system (N-ANFIS) based on an adaptive learning rate is proposed for an ANC system that overcomes the limitations of backpropagation. The adaptive learning rate based on the Lyapunov stability theory guarantees the convergence of the backpropagation algorithm. Thus, N-ANFIS ensures the faster adaptation of the system towards target noise thereby resulting in a quick response for efficient noise cancellation. newline |
| Classification Number | 621.38224 |
| Key Words | Active Noise Control ; Adaptive Neuro-Fuzzy Inference System ; Electrical and Electronics Engineering ; Neuro-Fuzzy Inference System |