Noise Reduction and Signal Estimation For 5g Anntenna Using Least Mean Square (LMS) Algorithm and Kalman Filter

The Kalman filter code used in the active control system is described in detail in this thesis. In order to react to variations in the primary noise and 5G environment, traditional active noise management techniques typically use an adaptive filter, such as the filtered reference least mean square (FxLMS) algorithm. However, the weak convergence features of the FxLMS algorithm typically affect how well dynamic noise is reduced. This research suggests utilizing the Kalman filter in the active noise control (ANC) system to enhance the efficacy of noise reduction for dynamic noise. The Kalman filter is used effectively by the ANC application using a new dynamic ANC model. The numerical simulation shows that the proposed Kalman filter works better than the FxLMS approach in terms of convergence performance for handling dynamic noise. This suggests that the transition of the control filter has a higher degree of confidence than the observation function. When the effects of various mu and K values were examined, it was discovered that the LMS algorithms had a slow rate of convergence for mu = 0.01 and high starting error signals for both echo and noise cancellation that subsequently decreased. Although there was some improvement in the noise and echo-canceled signals, the residual noise and echo persisted. Error signals dropped more quickly and the convergence rate was better with mu = 0.05 than with mu = 0.01; this suggests more efficient cancellation. There was less lingering echo and noise in the resulting clearer signals, with echo and noise cancelled. Hybrid LMS algorithms exhibited rapid convergence at μ = 0.1, with error signals declining sharply, indicating effective cancellation. With little lingering interference, the noise-cancelled and echo-cancelled signals were noticeably clearer. Although the convergence was quite quick with mu = 0.5, there was a higher chance of instability, particularly for the LMS method. It is often advised to use a hybrid algorithm with a mu value of between 0.05 and 0.1. Achieving the ideal balance between convergence speed and stability requires proper mu tuning, which guarantees efficient cancellation without creating instability.