Research On LED Nonlinear Post-Equalization Technology For Visible Light Communication

Visible Light Communication(VLC)is a wireless communication technology that employs the visible light spectrum emitted by Light Emitting Diodes(LED)for data transmission.VLC has several advantages,including environmental safety,reliability,and rich spectrum resources.In comparison to traditional Radio Frequency(RF)communication technology,VLC has significant benefits in mitigating the shortage of spectrum resources.However,similar to RF communication,VLC systems also face the issue of nonlinear distortion.Nonlinear distortion in the VLC system can originate from physical devices,such as LEDs,Digital-to-Analog(D/A)converters,and Photodetectors(PD),among which LED are the primary source of nonlinear distortion.Simultaneously,due to the Peak-to-Average Power Ratio(PAPR)of the Optical Orthogonal Frequency Division Multiplexing(O-OFDM)signal,the O-OFDM system is highly sensitive to the nonlinear distortion of the LED,which can severely deteriorate system performance.Therefore,it is imperative to mitigate LED nonlinear distortion to enhance system performance.Consequently,this paper undertakes an in-depth investigation into post-equalization technology to alleviate LED nonlinear distortion and introduces a single-stage adaptive postequalizer O-OFDM system.Subsequently,addressing the issue of error accumulation and the LED nonlinear memory effect not being considered when mitigating the nonlinear distortion of the LED in the single-stage post-equalizer,a Volterra hybrid post-equalizer O-OFDM system based on the memory polynomial is presented.Lastly,the neural network is employed in the post-equalizer,and a hybrid post-equalizer O-OFDM system based on radial basis function is suggested.The specific research contents are as follows:1.To address the issue of LED nonlinear distortion in the VLC system,a single-stage adaptive linear post-equalizer and adaptive non-linear post-equalizer are developed,based on the least mean square(LMS)algorithm,recursive least square(RLS)algorithm,linear equalizer,and nonlinear equalizer.In comparison with pre-equalization schemes,this approach reduces system costs by eliminating the need for high-precision estimation of nonlinear models and additional feedback physical circuits at the transmitting end.The improvement effect of the aforementioned equalizers on LED nonlinearity is evaluated using bit error rate as the performance index,and the influence of the number of equalizer taps and algorithm parameters on the equalization effect is simulated and analyzed.Results indicate that the three post-equalization strategies can mitigate LED nonlinearity,with the nonlinear equalizer showing the most significant effect.These findings form the foundation for the subsequent proposal of the hybrid post-equalization scheme.2.Propose a Volterra hybrid post-equalization O-OFDM system,which is based on memory polynomials.This solution effectively overcomes the high complexity of the traditional Volterra series and the limitations of single-stage equalizers in improving system performance,such as error accumulation and the non-linear memory effect of LEDs.Firstly,the traditional Volterra series is replaced with the Volterra series based on the memory polynomial.This reduces the complexity of the traditional Volterra series while compensating for the nonlinear memory effect of the LED.Secondly,the Volterra series based on the memory polynomial is applied to the equalizer,and a hybrid equalizer cascaded by two nonlinear equalizers is proposed to mitigate the nonlinear distortion of the LED.This equalizes the nonlinear signal twice and significantly improves system performance.Finally,the hybrid equalizer is quantitatively analyzed through Monte Carlo simulation experiments,and the optimal number of taps of the hybrid equalizer is determined.The single-stage equalizer and the linear cascaded nonlinear hybrid equalizer are combined to mitigate the LED nonlinearity distortion for comparison.Simulation results indicate that this scheme can effectively compensate for the nonlinear distortion and memory effect of the LED,and significantly improve system performance.3.The proposed approach involves the use of a neural network algorithm to mitigate LED nonlinear distortion in a post-equalizer,and a hybrid post-equalizer O-OFDM system based on radial basis function(RBF).The scheme employs a hybrid equalizer consisting of an LMS equalizer and a radial basis function-based neural network cascade,which overcomes the limitations of inaccurate center point selection and unclear data linearity in threedimensional space associated with radial basis function neural networks.The hybrid equalizer achieves superior mitigation of the LED nonlinear distortion.Monte Carlo simulation experiments are conducted to analyze the scheme in ACO-OFDM and DCO-OFDM systems,and compare it with a single-stage post-equalizer based on RBF neural network.The simulation results indicate that both the single-stage radial basis function-based post-equalizer and the proposed scheme can mitigate LED nonlinear distortion,but the proposed scheme demonstrates greater robustness in compensating for LED nonlinear distortion.