Comparison of Adaptive Equalization Methods for Improving Indoor Optical Wireless Communications Employing Few-Mode Based Uniform Beam Shaping

Abstract

Few-mode based uniform beam shaping exploiting the use of different spatial modes has been recently shown to improve the system resistance against pointing errors and link outage in indoor optical wireless communications. However, the inter-symbol interference arising from transmission of different spatial modes following different paths and mode crosstalk during the transmission becomes another important challenge. In this paper, we theoretically investigate and experimentally compare the performance of two popular adaptive equalization methods, namely least mean square and recursive least-square in both multi-equalizer and single-equalizer configurations, to improve the performance of indoor optical wireless communication system employing few-mode based uniform beam shaping. For both equalization methods, the multi-equalizer scheme achieves better equalization performance and higher convergence speed than the single-equalizer scheme for a similar cost of required computational complexity for each iteration. Moreover, recursive least-square method has a higher convergence speed than least mean square equalization while least mean square method has a lower computational complexity than recursive least-square method under the same desired equalization performance. More importantly, for a fixed reference BER threshold (KP4 FEC limit), the least mean square based single-equalizer scheme with a step size of 0.0035 and a tap number of 61 has the lowest computational complexity among the four types of equalization schemes. This kind of equalizer with relatively low computational complexity can provide guidelines for selecting an applicable adaptive equalization scheme in optical wireless communications with few-mode based beam shaping based on the comprehensive requirements of BER performance and computational complexity.