AdaBoost Algorithm-Based Channel Estimation: Enhanced Performance

A combination of a group of rules characterized by weakness and imprecision, to reach a prediction rule known for its high precision, forms the concept of boosting, a machine learning approach. From this concept, the Adaptive Boosting (AdaBoost) algorithm spun. It is the first of its kind, and remains in use, under study, and involved in practical applications in various fields to this day. This algorithm is involved in modulation techniques, as it works on improving the bit error rate (BER). Inputting a noisy signal received from a sender into AdaBoost yields the original signal after removing noise. This is done through the reconstruction of the constellation diagram of the modulation technique, by eliminating the noise filling the data’s signal space. The result is an enhancement of the BER of 8 quadrature amplitude modulation (8QAM) and 16 quadrature amplitude modulation (16QAM), through AdaBoost. The AdaBoost algorithm is then added to the channel estimation techniques like the Least Squares (LS), Least Mean Squares (LMS) and Recursive Least Squares (RLS) and is utilized to enhance the BER performance of different estimation techniques in a Rayleigh fading environment. The AdaBoost technique allows benefiting from its learning capabilities as a machine learning algorithm in overcoming the effect of noise and fading channel on the received signals.