Improving the performance of LMS algorithm using neural networks in WiMAX systems

Abstract

WiMAX is a broadband wireless technology that provides high-speed data over a wide area. The requirements for high-quality connections and high demand for high throughput in this wireless network and other networks have stimulated new improvements in wireless communications such as Smart Antenna Systems. Smart (adaptive) antennas enable spatial reuse, increase throughput, and communication range due to increased antenna array directivity. In this paper, the LMS algorithm that is widely used within smart antenna systems is discussed, then the neural network technology is used to give initial weights values ​​that enable us to obtain the lowest possible error, which increases productivity. After implementing the proposed algorithm, a small mean square error equal to less than 20% of the maximum mean square error was obtained as the initial response and needed less than 20 iterations to reach the stability state. In comparison with other research and improved algorithms in the same field, the proposed algorithm shows satisfactory performance with an initial error less than 35% of other algorithms’ error.