Research Of SNR Estimation Algorithm In Adaptive Frequency Hopping Communication

Real-time channel quality assessment is a key technology in adaptive frequencyhopping communication. Signal-to-noise ratio (SNR) estimation is a key part of thereal-time channel quality assessment, which performance of estimating will affect theoutcome of the channel quality assessment directly. This paper will study andanalyze the SNR estimation algorithm as the core issue.There are already many algorithms about SNR estimation, this paper studiesthree improved SNR estimation algorithms which based on the classic SNRestimation algorithms, then analyzes the performance of them through simulationfrom three aspects, such as the length of the received sequence, the estimation rangeof SNR and simulation repetitions. By comparing them from the accuracy, stabilityand computational, the performance of the SNR estimation algorithm based on theauto-correlation matrix eigenvalue decomposition is relatively best in these threealgorithms is found. But if the length of the received sequence is short, especially inthe conditions of low SNR, the performance of it will be poor.To solve the above shortcomings, this paper will study and improve it fromconstructing the autocorrelation matrix and determining the dimension of signal. Interms of constructing the autocorrelation matrix, an equivalent discrete-timemulti-channel model will be established to reduce the correlation between the data.In terms of determining the dimension of signal, the Otsu method will be used toreplace the minimum description length (MDL) criterion. Comparing theperformance between the improved algorithm and the original one throughsimulation, the result shows that the estimation of the improved one is more accurateand better stability at low SNR, especially the length of received sequence is short.The estimation performance is indeed better than the original one. It has been greatlyimproved in the actual complex battlefield environment, especially when the SNR isvery low or the number of samples is very small. In a certain extent, it will improve the quality of communications in the battlefield environment, then will protect theflow of communication. Then we study the Channel quality assessment algorithm,and compared the value of Kalman prediction between that based on the improvedSNR estimation and that based on the original one through simulation, the resultshows that the former has a better stability. In the last, this paper gives a completeflow of channel quality assessment algorithm.