Criterion of adaptively scaled belief for PDA in overloaded MIMO channels

Abstract

This paper proposes a new design criterion of adaptively scaled belief (ASB) in probabilistic data association (PDA), especially for overloaded multi-user multi-input multioutput (MU-MIMO) detection. The PDA is a kind of iterative detection algorithms. The most vital issue for improving the convergence property of PDA is how to deal with the soft symbol outliers in each iteration process, which are caused by modeling errors of the prior belief due to a fact that the law of large number does not work well when the large system limit is not satisfied. As the simplest way to mitigate the harmful impacts of outliers, adaptive belief scaling is proposed while stabilizing dynamics of random MIMO channels. Finally, we verify the validity of ASB regarding mitigation of the bit error rate (BER) floor level.