Information representation and fusion based on uncertain basis in complex environment

For the multiple source information fusion in complex environment problem, An information representation model based on the uncertain basis and a robust fusion algorithm are presented. First, a definition of uncertain basis is given and a realization method with kernel function is proposed. Based on the uncertain basis, an uncertain information description model is presented. Second, an adaptive kernel bandwidth robust fusion technique is proposed. The kernel bandwidth is initially decided by sample covariance and then is adjusted after each iterative to describe the estimated uncertainty. Theory analysis and simulation show that, the proposed method equals to the existing techniques in good scenario and takes advantage in complex environment.