NearOptimal Tracking of TimeVarying Digital Radio Channels Using apriori Statistical Channel Information

Synchronization algorithms for digital receivers operating over timevarying dispersive radio channels perform a variety of tasks for online identication of the environment In this paper we focus on nearoptimal rstorder tracking of the timevarying discreteequivalent impulse response of selective fading radio channels when certain apriori statistical channel parameters are known or have been estimated e g via repeated channel sounding experiments The optimal decisionaided channel estimator structure is shown to comprise a bank of innite order Kalman lters one for each individual channel tap For stability and complexity reasons simplied rstorder sublters should be used This is shown to result in a structure identical to the LMS algorithm with individual stepsizes for each tap Due to the mismatch between the inniteorder system channel and the rstorder estimator the tracker thus simplied yields suboptimal performance and the task of nding optimal estimator parameters Kalman gain factors becomes involved In this paper an analytical procedure of obtaining optimal Kalman gain factors and a lower bound on corresponding minimal tap error covariances is established for steadystate nearoptimal channel trackin