An RLS Tracking and Iterative Detection Engine for Mobile MIMO-OFDM Systems

This paper presents a tracking and detection engine for 2 x 2 spatial-multiplexing multiple-input multiple-output orthogonal frequency division multiplexing (MIMO-OFDM) systems under highly time-selective fading channels. A recursive least-squares (RLS) adaptive algorithm is adopted to track time-varying channel frequency response. Besides, iterative MIMO signal detection is used, including inter-carrier interference (ICI) cancellation, to enhance system performance. To avoid error propagation in the decision-directed RLS adaptation, ordered-selection mechanism and signal strength criteria are incorporated so that more reliable detection results can be selected for calculating channel statistics. The scheduling strategies and hardware reduction techniques are proposed for implementation of the computation-intensive RLS adaptive algorithm and ICI cancellation. This baseband processor has been fabricated in 90 nm CMOS technology and has 1023K gates. At the nominal operating frequency of 15.36 MHz, it provides throughput of 76.8 Mbps with 64-QAM up to 240 Km/hr. For constellation of 16-QAM and QPSK, the mobility of 360 Km/hr is supported. The power consumption is 11.8 mW at 0.7 V supply voltage. The operating frequency can be raised up to 103 MHz to offer 515 Mbps throughput with power consumption of 202 mW at 1.2V.