An OTFS Channel Estimation Scheme Based on Iterative Path Peak Search and Inter-Path Interference Mitigation

This paper proposes a channel estimation scheme based on iterative path peak search (IPPS) and inter-path interference mitigation for fractional Delay-Doppler (DD) channel in Orthogonal Time-Frequency Space (OTFS). The IPPS algorithm transforms the channel estimation into a path peak-seeking problem in the DD domain. The correlation of path response is used to estimate the initial position of the path peak, and the limited-memory Broyden-Fletcher-Goldfarb-Shanno (L-BFGS) method is employed to iteratively reduce the estimation error in the previous stage, bringing the estimated peak position closer to the actual peak position. To further improve the estimation accuracy, based on the IPPS algorithm, all identified paths, or the corresponding pilot responses, can be optimized to minimize the inter-path interference. Simulation results demonstrate that, compared with the related benchmarks, the proposed IPPS algorithm has quite low complexity while ensuring Normalized Mean Square Error (NMSE) performance, and the improved IPPS algorithm can significantly improve NMSE performance further at the cost of tolerable complexity.