Optimal Training Signals and Detection for OFDM Under Colored Noise

Based on convex programming, this paper presents the optimal training signal design for orthogonal frequency-division multiplexing (OFDM) under colored noise. An effective method for OFDM symbol detection with the use of preceding is also described. Both analytical and simulation results show that the proposed design and method can effectively estimate the channel and reliably detect the OFDM symbols. By working in the frequency domain rather than the time domain, the computation complexity of the proposed estimation algorithm is significantly reduced.