Optimal Low-Complexity Orthogonal Block Based Detection of OTFS for Low-Dispersion Channels

Orthogonal time frequency space (OTFS) modulation constitutes a promising technology for high-mobility scenarios. However, the detection of OTFS systems imposes substantial complexity. Hence, we propose novel orthogonal block (OB) based scheme significantly reducing complexity without any performance loss with integer Doppler shifts. This is achieved by recognizing that received signal can be partitioned into multiple parallel blocks. Therefore, data symbols within an only depends on signals this reduced dimension. Explicitly, graph theory identification algorithm, which models relationship between and original information as bipartite graph, where depth first search (DFS) algorithm invoked partitioning For each block, existing algorithms used. Since size blocks may much lower than signals, reduced. example, OB MMSE detector approximately factor 4096 traditional channel having two paths.