The MIMO and ISI Detection Problems: A Common Model and a Quasi Optimal, Quadratic Complexity Solution

A novel quasi-MLD solution denoted directional lattice descent (DLD in short) is proposed in this paper to the MIMO and ISI detection problems, which exhibits quadratic time complexity. We formulate the detection problem as that of finding the closest point in a lattice. This is solved by a proposed discrete space analogy to the continuous-space gradient descent method: an initial lattice point is selected and subsequent lattice points are iteratively calculated so that each resides at a smaller distance from the received point than its predecessor, till a termination point. The problem and its solution are formulated within a broader general framework. Implementation aspects and comparative simulation results are presented.