Roman Bychkov - Data-aided beamspace processing in Massive MIMO receiver

At the base station receiver, the processing of users’ signals in beamspace can significantly reduce the computational complexity by reducing the dimensionality of the processed signals. At the same time, there is no significant loss in detection accuracy due to the sparsity of signal in beamspace. Unfortunately, the choice of beams may not be optimal due to limited computational resources, the influence of noise, channel aging, and the misalignment between the selected directions of signal arrival and the actual signal subspace. It is proposed to solve these problems by accounting for the a prior distribution of users’ signals in space, and the beamspace choice based on the trainable structures with low computational complexity. Bio: Roman Bychkov received the . degree in applied mathematics and physics from the Moscow Institute of Physics and Technology, Moscow, in 2020. He is currently pursuing the Ph.D. degree with the Artificial Intelligence Technology Center, Skolkovo Institute of Science and