A Stochastic Singular Vector Based MIMO Channel Model for MAC Layer Tracking

TL;DR

This paper introduces a stochastic MIMO channel model based on singular vectors and values, enabling realistic eigenmode behavior simulation for testing device tracking capabilities without complex decomposition.

Contribution

It presents a novel eigen-domain modeling technique that directly manipulates singular vectors and values for realistic and stress testing of MIMO channel tracking.

Findings

Allows realistic eigenmode behavior simulation

Enables stress testing of eigenvector tracking

Produces narrow Doppler spectra for efficient updates

Abstract

A novel stochastic technique is presented to directly model singular vectors and singular values of a multiple input multiple output channel. Thus the component smodeled directly in the eigen domain can be adapted to exhibit realistic physical domain behavior when assembled. The model exploites natural paths of eigenmodes, such that a simple Doppler filter generator process can be used. Furthermore it is possible to directly manipulate the singular vector dynamics in a way that an unrealistic "stress channel" can be modeled in the eigen domain. This is particularly useful for testing the eigenmode channel tracking ability internal to a communication device such as a modem, where impairments in tracking will cause interference between eigenmodes. The model can also facilitate mode tracking testing as it directly produces tracked ungtangled eigenmodes, providing the narrowest possible singular vector Doppler spectra and consequently lowest required update rates of each eigenmode. The singular vector based model targets testing of the eigen domain functionality of MIMO modems/devices, an apparatus focus, without the need for including the decomposition stages.