Ambiguity Subspace Resolution for Digital Twin Parameter Estimation

In this work, we analyze the limits of parameter estimation of multipath components like their amplitude, delay and/or phase. Especially challenging is the estimation of such parameters when the multipath components have a low relative delay, i.e., less than half a tap delay with respect to the discrete channel impulse response. We will introduce the concept of the ambiguity subspace, which indicates that even for high signal to noise ratios and quite large RF bandwidths the Fischer information is limited and accurate estimation of the real world parameters becomes challenging. To overcome these theoretical limits, we propose the idea of ambiguity subspace resolution, ideally providing a solution close to the correct parameter values. There are a number of applications, which may benefit from the knowledge of the real world parameters like positioning, or, channel prediction. Here, we have in mind the building vector data map describing the environment of a digital twin of a mobile network, which is inferred from regular UE feedback of radio channel estimates. Such a digital twin is expected to support 6G use cases from optimized beam management, handover, load balancing up to channel estimation and/or channel prediction.