Interference Mitigation for Time of Arrival Estimation

In indoor environments, geolocation using Time of Arrival (TOA) estimation must use a wide bandwidth to overcome the effects of multipath propagation. Narrowband interference from other applications is a significant problem. In this paper we investigate a number of signal processing schemes to detect and overcome one or more relatively narrowband interferers. By making use of the limited time duration of the channel impulse response we are able to recover the TOA with high accuracy in most cases, even when the channel frequency response has large amount of missing data due to interference. Our proposed approach treats the corrupted channel frequency response as a non-uniform sampling of the true frequency response and draws on ideas from compressive sensing to recover an estimate of the missing frequency response samples. We also consider the computational complexity of a variety of potential TOA recovery methods and examine their effectiveness as a function of the interferer bandwidth. We demonstrate our method using real data measured in non line-of-sight conditions in a typical office building.