UCLA

Multipath is the dominant source of error in indoor positioning systems, because it can significantly distort the shape of the correlation function used for time-delay estimation. With a narrowband signal, the range resolution is often insufficient to decompose the overlapped received signals. However, if the signal has constraints, we can theoretically estimate the channel response more accurately. We assume the constraint that there is a known amount of nonzero values in the channel response, each of which represents a delayed version of the training signal in the receiver. We focused our work on the principles of the least mean squares equalizer, which tries to estimate the channel response, because it is fast and can be implemented in parallel and in real-time. Our modified algorithm uses the general principle of the least mean squares equalizer and sets the constraints. The algorithm is designed to be implemented in parallel and in real-time.