Investigation of challenges towards achieving sub-millimeter accuracy for UWB localization

We are studying the reasons preventing us from achieving sub-millimeter accuracy for UWB indoor localization which suffers from intense multipath interference. Using just the I-Channel offers better accuracy because of the I-Q alignment problem, but it requires great number of samples averaging, and causes unstable system operation. Non-coherent detection is an interesting alternative, since it eliminates the I-Q-channel problem. Also, our project operates indoor where attenuation is not a big problem, and the gain inherited in I-Channel detection is no more attractive to us. Moreover, a stable system is required to allow for pining out other sources of error that get lost due to averaging.