MCMC particle filter for tracking in a partially known multipath environment
B. Sachintha Karunaratne, Mark R. Morelande, Bill Moran · 2013
The principal difficultly in tracking in an urban terrain is the presence of multipaths. However by using proper modelling and signal processing techniques these multipaths can be used favourably. In this paper we consider a more robust model of the urban terrain by not assuming exact wall locations but rather allowing for small deviations. This is achieved by introducing a random phase shift to the radar equation. A MCMC based particle filter which uses an adaptive kernel to improve the mobility of the Markov Chain is proposed with supporting simulation results.