Pedestrian and two-wheeler detection using ultra-wideband vehicular radar

In this paper, pedestrian and two-wheeler detection is experimentally studied using an ultra-wideband (UWB) vehicular radar. The received echo of UWB vehicular radar generally suffers from heavy clutter, which makes it difficult to detect small target such as pedestrian and two-wheeler. Also it is not easy to utilize the Doppler since UWB is based on time-domain system. However, when small target is spatially discriminated from heavy clutter, the Doppler should be estimated using each echo's trajectory in time-range coordinate. As a result, pedestrian and two-wheeler are expected to be discriminated from clutter and detected using the trajectory estimation with Hough transformation. The field measurement was conducted at 24GHz and the feasibility is discussed.