A novel approach to UWB millimeter high resolution range detection

The Non-linear Gaussian Recursive Algorithm (NGRA) is a novel algorithm that solves some of the known short range radar impairments. These short range radar impairments include clutter, overlapping echo pulses, antenna pulse distortion, and poor distance resolution. The NGRA accurately models the source signal and subtracts it from the original signal which allows for additional peaks to be detected. In order to model the primary signal and its side lobes, a sum of Gaussian model was chosen. To estimate the models coefficients a non-linear fit algorithm is required using initial conditions generated from a peak detector. The coefficients from the model provide location information and distance resolution beyond the limitations of the sampling rate of the captured data. Through experimentation the NGRA algorithm was proven to be accurate and reliable, achieving between one to eight percent error rates.