Enhancing DME/N multipath rejection with tightened pulse waveform variation

This paper presents a methodology to suppress Distance Measuring Equipment (DME or DME/N) multipath using supervised learning and tightened DME pulse waveform variation. The current DME specifications allow relatively large variation in the DME pulse waveform in both of the interrogator and transponder. The allowed variation in the waveform makes it difficult to discern if the received signal is direct or deformed from added multipath. Therefore, the range of the allowed variation in the pulse shape needs to be tightened and that would facilitate multipath rejection using various algorithms. This paper applies a supervised learning technique to develop a time of arrival error estimator that compensates for the multipath impacts on the DME range measurements. The learning process utilizes the training data consisting of the numerous deformed DME pulse waveforms and the corresponding time of arrival errors. Considerable noise is also included in the training data such that the estimator behaves robustly in various situations. The theory behind and performance of supervised learning on the DME multipath problem are presented in this paper.