A nonparametric detector for signal detection and time delay estimation
Abstract
A nonparametric twoinput Kendall tau detector which detects and measures the dependency between two noisecorrupted signals received at two sensors separated by a known distance is discussed. The relation between the sampled values is derived and the correlation between the two received signals is obtained. The results are used to derive the detector statistics and analyze the performance. The crosscorrelation between the inputs is maximum when tau equals zero; for a given SNR, the Kendall tau statistics is maximum for zero time delay. The detector output statistics approximately follows a sine function. For specified signal and noise statistics the detector output can be used as a direct measure of the time delay. Computer simulation results show that the detector provides a better performance than that of Chan (1980) for all SNRs above 0.15.
 Publication:

ICC '82  The Digital Revolution, Volume 1
 Pub Date:
 1982
 Bibcode:
 1982icc.....1R...1H
 Keywords:

 Computerized Simulation;
 Correlation Detection;
 Nonparametric Statistics;
 Parameter Identification;
 Signal Detectors;
 Time Lag;
 Communication Theory;
 Signal Processing;
 Signal To Noise Ratios;
 Statistical Analysis;
 Communications and Radar