CHIRP-Like Signals: Estimation, Detection and Processing A Sequential Model-Based Approach
Chirp signals have evolved primarily from radar/sonar signal processing applications specifically attempting to estimate the location of a target in surveillance/tracking volume. The chirp, which is essentially a sinusoidal signal whose phase changes instantaneously at each time sample, has an interesting property in that its correlation approximates an impulse function. It is well-known that a matched-filter detector in radar/sonar estimates the target range by cross-correlating a replicant of the transmitted chirp with the measurement data reflected from the target back to the radar/sonar receiver yielding a maximum peak corresponding to the echo time and therefore enabling the desired range estimate. In this application, we perform the same operation as a radar or sonar system, that is, we transmit a “chirp-like pulse” into the target medium and attempt to first detect its presence and second estimate its location or range. Our problem is complicated by the presence of disturbance signals from surrounding broadcast stations as well as extraneous sources of interference in our frequency bands and of course the ever present random noise from instrumentation. First, we discuss the chirp signal itself and illustrate its inherent properties and then develop a model-based processing scheme enabling both the detection and estimation of the signal from noisy measurement data.