Using the clean portion of a CHIRP pulse to determine the time of arrival of a hydroacoustic signal under strong reflection conditions

Determining the time of arrival of a hydroacoustic signal is one of the fundamental problems in positioning systems that use propagation-time measurements, especially solutions based on Time of Arrival (ToA) and Time Difference of Arrival (TDOA). Under strong multipath conditions, estimation accuracy deteriorates because the direct signal overlaps with reflected waves. This study evaluates whether restricting the analysis to the initial, least distorted part of a CHIRP pulse can improve time-of-arrival determination. Experiments were conducted in a laboratory swimming-pool environment for ranges of 2 m and 4 m using 30–25 kHz CHIRP signals with durations of 1, 2, 5 and 10 ms and GNSS PPS synchronization. The analysis showed that the influence of reflections on the received-signal structure increases with pulse duration, and that for longer pulses the overlap between reflected waves and the direct signal becomes increasingly evident. The proposed approach, based on using the clean part of the pulse, substantially improved time-of-arrival accuracy. In representative configurations, the mean absolute error was reduced from 20.7 to 2.7 \\(\\mu \\) s and from 175.6 to 3.1 \\(\\mu \\) s, while maintaining 100% correct detections within a \\(\\pm 20\\,\\mu \\) s tolerance. For a range of 4 m and a 2 ms pulse, the mean absolute error was 1.7 \\(\\mu \\) s, also with 100% correct detections. The results confirm that using the initial, weakly distorted signal fragment can significantly increase the stability and accuracy of time-of-arrival estimation in strongly reflective environments.