Pulse-train based time-delay estimation improves resiliency to noise

Time-delay estimation accuracy in echolocating systems decays for increasing levels of noise until a breakpoint is reached, after which accuracy deteriorates by several orders of magnitude. In this paper we present a robust fusion of time-delay estimates from multiple pings that significantly reduces the signal-to-noise ratio corresponding to the accuracy breakpoint. We further show that a simple average of the time-delay estimates does not shift the breakpoint to a lower signal-to-noise ratio. The proposed fusion of multiple pings has the potential of improving the resilience to noise of echolocating systems such as sonar, medical ultrasound and geo-seismic surveys, hence increasing their potential operating range and reducing health/environmental hazards