A waveform model for bat echolocation

Bats use active sonar to navigate and forage. Performance of an active sonar, be it biological or man-made, degrades rapidly in the presence of competing (jamming) transmissions. It is in this environment that bats flying in dense swarms, such as an evening emergence from a large roost, nevertheless avoid colliding with conspecifics and other obstacles. We have previously investigated the hypothesis that Brazilian free-tailed bats (Tadarida brasiliensis) use variations spectro-temporal “shape” to facilitate successful echolocation in dense groups [J. Acoust. Soc. Am. 154, A48 (2023)]. The present work employs a compact analytical waveform model based on low-order Chebyshev polynomial series—Chebyshev Polynomial Series Frequency Modulation (CPSFM)—to represent frequency modulation functions observed in bat echolocation signals. Following this approach, we fit isolated calls from swarm recordings to the CPSFM model and show that diversity in model parameters facilitates call distinguishability. Standard signal detection methods, such as cross correlation, enable quantitative estimation of detection performance, including rejection of interfering emissions and echoes. Results demonstrate that subtle but specific variation in spectro-temporal shape, captured in CPSFM parameters, can enable rejection of acoustic signals from conspecifics in dense swarms which may aid in jamming avoidance by bats.