SonoTraceLab -- A Raytracing-Based Acoustic Modelling System for Simulating Echolocation Behavior of Bats

TL;DR

SonoTraceLab offers a GPU-accelerated, open-source acoustic simulation framework that combines raytracing for specular reflections with a curvature-driven Monte Carlo approach to diffraction, enabling realistic, post-processed ERTF-shaped signals for biological and robotic sonar scenarios. By operating on STL meshes up to about a million triangles and bypassing full PDE solvers, it provides a practical tool for rapid hypothesis testing in bat echolocation research and biosonar design. The authors validate the approach through ERTF fidelity tests, reflector-only scenes, and biosonar-driven prey-discrimination simulations, demonstrating biologically relevant cues and directional control. This work facilitates accelerated exploration of echolocation behaviors and sensor design, with potential impact on both biology and robotics.

Abstract

Echolocation is the prime sensing modality for many species of bats, who show the intricate ability to perform a plethora of tasks in complex and unstructured environments. Understanding this exceptional feat of sensorimotor interaction is a key aspect into building more robust and performant man-made sonar sensors. In order to better understand the underlying perception mechanisms it is important to get a good insight into the nature of the reflected signals that the bat perceives. While ensonification experiments are in important way to better understand the nature of these signals, they are as time-consuming to perform as they are informative. In this paper we present SonoTraceLab, an open-source software package for simulating both technical as well as biological sonar systems in complex scenes. Using simulation approaches can drastically increase insights into the nature of biological echolocation systems, while reducing the time- and material complexity of performing them.

Figures (9)

• Figure 1: A typical setup for an echolocation experiment. Panel A shows the Global X, Y, and Z axes, the local X, Y, and Z axes of the echolocation sensor (red, green, and blue arrows respectively), the 3D model of the echolocation sensor (in this case a model of the head of the bat Micronycteris microtus and the object to be ensonified, which in this case is a generic leaf with a model of a dragonfly. Panel B shows a still frame from a high-speed video showing a Micronycteris microtus performing the same task cosys-labMicronycterisMicrotisCapturing2018).
• Figure 2: Block diagram illustrating the sequential stages of the simulation implementation for synthesizing acoustic signals.
• Figure 3: Overview of the acoustic BRDF (Bidirectional Reflectance Distribution Function) applied to the loaded mesh using local curvature to modulate the parameters $\alpha$ and $k$. The top panel shows the opening angle of the BRDF (i.e., how specular or omnidirectional a reflection on the local vertex is) and the reflectance strength (i.e., how strong a reflection is from this vertex). It should be noted that the opening angle towards areas with high curvature shows a significantly wider opening angle, which is desired because diffraction echoes arise at locations with high curvature and manifest themselves as almost omnidirectional reflections. However, these diffraction echoes should be significantly weaker than the specular reflected echoes, which are reflected in the BRDF reflection strength plot.
• Figure 4: Panel a) shows the used coordinate system with two receiver arrays for the left and right ear, a single source, and the coordinate system attached to the sensor. Panel b shows an overview of the raytracing-based partial solution to the Helmholtz equation. In panel b1, a set of rays are cast into all directions of the frontal hemisphere. Panel b2 shows the intersection of one ray with the scene geometry, which is reflected around the surface normal. The ray further propagates (b3) and reflects from the scene geometry a final time (b4). The BRDF (red ellipse) is sampled by each microphone, depicted by two black lines intersecting the ellipse.
• Figure 5: Overview of the approach of solving the diffraction aspects of the Helmholtz equation. We calculate the curvature using local differential geometry and sample locations with high curvature to create diffraction echoes. The generated samples then sample the local BRDF properties and provide input to the impulse response generation module.