Vehicle microphones calibrated using balloon bursts to track sound sources
Spatio-Temporal Calibration of a Vehicle-Mounted Microphone Array for Autonomous Driving
Sound
Summary
Cars use cameras and LiDAR to see around them, but sometimes they miss things, especially sounds from other vehicles. The authors show how to line up microphones attached around a car with the car’s other sensors using balloon pops as sound targets. These pops are easy for the microphones to hear and the LiDAR to locate, helping to accurately place the microphones without needing complicated models of the car's shape. Their method worked well on a real car with microphones spread widely around it and needed only a few balloon pops to get a good calibration.
What this means in practice
- •For autonomous vehicle engineers: Calibrate microphones mounted around a vehicle using balloon bursts to align audio with LiDAR and camera data for better sound source tracking.
- •For urban traffic monitoring teams: Deploy calibrated vehicle-mounted microphone arrays to detect and locate vehicles based on sound in complex urban environments.
Authors
Marvin Seegert, Richard Taupitz, Korbinian Moller, Johannes Betz
Abstract
Microphones let an autonomous vehicle sense into areas its cameras and LiDAR cannot see, by picking up the sound that other traffic participants emit. Locating those sources and fusing them with the visual perception requires the microphone array to be calibrated and registered relative to the vehicle's other sensors. Existing calibration methods assume small, unobstructed indoor rigs, whereas a vehicle body casts an acoustic shadow that straight-line propagation cannot represent, and to our knowledge no calibration of an array distributed around a vehicle has been reported. We propose a pipeline that uses a balloon burst as a single cross-modal target, which the LiDAR locates in metric coordinates, every microphone hears as a sharp onset and a camera times. Each observation is weighted by its own timing variance, which discounts shadowed arrivals instead of discarding them and needs no model of the vehicle. On a full-size vehicle with eight microphones over a 4.9 m aperture, the held-out range residual of 6.37 cm approaches the 6.07 cm floor set by the onset timing noise, and the calibrated array tracks a moving source to within a few degrees of the vehicle's own LiDAR. A dozen bursts spread around the vehicle suffice, and spreading them matters more than adding more.