Sim2Real Transfer for Audio-Visual Navigation with Frequency-Adaptive Acoustic Field Prediction

TL;DR

This paper introduces a novel sim2real transfer method for audio-visual navigation by disentangling acoustic field prediction and waypoint navigation, utilizing a frequency-adaptive strategy to bridge the spectral gap between simulation and real-world audio data.

Contribution

It presents the first approach to sim2real transfer for audio-visual navigation, incorporating frequency-adaptive acoustic field prediction and real-world validation on a robot platform.

Findings

Improved navigation performance in real-world tests.

Effective spectral gap measurement and adaptation.

Successful transfer of policies from simulation to real robot.

Abstract

Sim2real transfer has received increasing attention lately due to the success of learning robotic tasks in simulation end-to-end. While there has been a lot of progress in transferring vision-based navigation policies, the existing sim2real strategy for audio-visual navigation performs data augmentation empirically without measuring the acoustic gap. The sound differs from light in that it spans across much wider frequencies and thus requires a different solution for sim2real. We propose the first treatment of sim2real for audio-visual navigation by disentangling it into acoustic field prediction (AFP) and waypoint navigation. We first validate our design choice in the SoundSpaces simulator and show improvement on the Continuous AudioGoal navigation benchmark. We then collect real-world data to measure the spectral difference between the simulation and the real world by training AFP models that only take a specific frequency subband as input. We further propose a frequency-adaptive strategy that intelligently selects the best frequency band for prediction based on both the measured spectral difference and the energy distribution of the received audio, which improves the performance on the real data. Lastly, we build a real robot platform and show that the transferred policy can successfully navigate to sounding objects. This work demonstrates the potential of building intelligent agents that can see, hear, and act entirely from simulation, and transferring them to the real world.