TACET: A Context-Appropriate Acoustic-Social Navigation Method for Quadruped Robots
New research presents a method for reducing locomotion noise in quadruped robots in social environments.
Quadruped robots operating in sensitive environments like hospitals and care homes must navigate quietly to respect social contexts. The TACET method integrates a vision-language reasoner with a reactive controller to optimize both movement and noise levels based on the robot's surroundings.
Key Features of TACET
- Utilizes a compact behavior token to manage gait, speed, and social cost.
- Maintains personal-space compliance while minimizing acoustic intrusion.
- Demonstrated a reduction in locomotion noise by up to 9.3 dBA.
This innovative approach ensures that robots can operate effectively without disturbing the social environment.