TACET: Quiet Navigation System for Sensitive Environments

Paper: TACET: Context-Appropriate Acoustic-Social Navigation for Quadrupeds

Authors: Sungsan Park, Young-Sik Shin, Sanghyun Kim

Executive Summary

Recent research has focused on improving the interaction of quadruped robots in sensitive environments such as hospitals, care homes, and quiet offices. The study introduces TACET, a novel approach to robotic navigation that not only considers where these robots should move but also how quietly they can do so. This dual focus on spatial and acoustic awareness makes TACET particularly relevant in settings where noise levels can impact comfort and privacy.

Traditional methods of robotic navigation often overlook the acoustic aspect of movement. They either treat the robot as a uniform source of sound or simply aim to reduce noise without considering the specific context of the environment. TACET addresses this gap by using a method that infers the social context of the robot's surroundings from its own perspective. This context is crucial for determining not just the appropriate path for the robot, but also the acceptable noise level for its locomotion.

The TACET system operates by integrating a vision-language reasoning component with a fast-reactive control mechanism through a compact behavior token. This token serves a dual purpose: it helps the robot decide both where to walk and how loudly to move, based on the social cues it perceives. Additionally, TACET maintains a memory of recently observed individuals, allowing the robot to continue responding to social dynamics even when these individuals are no longer in its direct view.

In practical tests on a real quadruped robot, TACET demonstrated a significant reduction in locomotion noise, achieving reductions of up to 9.3 decibels at matching speeds. This decrease is substantial, considering that even small changes in sound levels can affect human comfort in quiet environments. Furthermore, TACET effectively maintained compliance with personal space norms in all evaluated scenarios, ensuring that the robot operated within socially acceptable boundaries while minimizing acoustic disruption to less than 2.9 decibels.

The implications of this research are significant for industries where robotic integration can enhance service delivery without compromising the human experience. By ensuring that robots can navigate environments both quietly and respectfully, companies could explore applications in elder care, healthcare delivery, or even in office environments where maintaining a tranquil atmosphere is essential.

This research demonstrates not just a technical advancement in robotic locomotion but also emphasizes the importance of social context in the deployment of autonomous systems. By aligning robot behavior with human social norms, organizations could improve how these technologies are received and utilized in everyday settings. For further details on the TACET project, information is available at the project’s webpage.

Academic Abstract

Quadruped robots entering hospitals, care homes, and quiet offices must be context-appropriate not only in where they move but in how loudly they move: a legged robot's locomotion noise, dominated by foot-ground impacts, is itself a social variable. Prior social navigation respects human space but treats the robot as acoustically uniform, while quiet-locomotion methods reduce noise to an operator-specified, context-blind level. We present TACET, a context-appropriate acoustic-social navigation method that infers social context from the robot's egocentric view and decides both where it walks and how loudly, coupling a slow fine-tuned vision-language reasoner to a fast reactive controller through a single compact behavior token, . The same token conditions both a social costmap (where to go) and a quiet locomotion policy (how loudly to move), while a structured out-of-view memory keeps recently seen people in the reasoner's context after they leave the camera view. On a real quadruped, context-conditioned locomotion lowers locomotion noise by up to 9.3 dBA at matched speed, and across our scenarios the full method keeps personal-space compliance at 100% with low acoustic intrusion (<=2.9 dBA), jointly improving spatial and acoustic performance in the evaluated scenarios. The project page is available at https://rcilab.khu.ac.kr/tacet/.