"Good Time to Talk" KAIST Develops Situation Awareness and Mental Health Check Technology

A technology has been developed that finds the "right moment to start a conversation" to check the mental health of the other person. The feature of this technology is that it does not cause stress to the user with untimely questions and allows the user to actively record their psychological state.

Provided by KAIST

On the 24th, KAIST announced that a research team led by Professor Lee Euijin from the School of Computing developed a situation-aware "multi-modal smart speaker system" that supports users in recording their own psychological state.

This system is designed to ask mental health-related questions at the optimal moment by recognizing the user's surrounding situation in real time.

The speaker equipped with multi-modal sensors comprehensively analyzes various sensor data such as indoor movement, lighting, noise, and carbon dioxide to detect the user's presence and activity, then requests self-tracking surveys at moments suitable for the user to respond.

This reduces factors that can cause negative emotions such as stress and irritation from random surveys and maximizes the efficiency of survey responses.

Professor Lee Eui-jin's research team. Provided by KAIST

To evaluate the user experience of the developed system, the research team installed self-tracking smart speakers in 20 single-person households and conducted a one-month field study, building a total of 2,201 mental health survey response datasets.

Additionally, through dataset analysis, they identified survey response patterns according to response time and activity context, as well as which input method users prefer in different situations: voice input (VUI) or touch input (GUI).

The results confirmed that when the speaker recognizes the user's surrounding situation in real time and asks mental health-related questions at the optimal moment, the response rate is higher, and that most participants in the field study preferred touch input, which allows faster responses, over voice input.

Professor Lee said, "Based on the results of this study, our research team aims to develop a mental health management support smart speaker capable of performing roles similar to a human counselor."

Meanwhile, this research was conducted with the support of the LG Electronics-KAIST Digital Healthcare Research Center.

This content was produced with the assistance of AI translation services.