Study Chiang Mai, Thailand, December 23, 2025 – New wearable sensor technology shows promise for improving communication between nosework dogs and their human handlers through machine learning–based sniffing detection.
Nosework dogs carry substantial responsibility in public safety, from search-and-rescue deployments to narcotics and explosives detection. While their olfactory skills are exceptional, the subtle behavioral cues that accompany scent detection can take years for handlers to interpret reliably. A newly developed wearable device aims to bridge this communication gap by using audio and inertial measurement unit (IMU) sensors paired with machine learning algorithms.
The prototype, designed with canine comfort, handler usability, and environmental noise reduction in mind, incorporates a cardioid microphone and two IMU sensors. Researchers collected 20 minutes of labeled data from a trained nosework dog performing search tasks while outfitted with the device. In this preliminary phase, analysis focused on audio alone.
Using machine learning models trained on audio recordings, the system achieved 94.4% accuracy in classifying sniff-related sounds (idle vs. sniffing). This level of performance demonstrates strong potential for developing real-time tools that help handlers interpret subtle olfactory-search behaviors.
Future research aims to integrate IMU data to yield greater behavioral granularity—such as sniffing frequency, movement patterns, and search progression—providing deeper insights into how dogs work scent trails. Such systems could ultimately support training, improve deployment decisions, and strengthen the dog–handler relationship.
Source: Li, X., Hom, W., Wu, J., Verges, M., & Jackson, M. (2021). Wearable Sensors for Canine Nosework Sniffing Interaction. International Conference on Animal-Computer Interaction.
