Measuring Large Scale Human Social Behavior


mlshsbRecent research has shown that it is possible to sense whether people are talking or engaged in socially relevant activities (e.g. gesturing, laughing) using their body motion alone. This is particularly helpful in scenarios where relying on audio to measure socially relevant behavior is challenging because the ambient noise levels are too high.

The goal of this project is to develop devices that are able to capture human behavior related to their body movement at a very high level of sensitivity. The device must be small and light and be able to run for at least a day on the same battery while storing all sensor data as it records. The devices will be operated as part of a network with many nodes. The devices must be synchronized, have the capacity to scale to dense crowded scenarios and the potential to measure other body behaviors such as bio signals. For that purpose the devices must evaluated in a system configuration of at least 32 devices.

This device should be able to measure proximity and, if possible other features like distance between devices, acceleration, orientation, velocity, and the ability to store all the sensor data in a fully synchronized fashion. They also need to have audio recording capabilities as an experimental reference tool. If time is available, we would also like to explore the possibility for connecting physiological sensors to allow measurement of body states while socializing.

Regarding synchronization, we would also need to synchronize the sensor data with video cameras capturing people wearing the devices. Ideally, we'd be able to find solutions that can easily identify where a person is in the video given that we have someone's wearable sensor data.



Andre Bossche

Jeroen Bastemeijer