Smart Clothing and On-Device Gesture Recognition for Controlling Smart Buildings Systems

More and more devices are integrated into buildings, transforming them into intelligent structures that adapt to their occupants. Due to the large number of integrated systems and the continuously growing complexity, it is difficult for users to control these systems. This work explores the possibility of using smart clothes to control a building using hand gestures. A prototype device was developed that can be easily integrated into ordinary garments by making minor modifications. The design of the device prioritizes relatively small dimensions, to facilitate easy integration, low energy consumption, modularity, efficiency, and ease of use. Gesture acquisition was achieved using a 6-axis IMU (Inertial Measurement Unit). Classification of gestures was implemented directly on the wearable device using artificial intelligence (AI) techniques. Gesture recognition was achieved solely through real-time processing available on the device, adapted to limited computational power available, enabling controlling smart building systems without the need for a cloud service or internet. Wearing a smart device-integrated garment, the user can control various systems in a smart building, in an easy and more natural manner, using hand gestures.