Motion Aware Sensor Platform (MASP)

Type: Diploma thesis
Student: Gemma Colino
Advisor: Holger Junker, Pawel Lukowicz

MASP is a wearable platform with multiple sensors conceived to recognize human motion activities. To achieve this scope, 3 axis accelerometers, which measure static (gravity) and dynamic acceleration, are used. The concept of the platform is shown in next the next figure.

The structure of the platform is hierarchical. The system is divided in subnetworks, associating each subnetwork to a human extremity. The slaves get the information from the accelerometers and transmit it to the master of its subnetwork. The master pre-process these raw values, extracting features, and sends these features to the Main Core. The Main Core, after some processing, will produce the output of the system, that is, the activity the user is performing. The platform is also automatically configurable, that means, that, when adding a new device, MASP recognizes it and get the necessary information from it to perform the communication.
The communication between the devices that conform the platform is done under I2C Bus Protocol.
The implemented algorithm is able to distinguish between different positions of the limb, changes between positions, oscillations between positions and shaking around a position.