A foot-mounted indoor navigation system implementation is used in this project. The module has four MPU9150 IMUs (Inertia Measurement Units), Atmel ACR32UC3C microcontroller, Bluetooth and USB interfaces. Each IMU consists of a three-axial accelerometer and a three-axial gyroscope, which measure the accelerations and angular velocities in three coordinates. Details can be found in http://www.openshoe.org/?page_id=1205
To collect data, the modules can be connected to a PC either using USB cable or using Bluetooth wirelessly, with sample rates of 1000Hz and 100Hz respectively. To fully study FOG patterns, USB connection is used to sample more data. Studies of lower sampling rate for Bluetooth connection will be done in the future based on the performance of higher sampling rate algorithms.
To configure the device correctly and collect data, the computer needs to be installed with AVR Studio 5 and the environment needs to be setup. In the scope of this project, a pre-configured laptop and a user guide are provided to the physicians for convenient acquisition.
15 subjects were enrolled in this study at Washington University School of Medicine. All patients have reported a clinical history of FOG events, and recordings started with patients being in the “off” state according to neurologist. The acquisition protocol consists of 6 tasks: (1) walk straight at a normal pace from stationary for a distance of 20 feet and turn around by 180 degrees and walk back, (2) initiate and pause walking to follow instructions from physicians, (3) in-place 180 degree turns to the left and right, (4) walk through a narrow space with two sets of chairs on the sides that span a 3 meter track, (5) walk straight for 3 meters with a wood block of 6 inches in the middle, (5) walk backwards for 3 meters with no obstruction. Task (1) and (2) were repeated twice, and the rest tasks were repeated for 5 times. The physician labeled the FOG onset and offset time when the subject was experiencing FOG during the task.