We recently achieved our first milestone in our project. Aasif finished the skeleton of his sliding window algorithm with help from our TA Mentor, Aadu, and we tested it out with accelerometer data from our MBed. Aasif then created .arff files that we can use with a machine learning software package called WEKA.
I finished building the EMG circuit, but I'm waiting on a serial breakout to begin testing and fine-tuning the filters. I got inspiration for the circuit from the Advancer Technologies Muscle Sensor and from Vivek Menon on what op-amps and instrumentation amplifiers work well for EEG and EMG circuits.
Upon talking to Sid, the ESE Lab Manager, he told us we needed to use a BioPac system to isolate the electrodes from the power sources we use for the MBed for safety reasons. I then talked to Seville, the Lab Manager for the BE Lab, who provided us the BioPac MP35 system. We recently hooked up electrodes to our chin sites to see what signals we would get from chewing, and examined them on the BioPac software package, and we saw some very distinct signals from chewing and resting. As a result, we are pretty optimistic that with a good EMG circuit, we will be able to properly identify and characterize chewing vs. non-chewing events with Aasif's sliding window algorithm and several other sensors.
As of now, we also changed our project name from ShockWeight to VibeWeight. It may be a bit more applicable with the type of haptic feedback we will be providing.
-Neil
No comments:
Post a Comment