Progress
This week, we worked towards making an end-to-end system. We put together a full functional sensor system, with all of the trinity sensors feeding data to the Pi. The pi listens for the audio and then proceeds to extract load sensor data as well as its capacitive readings. We extracted features from the audio signal including number of peaks, maximum amplitude, spectrum spread, and spectrum centroid. We then use the feature vector to classify the object using our Support Vector Machine. We also created a preliminary data set as well as the functionality of adding more training data and retraining the SVM as data comes Garbougie's way. Alvin and Tushita worked on building the end-to-end system, while Ananya worked on the signal feature extraction.