/STMotionExploration

Under Distinguished Professor William Kaiser, our group has been exploring embedded Machine Learning through the STMicroelectronics SensorTile.

Primary LanguageMATLAB

STMotionExploration

Under Distinguished Professor William Kaiser, our group has been exploring motion classification through the STMicroelectronics SensorTile.

The Group:

  • Ryan Tabrizi
  • Arunan Elamaran
  • Keshav Vyas
  • Kavin Balakrishnan
  • Michael Davydov
  • Atiksh Rao

Within this repository is all the code we used to test and find features and test draft systems.

The folders and their purpose:

  • The MatLab folder has code that we developed on MatLab to graph the values collected on each of the axes for each motion. By doing this, we were able to spot clear, unique features.

  • The FeatureAnalysis folder has code that we used to help us find features that we would have not been able to observe with just our eyes. Additionally, using this code allowed us to quickly analyze the results for potential features after each training set.

  • In the DecisionTree folder is code that mirrors what we developed in the final SystemWorkbench algorithm. We used this code for quick testing of features and the techniques that we would use in our SystemWorkbench algorithm.

The final SensorTile SystemWorkbench algorithm is stored in this repository.

Our research paper is linked here.