Introduction

The summer of 2015 Siddharth Chandra again set out to locate a professor who would be able to help him improve VIBe during the prototyping stage. He reached out to Professor Durand of Case Western Reserve University and was ecstatic to have been given the opportunity to work for 4 weeks at the University. During this internship, Siddharth’s main goal was to focus on further developing Virtual Guardian, the algorithm that detects stimming behavior in VIBe.

Research

It is estimated that one third of autistic children also suffer from epilepsy and Dr. Chiang’s (a post doctoral student at Case Western Reserve University) research in Prof. Durand’s lab intrigued Siddharth. Dr. Chiang’s research focus is the effect of audio and visual stimuli on refractory epilepsy. Siddharth worked closely with Dr. Chiang on creating a prototype that would allow Dr. Chiang to test his hypothesis in human trials. Creating prototypes for Dr. Chiang allowed Siddharth to further improve his skills in rapid prototyping. Simultaneously, Siddharth learned about the promising use of other types of sensory stimulation to benefit autistic children. Currently, IPAT is investigating the possibility of utilizing similar technology in order to make VIBe a more effective tool for autistic children.

Siddharth had also contacted  Anil Damle, a PhD student in computational mathematics at Stanford University, about methods by which to improve the accuracy of Virtual Guardian. Siddharth remembered that Anil had mentioned that the Fourier Transform would be something worth looking into. Concurrently with research done for a write up assigned at school, Siddharth Chandra had begun to research the Fourier Transform so that he could apply it to VIBe. After months of studying the math and the purpose of the Fourier Transform, Siddharth concluded that it would radically improve Virtual Guardian’s processing abilities. His next step, which took place at CWRU, was to create an algorithm that utilized the transform. He soon realized that attempting to code the Fourier Transform by hand was far beyond his level of competency. Fortunately, at CWRU, he was given access to MATLAB. This new programming language is specifically designed for signal processing, which means that many of the functions (like the Fourier Transform) required for coding are already predefined in the library. After spending a few days understanding and learning the basics of the Fourier Transform, Siddharth began to design a smarter algorithm.

Conclusion

At the conclusion of his internship at Case Western Reserve University, Siddharth was able to successfully import data captured by the accelerometer and apply the Fourier Transform. From this calculated data, the program is able to identify the most prominent frequency. This algorithm is still in the development phase as small tweaks are constantly being made to make the algorithm more accurate.