Advisor: Prof. Jacob Whitehill
Co-Advisor: Prof. Gregory Fischer
Co-Advisor: Prof. Berk Calli
In the last decade, socially assistive robots have been used in therapeutic treatments for individuals diagnosed with Autism Spectrum Disorders (ASDs). Preliminary studies have demonstrated positive results using the Penguin for Autism Behavioral Intervention (PABI) developed by the AIM Lab at WPI to assist individuals diagnosed with ASDs in ABA therapy treatments. In recent years, power-efficient embedded AI computing devices have emerged as a powerful technology by reducing the complexity of the hardware platforms while providing support for parallel models of computation.
This new hardware architecture seems to be an important step in the improvement of socially assistive robots in ABA therapy. In this thesis, we explore the use of a power-efficient embedded AI computing device and pre-trained deep learning models to improve PABI’s performance. Four main contributions are made in this work. First, a robot-enhanced ABA therapy framework and a software architecture for PABI is designed and implemented. Second, a multifactorial experiment is completed to benchmark the performance of three popular deep learning frameworks over the AI computing device.
Experimental results demonstrate that some deep learning frameworks exploit the GPU power while others exploit the multicore ARM-CPU system of the device for its parallel model of computation. Third, the robustness of state-of-the-art pre-trained deep learning models for feature extraction is analyzed and contrasted with the previous approach used by PABI. Experimental results indicate that pre-trained deep learning models overcome the traditional approaches in some fields; however, combining different pre-trained models in a process reduces its performance. Fourth, a patient-tracking algorithm based on an open-set approach is developed.
Contrary to the traditional approaches where a set of images of the patient are taken and a deep learning model trained, the tracking algorithm developed combines patient data, a face encoding deep learning model, and unsupervised methods for the tracking algorithm improving the usability and autonomy of the robot during ABA therapies. Experimental results show that the developed algorithm can perform as well as an algorithm based on the traditional approaches.