Opioid Overdose Real-Time Wearable Monitor with Fuzzy Inference System

Vo, Dat Tien
Graduate Program:
Electrical Engineering
Master of Science
Document Type:
Master Thesis
Date of Defense:
June 26, 2018
Committee Members:
  • Robert Anthony Gray, Thesis Advisor
  • Sedig Salem Agili, Committee Member
  • Seth Wolpert, Committee Member
  • Opioid
  • overdose
  • fuzzy
  • inference
  • system
  • embedded
  • wearable
Opioid-involved deaths may be easily avoided if an injection of Narcan can be delivered on time to a person suffering from an opioid overdose. However, these victims of opioid users often do not find their way to the hospitals in time to receive the Narcan shots, and thus risk unnecessarily losing their lives. Therefore, this project proposes a system that can detect when a person is overdosing so that necessary action such as hypothetically contacting emergency responder, Narcan injection, etc. can be taken in a timely manner. Specifically, the device is aimed to be a wearable technology that performs constant non-invasive physiological and pathophysiological tests to detect when a person becomes apneic, which is the primary overdose reaction as opioid receptors in the respiratory centers become saturated. The innovation in our approach introduced the feasibility of Fuzzy Inference System embedded on a microprocessor and integrated with multiple sensors to diagnose true positive opioid overdoses with high reliability. Fuzzy Inference System features such as Fuzzy sets, IF-THEN rules and Fuzzy logical operation were leveraged in this project to overcome component flaws and the lack of clearly defined physiological thresholds for opoid overdose indicators. In the end, the finding shows that Opioid Fuzzy Inference system can be embedded on open- source hardware with less than 0.2% difference, when compared against MATLAB simulation.