All ETDs from UAB

Advisory Committee Chair

Mohammad R Haider

Advisory Committee Members

Emil Jovanov

Karthikeyan Lingasubramanian

Dalton S Nelson

Nasim Uddin

Document Type


Date of Award


Degree Name by School

Doctor of Philosophy (PhD) School of Engineering


Chronic wound recovery time is long and easy to recur, which brings a high burden on the patient's physical and psychological aspects. The increase in the cost of per capita medical services and the increase in the number of people in demand for medical services have led to an increasing requirement for real-time monitoring and healing in patient's health outside of the hospital. For this purpose, the smart bandage is being developed to monitor the real-time situation of the wound. With the development of wireless sensor networks, the increasing number of sensing units has placed a great burden on the central data processing unit. Because of the limitation of improvement in CMOS technology in device scaling, memory capacity, and power consumption, this research has been motivated to the CMOS based oscillator neural network (ONN) for analog or non-Boolean computing applications for energy-efficient computational units. In this dissertation, we have demonstrated an analog computational platform us-ing coupled ONN for real-time data acquisition, analysis, and processing systems suitable for smart bandage applications. When the sensor detects abnormal data, the ONN first processes the detected data through its analog processing unit. Traditional digital pro-cessing requires the transmission of all abnormal data. In contrast, the analog processing unit only needs to detect the convergence time and frequency as two indicators and com-pare it with the healthy skin storage mode to transmit only the lack of synchronization data. In this way, the ONN improves the computational efficiency in the application of the portable device. In the implementation of this hardware, through the simulation of both the Cadence and Simulink, we have achieved the correspondence between hardware circuits and mathematical models, which proves the feasibility of implementing the circuit to the smart bandage application. Two-layer hierarchical AM model is introduced, since the detected pattern will always need to compare with all storage patterns to search the nearest neighbor, compared to a single AM model, the two layers hierarchical clustering greatly reduces the pattern that needs to be stored through multiple layered processing.

Included in

Engineering Commons



To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.