Engineering Design Projects (ES 100)

Title: Optimized Pneumatic System for Wearable Actuators

Members: Cameron Maltman, S.B. '20, Mechanical Engineering

About: Wearable, underarm actuating balloons are currently being designed by the Harvard Biodesign Lab to help those who work above their heads or those who have had had strokes raise their arms at the shoulder. This project was concerned with the design, build, and verification of a pneumatic system that filled these underarm airbags quickly, efficiently, quietly, and at a low cost while remaining light. This was done through the creation of a mathematical model, experimental setup, and the final pneumatic system.

Title: HAND.IO: A Wearable Neural Interface Based on Forearm EMG.

Members: Charlie Colt-Simonds, S.B. '20, Electrical Engineering

About: The Hand Actuation Neuro-signal Detector for Intentional Object-control (HAND.IO) is a wearable device that translates hand gestures into meaningful input for 3D object control. The device uses 8-channel electromyography to detect muscle activation in the forearm associated with hand movement using dry, non-invasive electrodes printed onto a flexible band. The captured biosignals are amplified and filtered by an analog front-end circuit assembled on a custom printed circuit board before being digitized and streamed to a user’s computer via Bluetooth. Off-board computing classifies gestures that serve as user controls for movement of a simulated 3D object.

Title: A Hybrid Brain-Machine Interface and Control Paradigm for Prosthetic Limb Operation

Members: Rohan Jha, S.B. '20, Bioengineering

About: Individuals with upper limb loss or upper limb motor dysfunctions are unable to effectively manipulate their environment or carry out Activities of Daily Living (ADLs). Advancements in prosthetic limbs have allowed these individuals to regain their autonomy. Though there have been many advances, control schemes for prosthetic limbs are still limited in their scope, intuitiveness, and reliability. To improve upon the current control paradigms for high-level prosthetic limbs, a hybrid hierarchical Brain-Computer Interface approach based on electroencephalography (EEG) and electromyography (EMG) from residual shoulder muscles is proposed and implemented. In detail, EEG data is used to classify broad action intents of reaching and grasping. In a supplemental fashion, EMG data is used to modify these movements in x, y, z space to increase accuracy and resolution during actuation. EEG/EMG data acquisition systems are built, and various time-domain, frequency-domain, and time-frequency domain feature extraction methods are implemented to process self-acquired EEG and EMG data. Classification accuracy reaches 99% for 7-class EMG directional movement. Classification accuracy for EEG data reaches 81% for a 3-class problem and 97% for a 2-class problem.

Title: An In Vitro Model of the Aortic Arch to Improve Outcomes in the Repair of Aortic Defects

Members: Andrew Yang, S.B. '20, Bioengineering

About: This project consisted of the design and fabrication of an anatomically accurate in vitro model of an infant aorta using silicone. The model can be used to better understand the effects that patient-specific aortic defects can have on the pressure, vessel compliance, and complex impedance properties of the aorta. Current understanding of these aspects of the aorta is limited, resulting in poor outcomes and requiring multiple treatments for those with aortic defects. To create the model, patient CT scans were used to generate a computer-aided design (CAD) model. From this CAD model, molds were 3D printed, and molded silicone was used to model the vessels. Water and glycerol were used as working fluids to model blood. A ViVitro pulse duplicator pump was used to pump the working fluid throughout the model, and two compliance chambers and a resistive element were used to regulate the pressure. In tests with a 6 month old patient’s aorta, arterial pressure was found to be 65 mmHg at systole and 30 mmHg at diastole, compliance was found to be 0.48 %Vol/Pressure Change, and blood flow velocity was found to be 30.9 cm/s. Verification on these data was done by comparison with literature data for similarly aged patients, as well as two sample F-tests for variance to ensure consistency in the data. In making this in vitro model, the ultimate goal of this project was to allow for a streamlined, patient-specific process in treating aortic defects.

Title: Minimally Invasive Blood Glucose Prediction

Members: Colin Harvey, S.B. '20, Bioengineering

About: This project showcases the development of a blood glucose prediction algorithm for diabetic persons in order to help them better regulate their blood sugars and avoid adverse health effects. The algorithm uses data from a continuous glucose monitoring device already commonly worn by diabetics and a wrist-worn vital monitor akin to a Fitbit. The algorithm then uses this information to generate a prediction based on past experiences observed during training. All this is then implemented in a mobile application so that the user has access to the predictions at all times in a lightweight, fast app that presents the data in an intuitive, understandable format.

Title: Wearable Medical Device to Assist with the Treatment of Pediatric Cancer and Infectious Diseases

Members: Anna Raheem, S.B '20, Bioengineering

About: During ES100, I designed a wearable monitoring device to assist with the early detection of infections in the case of pediatric cancer or infectious diseases, and developed a temperature algorithm, an integrated circuit and an external housing for the proposed device. The wearable monitors and analyzes the user’s axillary body temperature in order to alert them if their body temperature is indicative of an infection. This device would allow high-risk users to constantly monitor their temperature to know when they have an infection and need to seek medical care without having to worry about contributing to an already overburdened health-care system.