Research

Multimodal Monitoring of Driver’s Alertness, University of Michigan

A research project to explore the potential of multimodal modeling for driving distraction detection and recognition Dec 2019-April 2020
• Simulating the driving process and collected data of driving performance and facial conditions under different distractors.
• Using Python with Open Face library to analyze the data of drivers’ facial reaction with different distractors and with Seaborn library to visualize the statistical data.

Generative heat sink design optimization, General Motors, Michigan, United States

A Multidisciplinary Design Project to research on algorithms to automatically generate and optimize heat sink designs
Jan 2020-June 2020
• Delivered an end-to-end process to synthesize and optimize dendritic (tree-like) heat sink designs automatically in a team of 7 students, ultimately reduced the design time and improved the energy efficiency of electronics for vehicles
• Developed an algorithm based on space colonization with Python to generate dendritic heat sink designs with less than 1% of parameters needed comparing with traditional way of 3D model representation
• Used PyAnsys to implement an automated process for thermal analysis of the generated heat sink CAD designs
• Applied differential evolution to iterate the heat sink designs based on their thermal performances