Projects </>

Connected Reference Marker (CRM) System Demonstration Development

I worked on the development of a Connected Reference Marker (CRM) System Demonstration. My responsibilities encompassed both hardware and software development for a positioning system utilizing Ultra Wideband (UWB) technology. I conducted testing for distance measurement and coordinate positioning, which included enhancing the performance of the underlying algorithms.

CARLA Automated Driving Simulation Project

In my leadership role, I guided a team of over 40 undergraduates in researching connected and automated driving concepts, with a particular focus on collaborative driving simulations within the CARLA environment. My responsibilities included directing the development and integration of roadside camera-based object detection and tracking systems. This process encompassed data collection, labeling, transfer learning, and integration into the CARLA simulation. Additionally, I was instrumental in designing and overseeing the development of vehicle control strategies, which included longitudinal and lateral control, deep reinforcement learning model development, and executing obstacle avoidance tasks.

City of Madison Connected Vehicle Corridor Project

I developed a data pipeline system using socket programming to connect with Roadside Unit (RSU) devices at intersections, enabling the acquisition of real-time traffic signal data. I processed this data in accordance with SAE J2735 standards and managed its storage using a PostgreSQL database. Additionally, I created an interactive dashboard to display both historical and real-time traffic signal data from RSUs. My work also involved conducting tests with On-Board Unit (OBU) devices in connected vehicles to receive and process information from RSU devices at intersections. Furthermore, I implemented a digital twin to simulate the collected data in a high-fidelity virtual environment, facilitating real-time monitoring, comprehensive analysis, and behavior simulation.

Enhancement of WisTransPortal V-SPOC Database and Web Application

My key contributions include transforming and enhancing existing MATLAB traffic data quality assessment algorithms into Python, which has significantly improved processing efficiency. I have also been responsible for managing daily data quality tests using detector data from the Milwaukee and Madison metropolitan areas. Additionally, I have upgraded web-based tools to facilitate more efficient traffic data analysis and extraction.

Automated Vehicle Detection and Tracking

I developed a vehicle detection model based on YOLOv4, tailored for both images and videos. To enhance the model's effectiveness, I implemented transfer learning techniques, which resulted in achieving a mean Average Precision (mAP) superior to standard benchmarks. I created a vehicle tracking model utilizing the Deep SORT algorithm, further advancing the capabilities in vehicle tracking technology.

Vehicle Lane-Changing Maneuver Recognition and Trajectory Prediction

I developed an LSTM-based model for identifying future lane-changing maneuvers and combined GNN and LSTM for predicting vehicles' future positions. My integration of lane change detection and trajectory prediction significantly outperformed benchmark error rates.

Other Projects </>

Steel Structure Frame Generator Based on Cellular Automaton and Genetic Algorithm

This project applies cellular automata (CAs), simple models adept at complex behavior generation, to simulate bracing systems in steel structures. These systems are crucial for resisting lateral forces like earthquakes and wind. Cellular automata's ability to evolve intricate patterns from basic rules offers a fresh approach to designing efficient bracings with minimal material use. The project employs Rhino and Grasshopper for modeling and optimizes the design through a Genetic Algorithm (GA) in Galapagos, blending computational design with structural engineering in an innovative way.

Data Mining in Music and Eurythmics Performance Evaluation

My primary focus was on analyzing Eurythmics performances to identify key features such as beats, tempo, and energy for evaluation models, which output performance grades. I also constructed an Arduino-based circuit equipped with geophones and microphones to conduct experiments for vibration and soundwave data collection. Furthermore, I developed and evaluated a model using MATLAB and Python, employing Support Vector Machine (SVM) methods for training and predicting performance outcomes.

Seismic Motion Simulation with Gaussian Processes Model

I employed a Gaussian Process model to analyze the 1940 El Centro earthquake accelerogram data. This involved simulating earthquake ground motions based on the recorded data and predicting pseudo-acceleration response spectra using Gaussian Processes Models, providing valuable insights into earthquake dynamics and impact.