Core Coursework

Indirect Signal Recognition of Visual Aided Satellite Positioning

Capstone Project | Team Leader

Combine the camera with a satellite signal receiver. Extract the spatial environment's contour from the image, convert it into the satellite's occlusion perspective, and determine the satellite's position relative to the device using mathematical models and simulations based on satellite map data. This enables the filtering of satellite signals with erroneous reflections, thereby improving positioning accuracy.

Indirect Signal Recognition of Visual Aided Satellite Positioning

Final Year Project

Explore the implementation process and application conditions of path planning methods such as the A* algorithm and RRT algorithm in UAV applications. Use MATLAB to model the two-dimensional and three-dimensional working environments of UAVs, simulate the operational status of the algorithms under obstacle conditions, and evaluate the search efficiency and computational overhead of the algorithms.

Core Coursework

AI Recommendation for Business Owners on Promotion Campaigns

Dissertation

Developed a generative AI system that creates tailored marketing campaigns based on coupon and membership data. The system was fine-tuned on a pre-trained model using the QLoRA technique with 4-bit quantization, striking a balance between computational efficiency and recommendation quality. RAG was integrated into inference to incorporate merchant-specific historical data and real-time insights, thus maintaining context and enhancing relevance in the generated promotion strategies.