Researches

Map-Matching, Sep.2023 - Now

• As part of generating Complete Trip data, the map-matching method based on the HMM model is used to perform map matching on human Location-based Service (LBS) data, aligning human movement trajectories with a digital map, such as Open Street Map (OSM).

Intelligent quality detection of lane rendering data,Aug.2022- Sept.2022

• Objective: To efficiently detect whether there are defects in the rendered generated map data for navigation
• Implemented pre-training with Masked Image Modeling using three hybrid CNN-LSTM neural network models
• Build the pre-training pipeline for the image recovery task and carried out the pre-training
• Tested different numbers of continuous sequence image inputs with different pre-training models
• Adjusted the pre-trained models for fine-tuning the image classification task
• Developed the pipeline for abnormal image detection to identify defected lane rendering image
• Delivery: An implementation report and participated in the HUAWEI 2022 Global AI Challenge

Robust Lane Detection through Self Pre-training with Masked Sequential Autoencoders,Sept.2021- Aug.2022

• Objective: To develop robust lane detection neural network model that can tackle challenging scenes
• Delivery: A research paper titled Robust Lane Detection through Self Pre-training with Masked Sequential Autoencoders and Fine-tuning with Customized PolyLoss

Figure.1
Figure.2
Figure.3

Research on the Traffic Capacity of Highway Region Ecological Environment—— A Case of G30 Zhangye,Sept.2021- Jun.2023

• Objective: To study the impact of highway traffic on the surrounding environment, calculate the carrying capacity of traffic
• Delivery: A postgraduate thesis, and a paper titled The Maximum Traffic Capacity of Highway Region: Under Road Region Environment Constrains

Figure.1
Figure.2

Research on ecological risk assessment and traffic carrying capacity of road areas in ecological function zones,Sep.2021 - Sep.2022

Lanzhou Jiaotong University | Research Assistant

• Objective: To study the impact of highway traffic on the surrounding environment, calculate the carrying capacity of traffic
• Wrote the research proposal.
• Arranged the required instruments and equipment for the research (portable weather stations, noise detectors, etc.), tested and debug the equipment
• Led the research team for a 5-day field data collection. Collected data includes regular air pollution data, meteorological data, traffic data, soil heavy metal data, noise data, etc.
• Established a road traffic and ecological risk evaluation index system for ecological function areas
• Determined the scope of the environmental impact of the highway is 600 meters using remote sensing datas
• Delivery: Completed a research report, and prepared a project proposal for the National Natural Science Foundation of China

Figure.1 Time series data (NDVI) using Landsat8 remote sensing images.
Figure.2 Correlation analysis between NDVI value and meteorology indexes.