Browsing by Author "Shi, Liang"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Driving Risk Assessment Based on High-frequency, High-resolution Telematics Data
    Guo, Feng; Qian, Chen; Shi, Liang (SAFE-D: Safety Through Disruption National University Transportation Center, 2022-04)
    The emerging connected vehicle and Automated Driving System (ADS), the widely available advanced in-vehicle telematics data collection/transmitting systems, as well as smartphone apps produce gigantic amount of high-frequency, high-resolution driving data. These telematics data provide comprehensive information on driving style, driving environment, road condition, and vehicle conditions. The high frequency telematics data has been used for several safety areas such as insurance pricing, teenage driving risk evaluation, and fleet safety management. This report study advances traffic safety analysis in the follow aspects: 1) characterize the high-frequency kinematic signatures for safety critical events compared to normal operations; and 2) develop models to distinguish and predict crashes from normal driving scenarios based on the high frequency data. Two deep learning models were developed. The first one combines the strength of convolutional neural network (CNN), gated recurrent unit (GRU) network and extreme gradient boosting (XGBoost). The second approach is based on a novel variational inference for extremes (VIE) to address the rarity of crashes. The models proposed in this project can benefit a variety of traffic research and applications including connected vehicles and ADS real-time safety monitoring, NDS data analysis, ride-hailing safety prediction, as well as fleet and driver safety management programs.
  • Thumbnail Image
    Enhanced Feature Representation in Multi-Modal Learning for Driving Safety Assessment
    Shi, Liang (Virginia Tech, 2024-12-03)
    This dissertation explores innovative approaches in driving safety through the development of multi-modal learning frameworks that leverage high-frequency, high-resolution driving data and videos to detect safety-critical events (SCEs). The research unfolds across four methodologies, each contributing to advance the field. The introductory chapter sets the stage by outlining the motivations and challenges in driving safety research, highlighting the need for advanced data-driven approaches to improve SCE prediction and detection. The second chapter presents a framework that combines Convolutional Neural Networks (CNN) and Gated Recurrent Units (GRU) with XGBoost. This approach reduces dependency on domain expertise and effectively manages imbalanced crash data, enhancing the accuracy and reliability of SCE detection. In the third chapter, a two-stream network architecture is introduced, integrating optical flow with TimeSFormer with a multi-head attention mechanism. This innovative combination achieves exceptional detection accuracy, demonstrating its potential for applications in driving safety. The fourth chapter focuses on the Dual Swin Transformer framework, which enables concurrent analysis of video and time-series data, this methodology shows effective in processing driving front videos for improved SCE detection. The fifth chapter explores the integration of corporate labels' semantic meaning into a classification model and introduces ScVLM, a hybrid approach that merges supervised learning with contrastive learning techniques to enhance understanding of driving videos and improve event description rationality for Vision-Language Models (VLMs). This chapter addresses existing model limitations by providing a more comprehensive analysis of driving scenarios. This dissertation addresses the challenges of analyzing multimodal data and paves the way for future advancements in autonomous driving and traffic safety management. It underscores the potential of integrating diverse data sources to enhance driving safety.