Trajectory Prediction and Research Datasets

This project thread covers data-driven perception research for autonomous driving, with emphasis on trajectory prediction, benchmark development, and research infrastructure for datasets. It includes survey work, multi-task benchmark design, and privacy-aware tooling for research data release.

Core Questions

• What limitations prevent current trajectory-prediction methods from generalizing to real traffic conditions?
• How can benchmark datasets better support multi-task autonomous-driving research?
• How can sensitive driving data be shared responsibly while preserving research value?

Methods

• Survey and taxonomy building for trajectory prediction
• Benchmark dataset design and curation
• Privacy-aware anonymization pipelines
• Data-driven evaluation for perception and scene understanding

Representative Outputs

• Trajectory Prediction for Autonomous Driving: Progress, Limitations, and Future Directions
• EMT: A Visual Multi-Task Benchmark Dataset for Autonomous Driving in the Arab Gulf Region
• Priv-DATA: Privacy-Aware Dataset Anonymization Tool for Autonomous Vehicles Research

Research Stack

• Dataset preparation and annotation workflows
• Benchmark design for multi-task learning
• Privacy-aware data processing pipelines
• Python-based evaluation tooling