Pedestrian trajectory and intention prediction: a comprehensive review of models, datasets, and challenges

Key Points

• Pedestrian intention recognition models show promise in enhancing safety measures and urban planning.
• Key findings indicate a significant reliance on machine learning algorithms across varied datasets.
• Observational analysis covers various approaches, identifying gaps in model evaluation and real-world applications.
• Challenges in integrating different datasets underscore the need for standardized metrics and robust methodologies.