Deep learning-driven pedestrian path prediction: Challenges, interpretability, and future research directions

Evangeline R. C. , Raviraj P.

Abstract

Pedestrian trajectory prediction is fundamental to the safety and efficiency of autonomous systems, intelligent transportation, robotics, and urban management. This paper presents a comprehensive survey of recent advances in deep learning methodologies for pedestrian trajectory forecasting. Focusing on architectures including Recurrent Neural Networks (RNNs), Long Short-Term Memory (LSTM) networks, Convolutional Neural Networks (CNNs), Graph Neural Networks (GNNs), Transformer-based attention models, and generative frameworks, this review covers how these techniques model spatial-temporal dependencies, social interactions, and contextual environment cues. It further discusses multimodal sensor fusion, evaluation metrics, benchmark datasets, applications, challenges such as occlusion and real-time deployment, and outlines future research directions emphasizing goal awareness, transfer learning, and explainability. A critical comparative analysis highlights performance improvements over the past decade and identifies gaps for ongoing innovation, providing researchers and practitioners with a structured understanding of current capabilities and emerging trends.

Key Words

convolutional neural networks (CNNs); graph neural networks (GNNs); long short term memory (LSTM) networks; recurrent neural networks (RNNs)

Address

Evangeline R. C. — Department of Information Science and Engineering, Nitte Meenakshi Institute of Technology, Nitte (Deemed to be University), Bengaluru, Karnataka, India/ Department of Computer Science and Engineering, GSSS Institute of Engineering and Technology for Women, Mysuru,Affiliated to Visvesvaraya Technological University (VTU), Belagavi, India
Raviraj P. — Department of Computer Science and Engineering, GSSS Institute of Engineering and Technology for Women, Mysuru, Affiliated with Visvesvaraya Technological University (VTU), Belagavi, India

PDF Viewer

Preview uses the same access rules as Full Text PDF (subscription, purchase, or open access).

← Back to Volume 11, No. 3