State Estimation and Motion Prediction of Vehicles and Vulnerable Road Users for Cooperative Autonomous Driving: A Survey

Ghorai, Prasenjit; Eskandarian, Azim; Kim, Young-Keun; Meh, Goodarz · 2022 · ROSA P / United States. Department of Transportation. National Transportation Library [distributor]

archive: archived pipeline: cataloged verified

Get this paper ↗ (full text — opens at the source; we link to it, we don't host it)

Summary

**State Estimation and Motion Prediction of Vehicles and Vulnerable Road Users for Cooperative Autonomous Driving: A Survey** This survey addresses the critical challenge of ensuring safety and reliability for autonomous vehicles (AVs) operating in mixed traffic environments containing manually driven vehicles, partially autonomous systems, and vulnerable road users (VRUs) such as pedestrians and cyclists. The authors are motivated by the limitations of current isolated AV systems, which rely solely on onboard sensors and often fail in complex scenarios due to sensor occlusion, extreme weather, or unpredictable human behavior. The paper aims to provide a comprehensive overview of state estimation and motion prediction techniques, emphasizing the shift toward Connected and Cooperative Autonomous Driving (CCAD) where vehicles share perception data via Vehicle-to-Everything (V2X) communications to enhance situational awareness. The study employs a literature review methodology, categorizing existing research into three primary domains: ego-vehicle perception, state estimation and motion prediction, and cooperative perception and navigation (CPN). The authors analyze exteroceptive sensors (cameras, lidar, radar), comparing their performance metrics, costs, and environmental robustness. They evaluate deep learning-based object detection architectures, distinguishing between two-stage frameworks (e.g., Faster R-CNN) and single-stage frameworks (e.g., YOLO, SSD), and assess their accuracy on benchmark datasets like KITTI and ImageNet. Furthermore, the survey reviews classical and modern methods for tracking and predicting the trajectories of vehicles and pedestrians, highlighting the use of sensor fusion techniques (e.g., radar-camera, lidar-camera) to mitigate individual sensor weaknesses. Key findings indicate that while deep learning has significantly improved object detection accuracy, substantial gaps remain in real-world applicability. Pedestrian detection accuracy drops significantly in 3D benchmarks compared to 2D datasets, and current methods struggle with abnormal behaviors and occlusions. The survey identifies four major challenges in ego-centric detection: physical sensor limitations in adverse weather, accuracy discrepancies between controlled datasets and real-world conditions, reliability issues regarding sensor faults, and time latency constraints where processing delays exceed safe braking distances at high speeds. Regarding cooperative perception, the authors note that while sharing processed perception data (such as occupancy grids) extends the field of view and aids in blind-spot detection, challenges persist regarding bandwidth limitations, data reliability, and the lack of standardized protocols for cooperative behavior. The significance of this work lies in its synthesis of disparate research areas into a unified framework for CCAD. By highlighting the limitations of isolated autonomy and the potential of cooperative systems, the paper provides a roadmap for future research. It underscores the necessity of developing robust sensor fusion algorithms, improving real-time computational efficiency, and establishing standardized communication protocols to enable safe, efficient, and scalable autonomous driving in heterogeneous traffic scenarios.

Key finding

Cooperative perception and navigation significantly enhance autonomous vehicle safety and efficiency by extending field-of-view and enabling seamless coordination, though challenges remain in sensor reliability, data latency, and handling abnormal pedestrian behavior.

Methodology

review

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via bulk_ingest_rosap on 2026-05-23 (6 acquisition events logged).

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• anticipation