Data-Quality Framework for Multi-source & Multi-modal Data Analytics
Embedded Files
The reliability and safety of AI systems, particularly in autonomous driving, depend fundamentally on the quality of multisource and multimodal data. Advancing from model-centric approaches toward a task-oriented framework that links data quality with system performance and decision accuracy is essential for building trustworthy and efficient intelligent systems.
Articles
📃 Zhou, Y., Chen, H., & Sha, K. (2025). "A Novel Multi-layer Task-centric and Data Quality Framework for Autonomous Driving." arXiv preprint arXiv:2506.17346.
📃 Zhou, Y., Tu, F., Sha, K., Ding, J., & Chen, H. (2024, July). "A survey on data quality dimensions and tools for machine learning invited paper." In 2024 IEEE International Conference on Artificial Intelligence Testing (AITest) (pp. 120-131). IEEE.
Federated Learning in Dynamic Connected Autonomous Vehicles
Federated learning (FL) enables Connected Autonomous Vehicles (CAVs) to continuously learn from diverse driving environments while keeping raw sensor data on each vehicle, preserving privacy and reducing bandwidth demands. It bridges algorithmic intelligence with real-world system constraints, linking sensing, communication, and computation to ensure adaptive, safe, and scalable autonomy under heterogeneous conditions
Articles
📃 Cherukuri, K. S., Sha, K., & Huang, Z. (2025). "Enabling Federated Object Detection for Connected Autonomous Vehicles: A Deployment-Oriented Evaluation." arXiv preprint arXiv:2509.01868.
📃 Cherukuri, K. S., Sha, K., & Ding, J. (2025, August). "Enabling federated learning for object detection in connected autonomous driving using yolo with the flower framework." In 2025 34th International Conference on Computer Communications and Networks (ICCCN) (pp. 1-6). IEEE.
An Edge-Based Approach to Robust Multi-Robot Systems in Dynamic Environments
Design an efficient edge infrastructure with intelligent collaboration mechanisms
Dynamic and efficient organization of heterogeneous edge resources.
Reinforcement-learning algorithms for resource allocation.
Make intelligent decisions and interact with a dynamic and uncertain environment
Models of environment
Estimation of motion intentions of people and adaption
Build a secure and robust computation framework
User-centric approach private-data and privacy-sensitive computation
Trusted edge-centric architecture
Human-robot collaborative decision making
Other Ongoing Projects
Blockchain technology for consortium applications (NASA)
Large language model (LLM) for intrusion detection
EKG based exercise detection, user identification
Attacks to multi-agent systems
Object detection for Connected Autonomous Vehicles (CAVs)
Edge computing for IoT and resource management
Collaborative multi-agent system using Large Language Model
Extended reality (XR) for Autism (NSF)
Page updated
Google Sites
Report abuse