Roadside Perception and Vehicle-to-Infrastructure (V2I) Communication

Multi-sensor fusion approach combining LiDAR, camera, and GPS technologies with advanced machine learning algorithms for robust traffic perception and real-time V2I communication in smart intersection environments.

• Hardware Integration: Roadside Perception Unit (RSPU) with high-resolution cameras, 3D LiDAR, GPS with Real-Time Kinematic (RTK) positioning
• Computer Vision: Hierarchical adaptive background subtraction algorithm for robust vehicle detection in varying lighting and weather conditions
• Sensor Fusion: Multi-modal data integration for enhanced detection accuracy and reduced false positives in complex traffic scenarios
• Communication Architecture: Low-latency V2I communication protocols with real-time traffic state broadcasting to connected vehicles
• Auto-calibration System: Automated sensor calibration and dynamic parameter adjustment for long-term deployment reliability
• Validation Framework: Dual validation approach using CARLA simulation for controlled testing and real-world deployment for performance verification

Roadside Perception Unit (RSPU) system architecture

V2I communication network topology

Real-time traffic detection and monitoring results

Successfully deployed smart intersection system with proven real-time traffic monitoring capabilities and reliable V2I communication performance validated through both simulation and field testing.

• Real-time Performance: Achieved sub-100ms detection and communication latency for time-critical traffic applications
• Detection Accuracy: Superior performance compared to existing background subtraction algorithms in both simulation and real-world scenarios
• Weather Robustness: Maintained consistent detection performance across clear, rainy, and low-light conditions
• Communication Reliability: Demonstrated stable V2I communication with 99.5% message delivery rate within operational range
• Field Validation: Successfully deployed and tested at UC Riverside Innovation Corridor with multiple vehicle types and traffic scenarios
• Scalability Demonstration: Proven system architecture scalable to multiple intersection deployments with centralized coordination