MSight Roadside Intelligence System
π What is MSight?
MSight is an edge-cloud intelligent infrastructure system developed by the Mcity and the Michigan Mobility Transformation Lab (MTL) of University of Michigan.
It enables next-generation roadside intelligence by tightly integrating with intersection facilities such as:
- Roadside sensors (cameras, LiDAR, radar)
- Traffic signal controllers and cabinets
- V2X roadside units (RSUs)
- Edge compute boxes and cloud services
MSight comes with state-of-the-art roadside perception, sensor fusion, trajectory prediction and nearmiss detection algorithms, providing a unified platform for perception, communication, and prediction across diverse roadside environments.

βοΈ What Can MSight Do?
The MSight System is a general-purpose, modular edgeβcloud framework designed to be highly customizable and support a wide range of roadside and cooperative-perception applications connecting roadside and vehicle systems.
It can, in theory, power any infrastructure-based intelligent transportation application.
Some representative use cases include:
-
π°οΈ Roadside Sensor Data Streaming
Stream multi-modal sensor data (video, LiDAR, radar, etc.) in real time with low latency to interested edge or cloud services. -
π Roadside Object Detection
Perform on-device or edge-based detection of vehicles, pedestrians, and other road users. -
π€ Cooperative Perception (CP)
Fuse and share detection results via RSU using standardized formats such as SAE J3224 SDSM, extending vehicle awareness beyond line of sight. -
β οΈ Near-Miss and Crash Detection
Continuously monitor and analyze interactions between road users to identify near-miss or crash-prone events in real time. -
π§ Proactive Safety and Hazard Prediction
Predict hazardous situations (e.g., potential collisions) in advance and send timely warnings to vehicles through V2X communication to prevent accidents. -
π¦ Traffic Signal Violation Detection
Detect red-light violations by analyzing vehicle trajectories in relation to signal phase and timing data. -
π And Moreβ¦
Thanks to its modular and extensible design, MSight can host custom analytics, data management, or AI modules for future roadside applications.
π Learn More
Explore the documentation using the left navigation panel or top tabs:
- Installation
- Quick Start
- System Architecture
- Deployments
- Tutorials
- Cloud Integration
- Developer Guide
- API Reference
Β© University of Michigan β Mcity & Michigan Mobility Transformation Lab