Facilities

At a Glace

Click any picture below to learn more.

Edney Innovation Center

The UTC Research Institutes maintains a strategic presence at the Edney Innovation Center in downtown Chattanooga. This space supports collaboration with industry, government, startups, and community partners while providing a visible front door for research engagement beyond campus. The primary tenant of this space is the Center for Urban Informatics and Progress.

Key Features:

• Flexible meeting and collaboration space
• Convenient access for community engagement and events
• Dedicated space for research translation and applied innovation initiatives including:
  • Driving Simulator
  • Traffic Lab
  • Traffic Management Center
  • 1/10th Scale CDA Vehicle Platform

High Performance Computing

Our High-Performance Computing (HPC) environment provides advanced computational resources to support research and innovation across multiple departments and centers at the university. Faculty, staff, and students leverage these resources for large-scale machine learning model training, artificial intelligence workflows, data-intensive simulations, and high-performance analysis. The cluster supports AI and machine learning workloads through Kubernetes/Kubeflow pipelines, alongside traditional Slurm batch-based workflows, enabling scalable, cross-disciplinary research.

Research Computing Infrastructure Overview

Scheduler: Kubernetes (Kubeflow)

• dgx nodes – Total of 2 nodes
  • Total Compute:  256 CPU cores, 2TB RAM, 1280GB VRAM
  • Access: Reserved for CUIP
  • Each node includes:
    • Two AMD EPYC 7742 64-core processors
    • 1 TB RAM
    • Eight NVIDIA A100 80GB GPUs

Scheduler: Slurm (Traditional HPC)

epyc nodes – Total of 18 nodes

• Total Compute: - 2,304 CPU cores, 9TB RAM, 2560GB VRAM
• Access: Open Access for University
• Each node includes:
  • One AMD EPYC 7662 128-core processor
  • 512 GB RAM
  • Two NVIDIA A100 80GB GPUs (16 of 18 nodes)

tennessine nodes – Total of 32 nodes

• Total Compute: 896 CPU cores, 4TB RAM, 512GB VRAM
• Access: Open Access for University
• Each node includes:
  • Two Intel Xeon E5-2680 14-core processors
  • 128 GB RAM
  • One NVIDIA P100 16GB GPU

firefly nodes – Total of 4 nodes

• Total Compute: 80 CPU cores, 768GB RAM, 512GB VRAM
• Access: Open Access for University
  • Each node includes:
  • Two Intel Xeon Gold 6148 20-core processors
  • 192 GB RAM
  • Four NVIDIA Tesla V100 32GB GPUs

lookout nodes – Total of 4 nodes

• Total Compute: 160 CPU cores, 1TB RAM, 256GB VRAM
• Access: Open Access for University
• Each node includes:
  • Two PowerPC POWER9 20-core processors
  • 256 GB RAM
  • Four NVIDIA Volta 16GB GPUs 

Mobile Traffic Unit

Our research infrastructure includes a mobile data collection trailer equipped with multiple NVIDIA edge computing units for on-site processing, along with an integrated sensor suite comprising LiDAR, optical cameras, and thermal imaging systems. The trailer operates off-grid via a solar power system with battery storage, ensuring sustained deployment in remote or field locations. All onboard systems are accessible remotely, enabling real-time data monitoring, retrieval, and system management without requiring physical presence at the deployment site. This mobile platform provides flexible, self-contained computational and sensing capabilities that can be rapidly repositioned to support diverse research needs across varying environments and conditions.

Traffic Lab

The CUIP Traffic Lab simulates intersections in a controlled environment. Researchers are able to validate their results before deploying to a real-world environment by utilizing the two traffic cabinets, mock utility pole, and mounted Roadside Units (RSUs) located in the lab. Additionally, the CUIP Traffic Lab allows for testing of interoperability for Intelligent Traffic Systems (ITS)

Key Features:

• Siemens M60 and Yunex Blade traffic controllers
• Several RSUs from leading manufacturers
• Simulated intersections, allowing for V2X preemption testing
• Legacy and modern traffic cabinets
• Mock utility pole to test mounting of devices
• Real and simulated GNSS signals

Driving Simulator

The Simulator provides a high-fidelity virtual environment for studying urban traffic, pedestrian activity, and human-in-the-loop driving behavior. By integrating the photorealistic CARLA 0.10.0 UE5 driving simulator with the traffic modeling capabilities of PTV VISSIM 2026, the system recreates key intersections along the MLK Smart Corridor in Chattanooga, TN, within a safe, repeatable, and fully controllable virtual setting. This co-simulation framework enables researchers and city partners to evaluate traffic operations, pedestrian movement, signal timing strategies, and vehicle–environment interactions without requiring disruptive or costly on-road deployments.

Key Features:

• Corridor-Scale, multi-intersection simulation representing key segments of MLK Smart Corridor
• Bidirectional CARLA–VISSIM integration synchronizing vehicle, pedestrian, and signal states while combining high-fidelity visualization and traffic modeling
• Realistic pedestrian–vehicle interactions at crosswalks and signalized intersections
• Traffic signal timing and control logic governed directly within VISSIM
• Support for interactive driving experiments through a controllable ego vehicle and customizable scenarios
• Four-camera driver-monitoring setup capturing steering actions, pedal activity, and driver behavior
• Safe, repeatable scenario testing without on-road deployment
• Flexible platform for traffic research, safety analysis, and smart-city experimentation

Traffic Management Center

CUIP's Traffic Management Center (TMC) serves as the operational heart of our Smart Corridor testbed. Designed as a live command environment, the TMC integrates data from cameras, LiDAR sensors, traffic signals, and environmental systems into a unified visualization wall. This space enables researchers, engineers, and partners to monitor infrastructure performance, analyze real-time traffic activity, and evaluate advanced transportation technologies as they operate in the field.

Fourteen high-resolution screens span the width of the lab wall, each displaying a different layer of the corridor's digital ecosystem — from live video feeds to AI-powered analytics dashboards.

Key Features:

• Live Camera Feeds: Real-time views from street cameras, thermal cameras, and traffic signal cameras.
• LiDAR Monitoring: Live detection of vehicles, pedestrians, and cyclists with safety event visualization.
• AI-Powered Traffic Analytics: Automated insights on traffic flow, crossings, and roadway activity.
• System & Infrastructure Status: Continuous monitoring of sensors, devices, and data systems.
• Environmental Data: Air quality and multimodal transportation information.
• Research & Demonstration Space: A live environment for testing, showcasing, and validating smart transportation technologies.

Vehicle Based Research Platforms

The AV Research Platform is a real-world, vehicle-based testbed designed to support research in perception, localization, control, and connected automation. The platform integrates advanced sensors, onboard computing, and open-source autonomy software to enable safe testing of automated driving functions under controlled and urban conditions.

It serves as a bridge between simulation and deployment, allowing researchers to validate algorithms on a full-scale vehicle equipped with real sensors and communication systems.

Key Features

• Full-scale research vehicle equipped with multi-modal sensors
• LiDAR, cameras, GNSS/IMU, and radar integration
• Onboard high-performance computing for real-time processing
• ROS 2-based autonomy stack
• Real-time perception, sensor fusion, and object detection
• Vehicle-to-Infrastructure (V2X) communication capability
• Manual and supervised autonomous driving modes
• Data logging for offline analysis and model validation
• Support for ODD definition and staged automation testing

1/10th Scale CDA Platform

The 1/10th-scale CDA platform is a high-fidelity, scaled autonomous platform designed to bridge the gap between simulation and full-scale cooperative driving research.

Evaluating autonomous systems on full-scale vehicles is often cost-prohibitive and logistically complex. Our 1/10th-scale Cooperative Driving Automation (CDA) platform provides a safe, repeatable, and scalable environment to study the most pressing challenges in urban mobility - specifically occlusion and cooperative perception. By integrating professional-grade sensing with edge computation, this platform allows us to develop and validate V2X (Vehicle-to-Everything) strategies and "digital twin" synchronizations before they are deployed in real-world.

Key Features:

• High-Performance Computing: Powered by the NVIDIA Jetson Orin Nano, enabling real-time onboard perception and GPU-accelerated sensor fusion.
• Advanced Sensing Suite: Equipped with RPLiDAR S2 (360° range sensing) and OAK-D Lite stereo cameras for high-resolution RGB and depth estimation.
• Multimodal Perception Stack: Integrated AI pipeline featuring object detection, instance segmentation, and Vision-Language Models (VLMs) for semantic scene understanding.
• Closed-Loop Control: Precision movement via NEMA stepper motors and AS5600 magnetic encoders for highly accurate trajectory following.
• Digital Twin Synchronization: Real-time mapping of physical vehicle data into 2D (Tkinter) and 3D (Blender) digital environments for live monitoring and post-run analysis.
• V2I Connectivity: Seamless communication with infrastructure-mounted cameras to "see" around corners and mitigate blind spots created by urban structures.