CS4326 AI on the Edge: Enabling Physical Intelligence

This course covers advanced concepts and practical implementations of artificial intelligence on edge-computing devices in support of naval-mission objectives. It covers both theoretical foundations and practical applications of AI at the edge, including hardware-acceleration techniques, model optimization, and deployment strategies. Students will learn about edge AI technologies including computer vision, natural-language processing, federated learning, autonomous robotics, intelligent networking, and generative AI models specifically designed for resource-constrained environments. This course builds on core AI and ML fundamentals to develop practical skills in deploying intelligent systems in the field using contemporary hardware (e.g., NVIDIA Jetson) and popular frameworks and tools (e.g., TensorFlow Lite/PyTorch, ONNX, and ROS).

Prerequisite

None

Lecture Hours

3

Lab Hours

1

Course Learning Outcomes

At the end of this course the student should be able to:

• Describe the key challenges and opportunities of deploying AI at the edge versus in the cloud.
• Demonstrate understanding of hardware acceleration techniques for AI inference on resource-constrained devices.
• Implement and optimize computer vision and natural-language processing models for edge devices.
• Apply federated-learning techniques to distribute model training across multiple edge nodes.
• Design and implement intelligent autonomous robotic systems using ROS and edge AI.
• Deploy and evaluate generative AI models on edge computing platforms.
• Apply AI agentic workflows on embedded devices (e.g. network analysis, automated red teaming, robot control/swarming).
• Develop reproducible research by producing appropriate NPS GitLab repository items with source code (using requirements files and object-oriented concepts) that will run on the course hardware, and an IEEE/ACM conference style five-page report describing their project outcomes.