IoT/Edge AI
AI at the edge: Real-time intelligence on resource-constrained devices
About This Project
This project focuses on implementing a real-time object detection or activity analysis system that processes video, sensor data, and other inputs on resource-limited or mobile/IoT devices. The solution is designed to operate with minimal latency, efficiently handling ML/vision tasks while providing real-time feedback, alerts, and edge analytics. The key challenge addressed is optimizing sophisticated AI models to run effectively on devices with limited computational resources while maintaining accuracy and responsiveness.
Core Concepts
- Edge computing architecture
- Model optimization for resource-constrained environments
- Real-time object detection algorithms
- Sensor data fusion techniques
- Low-latency inference pipelines
- Energy-efficient ML deployment
Key Knowledge/Skills
- Neural network architectures (CNNs, RNNs)
- Model compression techniques
- Embedded systems programming
- Real-time processing optimization
- Hardware acceleration (TensorRT, ONNX)
- IoT protocol integration (MQTT, CoAP)
Coursework Covered
Edge Computing & Embedded Systems for AI