Cybersecurity
Early Security Warning for Autonomous Systems
Code: ANX-RP-001
Develop an early warning system that detects security anomalies in autonomous vehicle sensor networks before they propagate to decision-making layers.
Why it matters
Autonomous vehicles process millions of sensor readings per second. A compromised sensor can lead to catastrophic failures. Current intrusion detection systems are too slow for real-time autonomous navigation.
Research gap
Existing IDS focus on network-level threats but ignore sensor-level adversarial attacks specific to autonomous driving pipelines.
Research questions
- 01 How can we detect adversarial sensor inputs in under 10ms?
- 02 What ML architectures are suitable for streaming anomaly detection on edge hardware?
- 03 How do we minimize false positives without missing novel attack vectors?
Suggested technologies
Required skills
Team and supervisor
2-4 researchers, supervised by Muhammad Qasim.
Suitable for: CS, SE, DS, AI, Robotics.
Expected deliverables
- Research paper on lightweight anomaly detection for AV sensors
- Open-source Rust library for CAN bus anomaly detection
- Benchmark suite comparing detection latency vs accuracy tradeoffs
Publication
Lightweight Anomaly Detection for Autonomous Vehicle Sensor Networks
Published · 2026-03-15
Work on this problem
Apply to the lab and this problem becomes your supervised research project.