RESEARCH
Applied research at the intersection
of AI, security, and embedded systems.
Investigating how neuromorphic computing can make embedded systems more secure, efficient, and intelligent.
ACTIVE RESEARCH
Neuromorphic Intrusion Detection
Problem
Modern vehicles rely on CAN buses for internal communication. These networks were designed without security and are vulnerable to intrusion attacks. Existing detection systems struggle with real-time constraints and resource limitations of automotive embedded systems.
Approach
Applying Spiking Neural Networks — a biologically-inspired computing paradigm — to detect intrusions through voltage fingerprinting. SNNs process information as discrete spikes, offering significant energy efficiency for resource-constrained automotive ECUs.
Methodology
Collecting voltage fingerprint data from CAN networks, designing SNN architectures for intrusion detection and ECU attribution, and evaluating against traditional ML approaches on automotive-grade embedded hardware.
My Contribution
Experimental design, SNN model implementation, and performance evaluation. Focused on bridging theoretical neuromorphic research and practical embedded deployment with millisecond latency budgets.
INTERESTS
PUBLICATIONS
Neuromorphic Voltage-Fingerprint Intrusion Detection for Automotive CAN Networks
Under Review
ECU Attribution Using Spiking Neural Networks
Under Review