Hardware Trojan Detection Circuit Using Behavioral Baseline

The patent performs preliminary actions by collecting and storing operational data, power consumption patterns, and timing information during the production test phase. This preliminary data collection enables detection of delayed-activation HTs in the field by comparing current behavior against the stored baseline, resolving the timing contradiction between fast verification and delayed detection.

The patent implements continuous monitoring of the IC's operational characteristics throughout its lifecycle. The system continuously collects data on power consumption, timing variations, and operational patterns, maintaining an ongoing comparison against the baseline. This continuous action ensures that HTs with delayed activation are detected when they become active, while maintaining fast detection response.

The patent segments the detection system into multiple independent monitoring components, each analyzing specific aspects of side channel emissions (power consumption, timing, operational patterns). Each segment processes and analyzes its specific parameter separately, then the results are combined for comprehensive HT detection. This segmentation reduces the measurement precision burden on any single component while maintaining overall detection coverage.

The patent introduces an intermediary processing layer that collects, normalizes, and correlates data from multiple side channel sources. This intermediary system aggregates information from power consumption sensors, timing measurements, and operational data, then compares the combined data against the stored baseline. The intermediary processing reduces the precision requirements of individual sensors while maintaining comprehensive detection coverage.

The patent implements self-service by using the IC's own operational characteristics (power consumption, timing variations, operational patterns) as the monitoring parameters. The system leverages existing signals already present in the IC during normal operation, rather than requiring separate dedicated monitoring circuits for each parameter. This self-service approach improves detection accuracy while minimizing the addition of complex monitoring hardware.

The patent creates a universal monitoring framework that can detect various types of HTs by analyzing multiple operational parameters simultaneously. The same baseline comparison mechanism handles different HT types (delayed activation, steganographic leakage, side channel attacks) by collecting and analyzing power consumption, timing, and operational data. This multi-functional approach improves detection accuracy across multiple threat vectors without proportionally increasing device complexity.

The patent extracts only the most critical operational characteristics and statistical patterns from the vast amount of available data during production testing. Instead of storing complete operational traces, the system extracts key features such as power consumption distributions, timing variations, and operational pattern statistics. This extraction process significantly reduces the volume of stored data while maintaining the capability to detect HTs through comparison with the extracted baseline.

The patent transforms raw operational data into statistical parameters and normalized representations during the baseline collection phase. The system converts detailed operational traces into summary statistics, probability distributions, and normalized patterns that capture the essential behavior characteristics. This parameter transformation reduces the quantity of stored data while preserving the information needed for HT detection, as the comparison process uses these compressed statistical representations rather than raw data.