Search for a command to run...
This dataset provides a processed behavioural device-fingerprinting representation designed to support research on privacy-preserving identity establishment and continuous device authentication. The dataset is derived from controlled execution-response measurements collected from 11 Raspberry Pi devices operating under fixed experimental conditions. Measurements originate from 3 Raspberry Pi 3 Model B+ devices and 8 Raspberry Pi 4 Model B devices. Data collection was conducted in a controlled headless execution environment with fixed CPU frequency configuration, CPU core isolation, and reduced scheduling interference to ensure reproducibility of behavioural response signals. Each device was evaluated across 4 CPU cores and multiple reboot cycles. The dataset contains execution-dependent behavioural signals, including: Temperature-dependent execution responses GPU-related performance counter behaviour CPU hash execution response signals Pseudo-random execution response signals True random number generation: behavioural signals A comprehensive feature engineering pipeline was applied, incorporating: Temporal rolling statistical features Exponential moving dynamics Signed behavioural interaction features Robust statistical transformations Behavioural deviation modelling Representation-learning-oriented feature construction Unlike the CD2A dataset, which utilises a limited handcrafted feature representation focused primarily on QPU frequency behaviour, the PP-IDE dataset provides an expanded behavioural feature space designed to support: Privacy-preserving federated learning research Behavioural identity establishment modelling Continuous authentication system design Adversarial robustness evaluation Hardware-rooted behavioural fingerprint analysis Persistent hardware identifiers have been removed and replaced with stable pseudonymous device labels to preserve device-level behavioural consistency while ensuring privacy protection. This dataset is intended for research in behavioural device authentication, privacy-preserving machine learning, federated behavioural modelling, and security evaluation of embedded and IoT systems.