Abstract

This document describes techniques and architectures for securing a computing device against physically coerced access by utilizing dynamic autonomic physiological state signatures. Processing circuitry of a computing device (e.g., a primary computing device, a wearable sentinel device, or an industrial control terminal) continuously determines a dynamic physiological baseline from sensor data (e.g., heart rate, galvanic skin response, inertial data). A physiological baseline and anomaly engine utilizes a machine learning model, which accounts for contextual states such as active exercise tracking to filter false positives, to detect a distress state corresponding to a non-voluntary autonomic deviation relative to the dynamic physiological baseline. When an authentication request is received, a keyguard manager intercepts the request based on the detected distress state, notwithstanding a valid physical identity match. In response to intercepting the authentication request, the computing device generates a simulated failure interface. This interface denies access to restricted data while outputting an error notification, masking a true security lockout state from an unauthorized user to provide plausible deniability. To secure this lockout state against circumvention, a duress flag is cryptographically asserted inside a hardware-isolated trusted execution environment (TEE) and committed to a hardware partition replay protected memory block (RPMB) to block unauthorized state reset attacks. Furthermore, localized wireless protocols, including Bluetooth Low Energy (BLE) and Ultra-Wideband (UWB), propagate state synchronization across paired peripheral devices, while the operating system returns specific hardware exception codes directly to application framework layers to seamlessly redirect execution paths to high-friction verification challenges.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.

Download

Share

COinS