MUDDLE: Multi-Modal Dynamic Detector Loopback Evaluator to Expose Trojans in Zero-Trust PCB Systems

Prashanth Krishnamurthy, Hammond Pearce, Virinchi Roy Surabhi, Joshua Trujillo, Ramesh Karri, Farshad Khorrami

Research output: Contribution to journalArticlepeer-review

Abstract

We propose the MUDDLE framework for hardware Trojan detection in Printed Circuit Boards (PCBs) in a golden-free and zero-trust setting (unavailable known-good PCB, measurements from processors on the PCB are not trustworthy). Trojans could be inserted by adversaries during/after manufacturing to impact safety, stability, and performance of cyber-physical systems (CPS). MUDDLE monitors side channels under defender-controlled software-driven excitations and I/O connections to create loopbacks. Side channel measurements are acquired by an external test rig that is configurable (using an FPGA). Using FPGA-based reconfigurable logic, the defender instantiates active dynamic loopbacks crafted to illuminate Trojans. Side channel time series are evaluated against models constructed from golden-free design-based characterizations to probabilistically detect Trojans. We demonstrate efficacy of MUDDLE on a reconfigurable PCB Trojan testbed (built atop the OpenPLC programmable logic controller (PLC) platform) combined with an external configurable test rig.

Original languageEnglish (US)
JournalIEEE Micro
DOIs
StateAccepted/In press - 2024

ASJC Scopus subject areas

  • Software
  • Hardware and Architecture
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'MUDDLE: Multi-Modal Dynamic Detector Loopback Evaluator to Expose Trojans in Zero-Trust PCB Systems'. Together they form a unique fingerprint.

Cite this