[Paper Review] How Physicality Enables Trust: A New Era of Trust-Centered Cyberphysical Systems

Stephanie Gil, Michal Yemini+4 more|arXiv (Cornell University)|Nov 13, 2023

Access Control and Trust4 citations

TL;DR

This paper proposes a trust-centered framework for multi-agent cyberphysical systems (CPS) that leverages physical layer characteristics—such as wireless channel behavior, sensing, and contextual awareness—to enable quantifiable, behavior-based trust and resilient coordination. By integrating physicality into trust inference and authentication, the approach moves beyond traditional identity-based security to ensure sustained functionality under adversarial conditions.

ABSTRACT

Multi-agent cyberphysical systems enable new capabilities in efficiency, resilience, and security. The unique characteristics of these systems prompt a reevaluation of their security concepts, including their vulnerabilities, and mechanisms to mitigate these vulnerabilities. This survey paper examines how advancement in wireless networking, coupled with the sensing and computing in cyberphysical systems, can foster novel security capabilities. This study delves into three main themes related to securing multi-agent cyberphysical systems. First, we discuss the threats that are particularly relevant to multi-agent cyberphysical systems given the potential lack of trust between agents. Second, we present prospects for sensing, contextual awareness, and authentication, enabling the inference and measurement of ``inter-agent trust" for these systems. Third, we elaborate on the application of quantifiable trust notions to enable ``resilient coordination," where ``resilient" signifies sustained functionality amid attacks on multiagent cyberphysical systems. We refer to the capability of cyberphysical systems to self-organize, and coordinate to achieve a task as autonomy. This survey unveils the cyberphysical character of future interconnected systems as a pivotal catalyst for realizing robust, trust-centered autonomy in tomorrow's world.

Motivation & Objective

• To address the limitations of traditional identity-based authentication in multi-agent cyberphysical systems, where agent behavior and physical interactions significantly impact trustworthiness.
• To develop a new paradigm for security in distributed, autonomous CPS by grounding trust in measurable physical-layer properties such as channel state, motion dynamics, and environmental sensing.
• To enable resilient coordination in multi-agent systems by integrating quantifiable trust metrics into distributed control and communication algorithms.
• To bridge gaps between communication, control, and robotics communities by establishing a shared nomenclature and framework for physics-based trust in CPS.

Proposed method

• Leverages physical layer characteristics—such as wireless channel state information (CSI), Doppler shifts, and round-trip time (RTT)—to infer inter-agent trust through signal propagation behavior.
• Introduces contextual awareness via onboard sensors (e.g., IMUs, LiDAR, cameras) to monitor agent behavior and environmental interactions, enabling real-time trust assessment.
• Employs AI-powered situational awareness to correlate physical interactions and communication patterns with trust levels, detecting anomalies and adversarial behavior.
• Designs distributed multi-agent algorithms that incorporate trust values as inputs to coordination protocols, ensuring resilience under attacks such as spoofing or jamming.
• Uses signal processing and control-theoretic tools to model trust as a dynamic, quantifiable metric that evolves based on agent actions and network feedback.
• Proposes a closed-loop architecture where application-layer performance and lower-layer physical behavior are tightly coupled to enforce trust and resilience guarantees.

Experimental results

Research questions

• RQ1 How can physical layer characteristics such as wireless channel behavior and motion dynamics be used to infer and measure inter-agent trust in multi-agent cyberphysical systems?
• RQ2 What are the key threats unique to multi-agent CPS that arise from the lack of inherent trust between autonomous agents, especially in the absence of human-in-the-loop supervision?
• RQ3 How can trust be formalized as a quantifiable, dynamic metric that reflects both agent behavior and network interactions, rather than relying solely on identity authentication?
• RQ4 What distributed coordination algorithms can integrate trust metrics to ensure provable resilience under adversarial attacks such as spoofing or denial-of-service?
• RQ5 How can a unified nomenclature and cross-disciplinary framework be established to align communication, control, and robotics communities in advancing trust-centered CPS?

Key findings

• Physical layer signals—such as CSI, RTT, and Doppler shifts—can be used to infer trust between agents without relying on centralized authentication, enabling decentralized trust assessment.
• Contextual awareness from onboard sensors allows real-time monitoring of agent behavior, enabling detection of anomalous actions that may indicate malicious intent.
• AI-driven situational awareness can correlate physical interactions and communication patterns to identify trust violations, such as spoofing or collusion, with high precision.
• Distributed coordination algorithms that incorporate dynamic trust metrics can maintain system functionality under adversarial attacks, achieving provable resilience guarantees.
• The integration of physicality into trust modeling enables a new paradigm of physics-based trust that surpasses classical security models by closing the loop between network behavior and application-layer performance.
• A shared framework across communication, control, and robotics is essential to enable scalable, verifiable, and practical resilience in future cyberphysical systems.