An Extended CIA-Based Multi-Level Model for AHP-Driven Safety and Security Decision-Making in Last-Mile Robotic Systems

The rapid growth of e-commerce and the increasing demand for efficient last-mile logistics have led to the rising use of last-mile robots. While these robots promise faster and cheaper deliveries, their operation in complex and dynamic urban environments introduces significant safety and security challenges. Sensor failures, communication disruptions, and cyber-physical attacks may affect the behaviour of the robots and influence human safety.

This work models and analyses these challenges using the Extended Multi-Level Model to represent the different components of last-mile robotic systems and their influence on the environment. We apply Multi-Criteria Decision Making (MCDM) for parallel safety and security risk assessment, focusing on the confidentiality, integrity, and availability (CIA) of the last-mile systems. Considering these three properties together allows priorities to be set within the CIA triad, which is essential for financial and economic decision-making when only limited resources for countermeasures are available.

We extend the model to an Extended CIA Multi-Level Model that enables detailed evaluation of safety and security risks across all system levels. A case study involving robots transporting critical parcel contents demonstrates how confidentiality, integrity, and availability concerns arise throughout the model and how their violation may affect human safety. The approach supports structured decision-making and contributes to improving the safe and secure deployment of last-mile robots.