AN EXPLAINABLE CYBER-PHYSICAL ARTIFICIAL INTELLIGENCE FRAMEWORK FOR CRASH PREDICTION AND PREVENTION OF HEAVY-DUTY VEHICLES ON HIGH-RISK TRANSPORTATION CORRIDORS IN NIGERIA

Abstract

This study proposes an Explainable Cyber-Physical Artificial Intelligence (CPAI) framework to support secure crash-risk prediction and prevention for heavy-duty vehicles on high-risk transportation corridors in Nigeria. The conceptual architecture integrates vehicle and infrastructure data, protected communication, data-trust assessment, machine-learning prediction, explainability and operator response. It is examined through a qualitative assessment of six threat scenarios involving location spoofing, telemetry injection, replay or impersonation, false-data manipulation, availability attacks and privileged-access abuse. The design identifies preventive, detective and recovery controls across the data-to-decision path, including source validation, message integrity, access control, anomaly monitoring, model protection, audit and controlled degraded operation, while a separate protocol specifies future evaluation of prediction performance, explanation quality and cyber resilience. The framework provides a security-oriented reference architecture for subsequent implementation and testing, but its architectural control coverage does not establish measured prediction accuracy, attack-detection effectiveness or crash reduction.