Integration of Artificial Intelligence Algorithms into On-Board Vehicle Diagnostic and Control Systems

( Vol-12,Issue-2,March - April 2026 )

Author(s): Madugula Abhiram

Download Full Text PDF
Total View : 104
Downloads : 1
Page No: 128-134
DOI: 10.22161/ijaems.122.15

Keywords:

on-board diagnostics, OBD-II, CAN bus, ECU, fault diagnosis, predictive maintenance, intrusion detection, AUTOSAR Adaptive, functional safety, edge–cloud deployment.

Abstract:

The article examines architectural and engineering approaches for integrating artificial intelligence algorithms into on-board vehicle diagnostic and control systems. Relevance follows from software-defined vehicles, Electronic Control Unit (ECU) density, and the need to convert On-Board Diagnostics II (OBD-II) and Controller Area Network (CAN) telemetry into dependable decisions under real-time and safety constraints. Novelty consists of a synthesis that links data acquisition and preprocessing, model families for fault prediction, multi-fault diagnosis, and intrusion detection, and deployment patterns spanning in-vehicle compute with cloud/edge services. The objective is to systematize design decisions that yield robust diagnostics, interpretable outputs for service workflows, and controlled mitigation actions. Methods comprise comparative analysis of recent publications, taxonomy building across task types, and architectural decomposition of pipelines. The source base covers OBD-II machine learning, deep-learning predictive maintenance, attention-based fault prediction, explainable multi-fault diagnosis, CAN intrusion detection, connected-vehicle platforms, deterministic AUTOSAR Adaptive communication, and ML-aware functional-safety extensions. The article targets automotive software engineers, data scientists, and platform architects.

Received: 20 Feb 2026; Received in revised form: 22 Mar 2026; Accepted: 26 Mar 2026; Available online: 01 Apr 2026

Cite This Article:

Share: