Integrating Augmented Reality and Artificial Intelligence in Vehicle Diagnostics: Applications for On-Board Diagnostics II Systems

1. Introduction

The use of electronics and sensors in vehicles has become critical component in performance, safety and emissions tracking, with On-Board Diagnostics II (OBD-II) being central. Since the mid-1990s, OBD-II has been standardised in the automotive industry and has the ability to instantly track data relative to fuel consumption, emissions, and the state of the engine - These data streams can be accessed via scanners that relay diagnostic trouble (DTC) and sensor data in real time. Unlike the older systems, OBD-II has the ability to discern major faults, but remains deficient in user-friendliness and accessibility (Michailidis, 2025; Singh, 2021). The streams of data coming from OBD-II are primarily in the form of numbers and letters which greatly complicates the understanding of the data for the layman. Traditional systems worked within the confines of metrics that were set in stone and did not take user specifications into thoughtful consideration, thereby decreasing the system in question’s ability to carry out preventative maintenance, and increasing idle time (Visconti, 2025; Hossain, 2024). 

The application of new principles such as Artificial Intelligence (AI) and Augmented Reality (AR) can be extremely beneficial for the issues presented. AI, on the side of the OBD-II and the attendant sensor streams, along with data from the other components associated with that then uses the techniques of machine and deep learning to detect patterns, forecast issues, and fine-tune schedules for maintenance (Michailidis, 2025). AR Systems can also improve the user experience by focusing the camera on a vehicle and then displaying information about the vehicle as well as being able to guide the user as to which component is faulty along with system visualisation (Patel & Gaikwad, 2020).

This review investigates AR-AI integration to augment OBD-II diagnostics by transforming raw data into actionable and visually actionable data and analyzes IoT’s capability for real-time data exchange between vehicles, diagnostic tools, and cloud-based AI (Maalik & Pirapuraj, 2021). This review discusses the current applications, research shortages, and future directions for intelligent, easy to use, and proactive vehicle maintenance.

2. Body

2.1   On-Board Diagnostics-II Overview

2.1.1 Evolution of OBD systems

The journey of On-Board Diagnostics (OBD) systems can be traced back to the late 1960s and early 1970s, when automakers first began embedding electronic controls in vehicles to regulate engine performance and emissions (He, 2019). The increasing adoption of electronic fuel injection and catalytic converters highlighted the need for monitoring and control mechanisms (Brown & Taylor, 2020). The first official step was the introduction of OBD-I in California in 1991, mandated by the California Air Resources Board (CARB). OBD-I provided the ability to store basic diagnostic codes and trigger the familiar “Check Engine” light; however, the system lacked standardisation, as diagnostic trouble codes (DTCs) and connectors varied widely between manufacturers, limiting its applicability across the industry (Ehsani, 2018).

To prevent these issues from occurring, OBD-II was developed in 1996 and buffed out and replaced all issues lacking from its predecessor. It standardizes, and sets, a universal system for diagnostic connectors, communication techniques, and trouble codes, allowing independent mechanics, shops, and even vehicle owners to use automated diagnostics without needing manufacturer-specific tools (Miller & Smith, 2021). OBD-II has somewhat adjusted its techniques, using CAN (Controller Area Network) for instant monitoring of parameters such as fuel trims, air-fuel ratio, ignition timing, and first-stage sensors. OBD-II is globally accepted, and necessary, in, North America, Europe, and the explicitly growing Asia. This showcases the primary universal diagnostic backbone. This improvement depicts the ongoing trend of cars being “computers on wheels”. This is when the sensors, ECUs, and standardized interfaces on the vehicle all support the intelligent supervision and control of the car.

Figure 1: Timeline of On-Board Diagnostics (OBD) System Evolution

2.1.2 Current role in diagnostics

In its current form, OBD-II serves as the primary interface between the vehicle’s internal systems and external diagnostic tools. The system continuously monitors essential functions, including emissions control, fuel efficiency, and overall engine performance. When anomalies are detected, the system logs a diagnostic trouble code, which can be retrieved using a scan tool or smartphone-based OBD readers via Bluetooth or Wi-Fi. For example, a P0300 code indicates random cylinder misfires, alerting the technician to investigate spark plugs, ignition coils, or fuel injectors (Michailidis, 2025; Singh, 2021).

Beyond emissions compliance, OBD-II contributes significantly to preventive and predictive maintenance. Modern vehicles use OBD-II to track the performance of critical components such as oxygen sensors, mass airflow sensors, and catalytic converters. Fleet managers, in particular, rely on OBD-II data to optimise fuel economy, reduce downtime, and lower operational costs by addressing issues before they escalate into catastrophic events (Maalik & Pirapuraj, 2021). With the rise of telematics and IoT platforms, OBD-II data can now be transmitted in real-time to centralised systems for large-scale monitoring, supporting applications such as usage-based insurance, driver behaviour analysis, and smart fleet management (Patel & Gaikwad, 2020).

In addition, OBD-II has become more consumer-friendly. Affordable OBD-II dongles connected to mobile apps provide everyday drivers with access to live vehicle performance metrics, trip data, and even fuel efficiency reports. This democratisation of diagnostics reflects the broader trend toward data-driven decision-making in personal vehicle ownership (Visconti, 2025).

2.1.3 Challenges (raw data, lack of visualisation)

Despite its broad utility, OBD-II still faces significant limitations in data usability and interpretability. The information retrieved from an OBD-II port is highly technical—primarily a list of fault codes and raw sensor values. While expert technicians can decode these numbers into meaningful insights, non-specialists often struggle. For example, a code like P0420 (“Catalyst System Efficiency Below Threshold”) only indicates a general issue with the catalytic converter but does not specify whether the cause is a faulty oxygen sensor, a clogged converter, or a wiring issue (Michailidis, 2025; Singh, 2021). This lack of specificity can result in diagnostic ambiguity, requiring additional manual troubleshooting and extending repair times.

Moreover, OBD-II is inherently a text-based system. It does not provide built-in visualisation capabilities, meaning that parameters such as fuel trim fluctuations, oxygen sensor waveforms, or coolant temperature variations are displayed as raw numeric data rather than user-friendly graphs or dashboards. This hinders the ability to spot trends or anomalies in real time (Visconti, 2025). For fleet operators and technicians working with large datasets, the lack of visualisation tools makes analysis cumbersome.Understanding OBD systems requires more than just identifying an issue after it happens, and OBD-II systems’ limitations prevent predictive analyses and actions from being taken on issues that arise, and the inability to extrapolate and integrate data across multiple models/brands adds to the problem (Patel & Gaikwad, 2020). Artificial intelligence, as well as machine learning which helps make real-time predictive analyses, coupled with augmented reality systems which provide contextually adaptive interface aids, and support through IoT connected devices, can reinforce OBD-II systems like advanced predictive and diagnostic support systems (Maalik & Pirapuraj, 2021).

Table 1: Positioning OBD-II, AI, AR, and IoT in Automotive Diagnostics

2.1   Augmented Reality in Automotive

2.2.1 Applications in maintenance and training.

AR has been utilised in diverse ways in car repairs. In garages, mechanics are able to see the precise location of malfunctioning parts using AR-capable headsets or tablets. For example, when a malfunctioning oxygen sensor is fitted, AR can display its precise position, give real-time sensor readings, and guide the mechanic through replacement steps by step (Maalik & Pirapuraj, 2021; Patel & Gaikwad, 2020). Advanced AR platforms can even simulate virtual overlays for wiring diagrams, torque specifications, and assembly instructions, reducing errors because of misreading of traditional manuals (Azuma, 2018).

In training and education, AR enables students and young engineers to interact with high-fidelity 3D models of an engine, transmission, brake system, or electronic control unit. Parts can be disassembled, repaired, and reassembled virtually by trainees, which minimises the need for real parts and lowers operational costs. Training with AR can also provide access to rare or complex faults that are not always easily identified in physical cars, thereby maximising learning effectiveness and memory retention.

Furthermore, combining AR with gamification methods—i.e., with tracking progress, scoring, or instant feedback—can promote engagement and accelerate the development of skills more efficiently (Zhou et al., 2021). AR is utilised in remote support software, where professionals can see a mechanic's surroundings and provide live instructions. It is particularly helpful for geographically spread or on-location teams as it facilitates expertise being offered without necessarily having to physically travel (Kaleem et al., 2022).

2.2.2 AR advantages for mechanics and drivers.

AR significantly accelerates diagnostic speed and accuracy for technicians. Through the superimposition of OBD-II fault codes onto actual parts, AR eliminates the need for manual code-to-actual-part correlation. Techs have live sensor readings, temperature, or voltage fluctuations in real time, allowing more efficient fault finding and reduced repair times (Patel & Gaikwad, 2020; Maalik & Pirapuraj, 2021).

For drivers, AR provides enhanced security and operational awareness. AR-enabled head-up displays (HUDs) can project predictive maintenance alerts, fuel consumption information, or potential hazards directly onto the windshield, diminishing distraction and keeping the driver road-conscious. On connected vehicles, AR may be combined with IoT platforms to show car health notices remotely on mobile phones or AR spectacles, yielding an anticipatory car maintenance approach (Visconti, 2025). In addition, AR allows for collaborative diagnosis.

It enables multiple users, such as mechanics, engineers, and trainees, to have the same AR overlay simultaneously, allowing them to make decisions and analyse together. The feature comes in handy for fleet operations, complex repair, and training (Kaleem et al., 2022). Overall, AR turns car repair into a data-driven, visually augmented, and collaborative process instead of a human, experience-based one, reducing errors, costs, and downtime and improving overall safety and efficiency (Azuma, 2018).

2.2   Artificial Intelligence in Vehicle Diagnostics

Artificial Intelligence (AIt in the form of Machine Learning (ML) and Deep Learning (DL) has pushed the boundaries of OBD II diagnostic software and its functionality well beyond the fault code listing that was previously the standard. Algorithms developed using ML, such as SVM, Random Forests, and versatile Neural Networks, diagnose faults within vehicles and enhance efficacy (Zhang et al. 2023; Michailidis et al. 2025). DL frameworks such as CNNs and RNNs are able to pinpoint time-series sensor data and catch anomalies during interval periods (Hussain et al. 2022). Predictive maintenance also uses AIl to help estimate the repair time, thus minimizing the cost, and estimating the component’s lifespan (Lee et al. 2024; Mahale et al. 2025). AIl’s exceptional skill of pattern recognition (Lee & Park, 2025; Verma & Kumar, 2024) gives OBD II the ability to be proactive in the realm of Decision Support Systems.

2.3   Combining AR and AI for OBD-II Visualisation

The combination of Artificial Intelligence (AI) and Augmented Reality (AR) is transforming OBD-II automobile diagnostics. While AI forecasts and announces sensor failure and sensor input and proactively streams them, AR provides real-time, multi-dimensional, context-aware presentation of the issues and solutions (Verma & Kumar, 2024; Hussain et al., 2022). AR-AI medicine and flight, for example, illustrate how the collaboration reduces human mistakes, enhances training, and enables efficient maintenance (Azuma, 2018; Kaleem et al., 2022). As a whole, the integration of Augmented Reality and Artificial Intelligence (AI) transforms data achieved from OBD-II ports and makes them intelligent, animated, usable data, thus maximizing the accuracy, working speed, and dependability of car diagnostics (Chen et al., 2024; Michailidis et al., 2025).

2.4   IoT as a Supporting Technology

Table 2: Simulation-Based Vehicle Diagnostics – Literature Insights and Gaps

The Combination of IoT, AI, and AR makes it possible to future-proof vehicle diagnostics using OBD-II. IoT retrieves real-time sensor data from vehicles, while AI translates the same into fault prediction, anomaly, and maintenance scheduling (Zhou et al., 2021; Verma & Kumar, 2024). AR projects actionable information onto vehicles with the ability of speedy repair and distant guidance (Kaleem et al., 2022; Patel & Gaikwad, 2020). Cloud infrastructure makes possible end-to-end scaling, collective diagnostics, and pervasive learning of AI (Chen et al., 2024; Mahale et al., 2025). The three unities convert the conventional reactive diagnostics into interactive, proactive, and smart ones with end-to-end fleet efficiency, safety, and vehicular reliability (Lee et al., 2025; Michailidis et al., 2025).

3. Challenges and Research Gaps

Despite these technological advancements, the convergence of OBD-II and AI with augmented reality remains challenged by the absence of standard PIDs, real-time data processing constraints, ui/ux challenges, and system security. However, there are no universal diagnostics for proprietary PIDs (Verma & Kumar, 2024; Michailidis et al., 2025). Latency and inadequate edge power limit the real-time intelligent AR processing (Lee et al., 2025; Chen et al., 2024). Poorly implemented augmented reality systems can overwhelm the user with information (Billinghurst et al., 2020). The security risks of IoT connectivity can bring about privacy risks (Hussain et al., 2021; Li et al., 2024). There is lack of research work that combine AR, AI and OBD-II as one unit and this demands standardization and secured the frame works and usability driven research (Patel & Gaikwad, 2020; Mahale et al., 2025).

4. Future Directions

There is also a need of standard OBD-II visualization framework for future vehicle diagnostics to provide the ability of interoperability and cross platform tools (Verma & Kumar, 2024; Li et al., 2024). Further, AR enhanced with AI can offer context-aware overlaying, to assist user decisions and interaction (Mahale et al., 2025; Sharma & Singh, 2022). Secured, standardised data protocols in the context of predictive maintenance are also needed for autonomous and connected vehicles (Zhang et al., 2023; Hussain et al., 2021). Edge computing can reduce the delay by allowing real time processing at the edge, which is vital for the AR-AI-OBD-II integration (Shi et al., 2022; Patel et al., 2023). The industry acceptance depends on the cybersecurity, technician education, and cooperative standardisation (Gupta & Jain, 2021; Kumar et al., 2023).

5. Conclusion

By combining AR, AI and IoT, the future of vehicle diagnostics can overhaul trajectory by converting the traditional means of fault detection into an intelligent, proactive and connected process. AR improves visualization by superimposing diagnostic data onto vehicle parts, thereby minimizing error and downtime (Gupta & Jain, 2021; Liang & Wang, 2022). AI allows predictive analysis, thus helping in early failure detection and predictive diagnostics of vehicles in terms of OBD-II data analysis in real-time (Xu et al., 2023; Mahale et al., 2025). With IoT technology, the interconnection can be maintained perfectly, which enables the remote diagnosis, the fleet management, and the OTA(Over-the-Air) (Rathore et al., 2023; Zhang et al). With progress not going beyond the prototype stage, they are hampered by latency, a lack of standardisation, and security vulnerabilities (Abreu et al., 2024; Pourrahmani et al., 2022). Further research is required for validation in real-world scenarios, and to develop scalable-edge-computing frameworks, adaptive AR interfaces and secure IoT data exchange (Hossain, 2025; Patel & Gaikwad, 2020). Collaborations with industry and pilot tests are going to be indispensable for the success of intelligent diagnosis systems that are efficient, user-friendly, widespread (Saki et al., 2015; Capacho et al., 2015).

Author Contributions: All authors contributed to this research.

Funding: Not applicable.

Conflicts of Interest: The authors declare no conflict of interest.

Informed Consent Statement/Ethics approval: Not applicable.

Declaration of Generative AI and AI-assisted Technologies: This study has not used any generative AI tools or technologies in the preparation of this manuscript.