The rapid growth of the automotive industry has led to a significant increase in the number of vehicles on the road, resulting in higher demand for efficient vehicle diagnostics and maintenance. Conventional diagnostic approaches require vehicles to be transported to service centers, leading to increased downtime, operational costs, and dependency on specialized diagnostic equipment.
This white paper presents the design and development of an On-Screen Intelligent Maintenance System (OSIMS) for automotive vehicles. The proposed system enables real-time monitoring, diagnostics, and predictive maintenance directly within the vehicle. By leveraging CAN (Controller Area Network) communication, Non-Volatile Memory (NVM), and machine learning-based analytics, the system provides actionable fault detection and resolution guidance to vehicle operators through an integrated on-screen interface. The solution enhances vehicle reliability, reduces maintenance costs, and empowers drivers with self-diagnostic capabilities.
1. Introduction
Modern vehicles consist of multiple Electronic Control Units (ECUs) responsible for managing engine performance, safety systems, comfort features, and driver interfaces. As vehicle complexity increases, traditional off-board diagnostics become inefficient, costly, and time-consuming.
The proposed On-Screen Intelligent Maintenance System addresses these challenges by embedding diagnostic intelligence directly into the vehicle. The system continuously monitors critical vehicle parameters, detects anomalies at an early stage, and provides guided maintenance instructions in textual and video formats via an on-board display.
This approach minimizes vehicle downtime, enables proactive maintenance, and reduces reliance on external diagnostic centers.
2. System Objectives
The key objectives of the on-screen maintenance system are:
- Provide real-time vehicle health monitoring
- Enable early fault detection and predictive maintenance
- Reduce service center dependency and operational costs
- Improve vehicle availability and safety
- Enhance driver awareness and decision-making
- Offer intuitive, multimedia-based repair guidance
3. System Architecture Overview
The proposed system follows a layered and modular architecture, ensuring scalability, reliability, and ease of integration with existing automotive platforms.
3.1 High-Level Architecture Components
- Vehicle Sensors and ECUs
- CAN Bus Communication Layer
- Data Acquisition and Storage (NVM)
- Data Processing and Analytics Engine
- Machine Learning-Based Fault Prediction Module
- On-Screen Human-Machine Interface (HMI)
4. Detailed System Design
4.1 Vehicle Sensors and ECUs
The vehicle is equipped with various sensors connected to multiple ECUs that monitor critical parameters, including:
- Vehicle speed and acceleration
- Fuel level and fuel consumption
- Engine temperature and pressure
- Tire pressure monitoring system (TPMS)
- Internal cabin temperature
- Battery voltage and charging status
- Driver input and vehicle interface data
Each ECU collects sensor data and transmits it over the CAN bus.
4.2 CAN Communication Layer
The Controller Area Network (CAN) protocol serves as the backbone of communication between ECUs. It provides:
- High-speed and reliable data transfer
- Efficient handling of multiple parameters
- Robust error detection and fault tolerance
- Real-time data exchange
All sensor data and diagnostic messages are transmitted through the CAN bus to the central processing unit of the on-screen maintenance system.
4.3 Data Acquisition and Non-Volatile Memory (NVM)
Raw data received via the CAN bus is:
- Filtered and timestamped
- Stored in Non-Volatile Memory (NVM) for persistence
- Used for both real-time diagnostics and historical analysis
The use of NVM ensures:
- Data retention even after power cycles
- Support for trend analysis and fault history tracking
- Improved reliability of predictive maintenance models
4.4 Data Processing and Analytics Engine
The analytics engine continuously decodes raw CAN data and converts it into meaningful diagnostic parameters. Key functions include:
- Signal decoding and normalization
- Threshold-based fault detection
- Pattern recognition for abnormal behavior
- Correlation across multiple subsystems
This engine acts as a bridge between raw vehicle data and intelligent decision-making.
4.5 Machine Learning-Based Fault Prediction
To enhance diagnostic capabilities, the system integrates machine learning algorithms that analyze both real-time and historical data stored in NVM.
Key Capabilities
- Early detection of system degradation
- Prediction of potential failures before critical breakdown
- Continuous learning from vehicle usage patterns
- Reduction of false positives compared to rule-based diagnostics
By identifying anomalies in an early phase, the system enables preventive maintenance and avoids major failures.
5. On-Screen Maintenance Interface (HMI Design)
The Human-Machine Interface is designed to be intuitive, responsive, and driver-friendly.
5.1 Fault Notification Workflow
- A fault or anomaly is detected by the analytics engine.
- The system decodes the faulty data.
- A visual and/or audible warning is displayed on the screen.
- The fault is categorized by severity (informational, warning, critical).
5.2 Interactive Fault Resolution
When the operator selects a fault from the screen:
- The system displays:
- Fault description
- Possible causes
- Severity level
- Step-by-step solutions are presented in:
- Text format for quick reference
- Video format for detailed guidance
These instructions guide the operator through basic troubleshooting or corrective actions, enabling safe continuation of the journey whenever possible.
6. End-to-End System Workflow
- Sensors collect real-time vehicle data.
- ECUs transmit data via CAN bus.
- Raw data is stored in NVM.
- Analytics engine decodes and processes data.
- Machine learning module predicts faults.
- Detected issues trigger alerts on the screen.
- Operator views causes and solutions.
- Fault history is updated for future analysis.
7. Benefits of the Proposed System
- Reduced Vehicle Downtime
- Lower Maintenance Costs
- Early Fault Detection
- Improved Safety and Reliability
- User Empowerment and Convenience
- Scalable and Future-Ready Architecture
8. Use Case Scenario
A vehicle experiences gradual tire pressure loss. The system detects abnormal pressure trends over time using stored NVM data and machine learning analysis. Before reaching a critical level, the system alerts the driver with a warning and displays a video explaining how to inspect and inflate the tire. The driver resolves the issue without visiting a service center, saving time and cost.
9. Conclusion
The On-Screen Intelligent Maintenance System represents a significant advancement in automotive diagnostics and maintenance. By combining CAN-based data acquisition, persistent storage, machine learning analytics, and an intuitive on-screen interface, the system enables proactive, cost-effective, and user-friendly vehicle maintenance.
This approach not only reduces dependency on traditional diagnostic centers but also enhances vehicle safety, reliability, and customer satisfaction. The proposed design lays a strong foundation for future integration with connected vehicle ecosystems and over-the-air updates.