Advanced Detection of Failed LEDs in a Short Circuit for Automotive Lighting Applications

Area of Science:

• Automotive Engineering
• Electrical Engineering
• Artificial Intelligence

Background:

• Automotive LED lighting systems are crucial for safety and signaling.
• Current diagnostic methods struggle to detect subtle LED short-circuit faults, especially single-LED failures.
• Reliable fault detection is essential for vehicle safety and maintenance.

Purpose of the Study:

• To develop advanced algorithms for accurate LED short-circuit fault detection in automotive systems.
• To overcome the limitations of conventional voltage-drop-based diagnostic methods.
• To enhance the reliability and safety of automotive LED lighting.

Main Methods:

• Proposed two novel, easily implementable algorithms for the vehicle's control unit.
• Developed a mathematical prediction model for fault analysis.
• Utilized a neural network-based approach for fault detection and prediction.

Main Results:

• Achieved a 100% LED short-circuit fault detection rate in most analyzed cases.
• Demonstrated significant improvement over conventional methods, even for single malfunctioning LEDs.
• The neural network model accurately predicted the number of failed LEDs.

Conclusions:

• The proposed algorithms offer a superior solution for automotive LED short-circuit fault detection.
• These methods enhance diagnostic capabilities, particularly for challenging single-LED failure scenarios.
• The neural network approach provides an added benefit of quantifying the extent of LED failures.