MUMBAI, India, Nov. 2 -- Intellectual Property India has published a patent application (202641125885 A) filed by Karpagam Academy Of Higher Education; Karpagam Institute Of Technology; Mrs. S. Sumathi; and S. Santhana Raj on October 22, 2026, for Advanced Driving Assistance System Using Deep Learning Techniques.
Inventors include Mrs. S. Sumathi; and S. Santhana Raj.
The application for the patent was published on October 30, 2026, under issue no. 44/2026.
Abstract: Driver drowsiness and fatigue are among the leading causes of road accidents worldwide, significantly compromising public safety and resulting in severe economic and human losses. Prolonged driving durations, irregular work schedules, sleep deprivation, and underlying health conditions often reduce a driver’s alertness, leading to delayed reaction times and impaired decision-making. Conventional safety mechanisms and awareness programs alone are insufficient to address this issue, creating a strong need for intelligent driver monitoring systems capable of detecting fatigue in real time and providing timely intervention. This work presents an Advanced Driver Assistance System (ADAS) based on deep learning techniques for real-time detection of driver drowsiness. The proposed system employs computer vision and convolutional neural networks (CNNs) to analyze visual cues related to driver alertness in a non-intrusive manner. An in-vehicle camera continuously captures facial video data of the driver, which is processed to extract regions of interest, particularly the eyes and facial features, as these physiological indicators are closely associated with fatigue and reduced vigilance. The system incorporates image preprocessing techniques such as face detection, facial landmark localization, and eye-region segmentation to enhance feature extraction accuracy. The extracted features are then fed into a CNN model trained to classify driver states into alert and drowsy categories. Key indicators such as eye closure duration, blinking frequency, yawning behavior, and facial expressions are analyzed to determine the presence and severity of drowsiness. Temporal evaluation across consecutive video frames is performed to ensure reliable detection and to minimize false positives caused by transient facial movements or lighting variations. One of the primary advantages of the proposed approach is its robustness under diverse real-world driving conditions, including variations in illumination, head pose, facial occlusion, and individual driver characteristics. Unlike traditional fatigue detection methods that rely on vehicle dynamics or wearable physiological sensors, the proposed system is cost-effective, comfortable, and suitable for continuous monitoring without causing driver distraction or discomfort. Upon detecting sustained drowsiness beyond a predefined threshold, the system generates an audible alert to immediately notify the driver and prompt corrective action. The system is designed to operate in real time with low computational latency, enabling seamless integration into embedded automotive platforms and modern vehicles equipped with ADAS functionalities. Overall, the proposed deep learning–based driver drowsiness detection system provides an effective and scalable solution to reduce fatigue-related accidents. By combining real-time visual analysis, intelligent classification, and timely alert generation, the system contributes to enhanced road safety and supports the advancement of intelligent transportation systems and next-generation automotive safety technologies.
Disclaimer: Curated by HT Syndication.
