The Smart Automobile (SAM): An Application Based on Drowsiness Detection, Alcohol Detection, Vital Sign Monitoring and Lane based Auto Drive to avoid Accidents

摘要

In today's rapidly developing world where every industry's progress is contributing to adding comfort and safety to human life, the automobile industry is focusing on introducing safe, reliable and autonomous vehicles to save humans from some unbearable losses. Our research aims to serve the purpose of avoiding road accidents caused due to driver's hypo-vigilance, drunk driving or sudden health issues. The proposed system is an SoC (System on Chip) that can be easily installed in the car, it is able to accurately monitor the driver's state while driving and upon observing incapability of driving, the system switches to auto drive mode followed by a safe parking in the left lane. The developed system comprises of four parts. The drowsiness detection sub-system monitors the driver's drowsiness by employing the image processing techniques to detect his face and eyes and determine his drowsy state on the basis of Eye Aspect Ratio (EAR). Alcohol is detected by analyzing the Blood Alcohol Concentration (BAC) levels of driver's continuously, vital signs i.e., heart rate and electrocardiogram (ECG) is measured within the car and if it is observed as being abnormal, the information is coordinated with the driver and with the aid of GSM/GPRS module, information is also forwarded to some concerned family member and emergency helpline. Upon detection of any abnormality by any of the sub-systems, the car switches to auto drive mode. This purpose is served by the auto driving sub- system. Each sub-system developed for the realization of this project is user friendly and easily adaptable within the car. The sub-systems are developed keeping cost, size, reliability and real time response constraints in consideration. Upon testing the system in real world driving scenarios, considerable accuracies are achieved by each sub-system. However, every proposed solution is subjected to some new challenges. The future research may contribute to monitor and avoid more causes of accidents for ensuring human safety and realize a system capable of wide range, safe and reliable autonomous driving.