DESIGN AND DEVELOPMENT OF AN ARTIFICIAL INTELLIGENCE-BASED ALCOHOL DETECTOR SYSTEM FOR VEHICLE ACCIDENT PREVENTION AS A REPLACEMENT FOR THE MANUAL SOUTH-SMELLING TECHNIQUE

Artificial intelligence (AI), sometimes called machine, device or system intelligence, is intelligence demonstrated by machines, in contrast to the natural intelligence displayed by humans and other animals. In other words, it is an area of computer science that emphasizes the creation of intelligent machines that work and react like humans. Until recently, computers could do only what we already programmed them to do. But now, they can learn it themselves and can show us the result that we have never told them before. The alcohol detector system works like human to detect by itself the presence of alcohol in a drunken driver. There was no technology to check the driver either drinks alcohol or not by sensing the alcohol consumption by the driver, which was considered to be the main cause of the accidents on our roads. There was manual checking after particular distance on the roads or the highways but still these checks were not sufficient to stop the happening of the mishaps. Besides, it also involves smelling the mouth of the suspected driver, which is capable of spreading the deadly virus and this can expose our dear law enforcement agents to great risk. Hence, the manual methods must be discouraged as we face the reality of covid19 new normal. So, to avoid these problems, alcohol detector system is developed. This system is developed by integrating the alcohol sensor with the microcontroller 16F877A. The alcohol sensor used in this work is MQ-3 which is to detect the alcohol content in human breath, liquid crystal display (LCD) helps to display information about the device and buzzer sounds to indicate presence of alcohol. When the device remains silent, it means no alcohol but when buzzer sounds, it means there is alcohol detected. The main purpose behind this work is “Drunk driving detection”. Now a day, many accidents are happening because of the alcohol consumption of the driver or the person who is driving the vehicle. The work aims at designing a system that detects alcohol leakage and alerts the subscriber through alarm and status display besides turning off the alcohol supply valve as a primary safety measure, to enable real-time alcohol detection with minimal latency, to provide immediate alerts or notifications when alcohol is detected above predefined thresholds, to seamlessly integrate with advanced sensors (e.g., breathalyzers, infrared, or gas sensors), to collect and analyze data, to support multi-sensor data fusion for improved detection accuracy, incorporate AI algorithms to distinguish between genuine alcohol consumption and incidental exposure (e.g, from hand sanitizers or cleaning agents) and adapt detection sensitivity based on user profiles or situational contexts. The system is an intelligent system, as it does not create nuisance by continuously sounding alarm but the alarm stops beeping once the concentration of the alcohol in the atmosphere after leakage goes below the set point. This work will minimize losses occasioned by explosions, accident while driving due to alcohol leakages and improve safety of life. Alcohol leakage detector system consists of buzzer, which warns by sounding. The suggested system uses sensor technology and machine learning techniques to detect alcohol intake in a non-invasive, accurate, and dependable manner. The system is a reliable and efficient tool for a range of applications, such as law enforcement, healthcare, and transportation, because of its capacity to learn from data and adjust to novel situations. The design and development of the AI-based alcohol detector should serve as a basis for further research and development. For instance, more research and development should be done to further enhance the system's sensitivity, specificity, and accuracy. Additionally, conduct thorough field testing and validation, for instance, to ensure the system's dependability and effectiveness in real-world scenarios.