Design and Construction of a Microcontroller-Based Driver Alcohol Detection System (MDADS)

Abstract
Aims: Vehicle accidents on most highways had caused a lot of losses. Many sustained injuries that marred them and left families helpless. Nigeria highways are not exempted. Drunk-driving increases tendency, severity and causality of crashes. Effects of auto crash damage to lives and properties necessitated the development of the Microcontroller-based Driver Alcohol Detection System (MDADS). Study Design: The system employed ATMega328p microcontroller (CU) which coordinated operations of 7 units that made the MDADS. The units are: Sensor Unit (SU), Switch (S), Power Unit (PU); LCD Indicating Unit (LIU), Alarm Unit (AU), DC motor (Ignition) Unit (IU) and Liquid Crystal Display Unit (LCDU). Place and Duration of Study: The study was conducted for 7 months in the Department of Computer Engineering, Federal Polytechnic Ile-Oluji (FEDPOLEL), Nigeria. It was conducted between October 2020 and July 2021. Methodology: Once the MDADS is ON, it assesses presence of alcohol in the endogenous alcohol molecules from the driver with the help of the SU. The SU sends signal to CU to control and sends signal to trigger the IU, AU and the LCDU of the MDADS, if the Blood Alcohol Content (BAC) exceeds the stipulated threshold 0.29ml/l. 60s tolerance was given to driver to switch OFF the ignition. If driver refuses to comply by switching OFF the ignition, the CU sends a “SWITCH OFF” signal to the IU, the LCDU displays “Drunk” and the buzzer continuously sounds alarm. The designed system was tested and parameters for evaluation were taken. The parameters among other includes True Acceptance Rate (TAR), False Acceptance Rate (FAR), Unable to Accept Rate (UAR) and Detection Accuracy (DA), Results: TAR were 0.81, 0.79, and 0.77 for man, alcoholic drinks and herbal mixture respectively. FAR were 0.03, 0.00, and 0.00 for man, alcoholic drinks and herbal mixture, respectively. For human being, Precision (P) and Recall concept (R) were 0.04 and 0.15 respectively while for P and R for others were negligible. Conclusion: The results reveals that the system can be profitably employed for and improved safety on the highways through precise warning before “switching off” of car engine. A further design should be done to differentiate vividly between drunk drivers and presence of other alcoholic substances such as drugs that contain some alcoholic contents, petrol, methylated spirit and alcoholic drinks.