Analysis of Driving Behaviour through Instrumented Vehicles

India ranks among the countries with the worst road safety record. With over 150,000 road fatalities in the country in 2019, safety is one of the top priorities of the government in the road sector. “Fault of Driver” was reported to be the cause of the accident in about 84% of the cases. This assumes significance in the context of the potential for technology applications in assisting and warning drivers for minimizing driving errors. This paper evaluates driving behavior under various scenarios of road conditions and traffic flow. Essentially, portable sensors were deployed in the vehicles to monitor the vehicle kinematics and assess the drivers' driving behavior. Test runs were conducted in Chennai city in cars mounted with fabricated Inertial Measurement Unit (IMU) interfaced with Real-Time Clock (RTC) and Arduino board along with GPS. Data on vehicle accelerations, decelerations, speeds, and jerk were collected for 44 trips (test runs) in the city. These data were synthesized and utilized to conduct detailed analyses of driving behavior. Acceleration noise, number of accelerations or decelerations per kilometer, speed drops, and jerk characteristics were analyzed in detail. Conclusions were drawn on various driving parameters of the drivers. Regression models relating jerk with various independent variables were built and validated. The drivers were classified into non-aggressive or aggressive driving styles with the help of k-means clustering. This study's findings can serve as useful insights in developing technology solutions in the form of dashboard signs, indicators, messages, warnings, etc. to alert and assist the driver, which will likely result in corrective driving behavior over time. The emergence of connected and autonomous vehicles in developed countries is making headline news these days. This research effort can be viewed as a precursor to such thrust areas, focusing on the safety dimensions.