Development of smart tachograph with a novel algorithm detecting and recognition of driver behaviour

Due to the increase in the number of vehicles on the roads, traffic safety becomes more important. Losses in accidents involving heavy-duty vehicles, where the use of tachograph devices are mandatory, are higher both in terms of death, injury and cost. Although tachographs have made significant progress especially in terms of safety and communication from past to present, they cannot provide any data about inattention, insomnia, fatigue, aggressive driving and to detect and recognise driver behaviors. In this study, a new expert system is proposed. A low-cost 3D gyroscope and an IMU sensor module with a 3D accelerometer are included in the tachograph for detection of lateral and linear maneuvers with their degree of aggressiveness. For highly accurate detection of lateral maneuvers such as right-left turns and lane changes, the edges of events are first captured with gyroscope-Z data, then the last two edges are inspected for lane changes. Edges that do not have the beginning or end of a turning maneuver and that conform to the lane change pattern are recorded as lane changes, others as the start or end of the maneuver. With the new algorithm, for the turning maneuvers GAS magnitude including the gyroscope-Z, accelerometer-X and the speed data, and for the lane changes GSS magnitude including the gyroscope-Z slope and the speed average, are calculated. In practice, it was observed that the algorithm reached 100% accuracy in detecting turns and 76% accuracy in lane changes. The average accuracy of lane changes increases to 85% when the lane changes recorded as turns which have correct scores are included. After the start and end points of longitudinal maneuvers such as braking and acceleration are determined by 3-stage control of the speed data, the score is calculated with the speed change and the maximum-minimum difference of the accelerometer-Y data in the range. All of the braking and acceleration maneuvers in the video recording of the last trip were detected with 100% accuracy by the tachograph. At the end of the journey, the aggressiveness scores of the journey are displayed on the tachograph screen in 3 categories: turning, lane change and linear.