Transportation is a component of supporting human mobility to move from one place to another quickly. Along with the times, the technology embedded in cars is getting more advanced, such as technology for security systems and car control systems. This study designs an autonomous car prototype by implementing a lane guard system, object detection system, and automatic braking, thereby reducing the risk of road accidents. The method used is deep learning with a pretrained ssd-mobilenet-v2 model and jetson-inference library on an object detection system with 100% accuracy on testing 3 images of people and 3 images of cars. The use of pure pursuit control that brings the car to maneuver on the track well in 3 experimental scenarios, namely straight, turn right, and turn left. The average steering angle on the straight line is 90,44º, right-turning lane is 65,4º, and on the left-turning lane is 113,1º. In the automatic braking system, the car can estimate the distance with the largest error of 0,02 m and the smallest 0 m. The car stops at an average distance of 0,48 m with the highest distance value of 0,57 m, and the lowest value of 0,40 m with an average car speed of 0,399125 m/s.
Keywords: Lane guard system, object detection system, automatic braking system, pure pursuit control