Design and Implementation of Intelligent Medicine Delivery Car Based on K210

Abstract

The technology of self-driving automobiles has developed rapidly. Line tracking and path planning are essential to driverless vehicles. At the same time, intelligent robots are widely used, especially for hazardous and repetitive work operations. This paper introduces the design and realization of an intelligent medicine delivery car, which simulates the delivery and pickup of medicine between the pharmacy and wards. A K210 development board serves as the car system's control center. The pressure sensor RP-S40-ST detects whether the car is loaded with medicine. Aiming to control the motion state of the car, the PID incremental algorithm and L298N are used, while the STM32F103C8T6 with a 7-road grayscale sensor module achieves automatic road tracking. The OV5640 camera in the design collects images to decide the target ward's location. By applying the YOLO v2 training AI model, the car is able to identify numbers and acquire information on identification probability and number position. The intelligent drug delivery car is capable of loading detection, number recognition, automatic tracking, path-selecting, and transporting medicine to the designated ward and back to the pharmacy. The time for identifying ward numbers is within 3.01 seconds.