Integrated Systems Development: Human Detection and Goods Control with TensorFlow and IoT

This research comprises the implementation methods used to design and develop a human detection system and goods control system. The development of the TensorFlow Machine Learning algorithm for human detection is described in this work. The use of IoT devices, namely ESP32 CAM for data capture, ESP32 for controlling the overall system, and establishing Firebase Database for communication between the TensorFlow development platform, PyCharm, and ESP32 are explained and justified in this paper. The development of the goods control system using ultrasonic sensors and ESP32 as a micro controller, to control the stepper motor, is also explained and justified. Each system was tested individually first before integrating them. Five tests were performed, namely the response time to activate the stepper motor, the human detection accuracy test, the precision of the ultrasonic sensor responsible for height control, the precision of the ultrasonic sensor responsible for motion control, and the stress analysis test of goods lift. The tests present coherent data, but limitations were still found during the testing phase and had to be readjusted before the final integration of both systems.