Camera-Based Falling Detection and Human Following Robot

Lee, Chen Ling (2016) Camera-Based Falling Detection and Human Following Robot. Final Year Project (Bachelor), Tunku Abdul Rahman University College.

Text Lee Chen Ling Full Text.pdf Restricted to Registered users only Download (3MB)

Abstract

Nowadays, they are many cases happened about elderly or patient fall in their house while no one are around them. Therefore, to assist the elderly or patient, the aim of my research and build a prototype for Camera-Based Falling Detection and Human Following Robot as my final years’ project. This research deals with the development of variety sensor fixed on the robot’s sensing system, type of material used to build the robot’s chassis, movement and control system which allow the robot to monitoring user’s motion and detecting environment around it by using ultrasonic sensors to detects obstacles and a task of human following. A part from that, the computer USB webcam and PlayStation 3 Eyes camera are used to detect human motion’s condition in this mini project. The computer USB webcam are used to follow human movement, keep update for their current position and detect human movement’s condition for upper view either walk or fall down. Next, the PlayStation 3 Eye camera is uses to detect bottom view of human movement’s condition either fall down or sleep. The extra feature such as alarm, send an emergency message to family member are added in the robot when detect of falling occur. In addition, the Raspberry Pi 2 microcontroller are used as control system which process and make decision from sensor input signal and passes signal data to the output actuator which control the robot movement. The mini project was successful in the sensing and control system that met the specification given. The using of Solidworks software in the design stage is adding of advantages to speedy and easy to build an actual prototype of robot. The use of Solidworks software can be easily modifying and the design of hardware structure can clearly view in 360° of angle.