Dernière édition Attention: Le pdf pèse environ 17 - BFH-TI - Berner ...

Master of Science in Biomedical EngineeringBIBEBUVALife saving robot for elderly peopleBiomedical Engineering / Thesis advisor: Prof. Dr. Volker M. KochExpert: Prof. Dr. Elham FirouziElderly people often suffer from isolation and fear to have a serious problem when being alone at home.One hard fall, a hypoglycemic coma or a stroke may cause the person to die because nobody is there tohelp. The goal of this project is to build a prototype robot which can accompany his or her owner and detectphysiological anomaly. If a problem has been detected, the robot can send an alarm to one or more trustedpeople to ask for help. The robot can then be remotely operated from a PC or smartphone such that thealarmed person can check what happened.Project ScopeThe aim of the project is to build amobile platform that can navigateautonomously through an apartment,as well as be remotelysteered over the Internet with awebcam. A web server is used todisplay all medical information ona website and to control the robot.Wireless medical sensors areused to measure physiologicaldata like heart rate, breathing rate,skin temperature and movements.An algorithm was implemented todetect anomalies in these signals.ResultsThe developed robot can do almosteverything that was planned.Its most impressive feature is thatit can stand and drive on twowheels using a self-balancing controlalgorithm. This has beenachieved by using an accelerometerand a gyroscope to know theangle, and by using an incrementalsensor on the wheels to know thespeed of the robot. Once thespeed and angle are known, thepower on the motors is calculatedusing a PID controller.To know about the health of thepatient, a sensor from ZephyrTechnology, which is attached toa chest belt, has to be worn by thepatient. This sensor sends dataabout heart rate, breathing rate,skin temperature and movementsto the robot via Bluetooth. The robotthen monitors the health of thepatient and can send an email topre-defined persons.On the robot’s computer, a Javaapplet has been developed whichcontrols everything: the remotecontrol, the health data, the sendingof emails and the communicationwith the sensor.The video streaming has beendone with an open-source softwarerunning on Linux called FFserver.This is the part that consumesmost processor power andenergy.Discussion and ConclusionA very interesting additional featurewould be to add artificial intelligencesuch that the robot canlearn new skills on its own or byteaching it. The results of this workshow good perspectives for a futurecommercialization of the robot.There is an abundance ofpossible clients and their numberis expected to grow within the nextfew years. To achieve commercialization,an industrial partnershould be found.Lukas FreiRobot without dressPhysiological monitoring sensor from Zephyr Technologyti.bfh.ch43