Lack of healthcare assistance is one of the issues that European countries such as Sweden face due to the increase of the ageing phenomenon which leads to a higher demand for personnel in hospitals, basic attendance, and housing. Therefore, a tool is clearly required to fulfil these social needs. The thesis focuses on the elderly but especially on those with reduced mobility and develops a wheeled domestic mobile robot whose objective is to deliver their pills at the right time of the day. The main goal is the implementation of automatic drug delivery. The project presents a selection of the most suitable wheel-based robot, and a study and evaluation of different techniques and algorithms used for indoor positioning, navigation, facial recognition, and a communication interface between the patient and the robot.
The robot used for the development of this thesis was the TurtleBot3 Burger and was evaluated upon a map created using the Hector SLAM method. Results showed that the Bluetooth technology (iBeacon), as well as the trilateration algorithm, are suitable choices for detecting a human in an indoor environment; a successful average drift error of 0.83 metres was obtained for indoor localization, and further results showed the facial recognition system achieved an accuracy of around 90%. It is concluded that the robot is capable of reaching and identifying the patient in an indoor environment, and so this project can be considered as the first step of implementation of a totally autonomous medical robot ready for domestic use.