AGOJO, S. Sensor Based Motion Tracking and Recognition in Martial Arts Training [online]. Brno: Vysoké učení technické v Brně. Fakulta elektrotechniky a komunikačních technologií. 2017.
The aim of the project is the design and development of a system for tracking and recognition of human motion. Student designed acquisition system based on Raspberry Pi 3. He wants to use the Raspberry for signal processing, but student decides to analyze data offline on a computer with Matlab due to some issues with programming language on Raspberry. The student also uses only basic filtration of recorded signal and classification should be described in details. Despite these complaints, the assignment was largely met.
The aim of master thesis of Stephan Agojo is design and build Sensor Based Motion Tracking and Recognition in Martial Arts Training. The theoretical part describes motion capture technology by Vicon and mechanical (non-visual) systems. Describes ways of noise filtering, even though there is no apparent context with in the work assembled software. It also describes the use of the Raspberry Pi3 platform and gyroscopic sensors. The work itself consists in realizing a Raspberry and sensor connection, assembling the software and building a simple program in the Matlab environment used for loading data and basic signal processing. In most of the work, the student uses available demo programs and known procedures. The master‘s thesis does not contain many inventions. Formally I have reservations about the quality of some pictures, the charts are not well readable even in the electronic version of the thesis. References and links are not written according to technical standards. It is also useless to describe how to set the environment and processor programming. I evaluate the thesis by Stephan Agojo with the grade "E”.
eVSKP id 104354