An inertial motion capture framework for constructing body sensor networks
thesisposted on 2023-06-08, 20:28 authored by Tudor Pascu
Motion capture is the process of measuring and subsequently reconstructing the movement of an animated object or being in virtual space. Virtual reconstructions of human motion play an important role in numerous application areas such as animation, medical science, ergonomics, etc. While optical motion capture systems are the industry standard, inertial body sensor networks are becoming viable alternatives due to portability, practicality and cost. This thesis presents an innovative inertial motion capture framework for constructing body sensor networks through software environments, smartphones and web technologies. The first component of the framework is a unique inertial motion capture software environment aimed at providing an improved experimentation environment, accompanied by programming scaffolding and a driver development kit, for users interested in studying or engineering body sensor networks. The software environment provides a bespoke 3D engine for kinematic motion visualisations and a set of tools for hardware integration. The software environment is used to develop the hardware behind a prototype motion capture suit focused on low-power consumption and hardware-centricity. Additional inertial measurement units, which are available commercially, are also integrated to demonstrate the functionality the software environment while providing the framework with additional sources for motion data. The smartphone is the most ubiquitous computing technology and its worldwide uptake has prompted many advances in wearable inertial sensing technologies. Smartphones contain gyroscopes, accelerometers and magnetometers, a combination of sensors that is commonly found in inertial measurement units. This thesis presents a mobile application that investigates whether the smartphone is capable of inertial motion capture by constructing a novel omnidirectional body sensor network. This thesis proposes a novel use for web technologies through the development of the Motion Cloud, a repository and gateway for inertial data. Web technologies have the potential to replace motion capture file formats with online repositories and to set a new standard for how motion data is stored. From a single inertial measurement unit to a more complex body sensor network, the proposed architecture is extendable and facilitates the integration of any inertial hardware configuration. The Motion Cloud’s data can be accessed through an application-programming interface or through a web portal that provides users with the functionality for visualising and exporting the motion data.
- Published version
Department affiliated with
- Informatics Theses
InstitutionUniversity of Sussex
Full text available