A wireless body area network (WBAN) is a collection of miniaturized and energy efficient wireless sensor nodes which monitor human body functions and its surroundings. It has been observed that WBANs perform single application per network, computation and storage capacities are scarce and there is no or limited mobility support. Technically complex WBAN application solutions today, find refuge in processing computationally complex data external to WBANs, i.e., processing sensor data on a conventional PC which is impractical and clumsy. There is a strong need for WBAN platforms which can perform computationally complex tasks on their own having enough resources in terms of computation and memory but still consuming as low power as possible in order to prolong network uptime.
In this thesis work, an improved WBAN named multipurpose-BodyNet (MPBodyNet) is implemented. It has enough computational and memory resources and compact software solutions to achieve high performance and fidelity. MPBodyNet is a self-configuring, multipurpose WBAN which can perform multiple applications and user can switch between applications by a mere push of button. It supports mobility and it acts like an agent network to other networks. MP-BodyNet forms a hierarchy where low-capability networks are supported by higher-capacity networks.
Hardware used for MP-BodyNet has been designed by WSN-Team at Centre for Wireless Communications, University of Oulu and this thesis proposes two application scenarios. Senior citizen protection mode (SPM) deals with a very hot health care issue for elderly people and patients. An algorithm is proposed and implemented that can detect falls or if the subject/patient has fainted. In SPM, MP-BodyNet can generate alarms in case of emergency and events can be seen on a central server as well as a special alarm is generated to the user’s phone (android app.) which can in turn establish an emergency call automatically. Algorithmic efficiency achieved is 100%.
Silent communication mode (SCM) deals with a military hand signal/gesture recognition application. A quite complex pattern recognition algorithm has been proposed with two novelties in it i.e., a sampling process is introduced in the algorithm and the whole algorithmic processing is supposed to be done on the sensor node itself, no processing is supposed to be happening external to the WBAN. Algorithm for SCM is only presented here conceptually after rigorous research about the subject at disposal. It is not implemented in this thesis due to lack of time and is saved for future development. After a gesture would be recognized, an audio message mapped to the gesture will be heard over a headphone.
| Identifer | oai:union.ndltd.org:oulo.fi/oai:oulu.fi:nbnfioulu-201306061569 |
| Date | 14 June 2013 |
| Creators | Virk, M. (Muhammad) |
| Publisher | University of Oulu |
| Source Sets | University of Oulu |
| Language | English |
| Detected Language | English |
| Type | info:eu-repo/semantics/masterThesis, info:eu-repo/semantics/publishedVersion |
| Format | application/pdf |
| Rights | info:eu-repo/semantics/openAccess, © Muhammad Virk, 2013 |
Page generated in 0.0022 seconds