Performance analysis experiments for the wireless sensor networks integrated into the C6 virtual reality environment
Lwakabamba, Bernard (2004) Performance analysis experiments for the wireless sensor networks integrated into the C6 virtual reality environment. Masters thesis, Iowa State University.
Full text available as:
Effective VR systems require a variety of user interaction methods with the environment. Traditional systems used wired devices to implement the interaction devices. This, however, is not possible within the fully enclosed space of the C6, which is a fully enclosed virtual reality (VR) system. For this reason, the interaction devices within the C6 must be wireless devices. The current wireless systems are not able to meet the performance needs of real-time interaction in an immersive environment. Problems with the current system include: instability, interference, difficulty in adding new devices, and unacceptable system latency. To provide a better immersive environment, a robust wireless sensor network (WSN) is necessary to handle interactive devices provided by vendors in the future. The WSN was implemented using off-the-shelf devices developed by Crossbow. The Mica2 and Mica2dot sensor node devices are referenced to as Motes. Optimization methods for the Motes sensor network communication architecture are verified in practice. The individual communication layers of the Motes and C6 VR environment affect the packet delivery performance of the WSNs. Since the integrated sensor nodes are linked in a single hop network, focus was placed on the physical layer and data link layer of the Motes. The metrics used to determine the performance of the physical layer were the packet acceptance rate (PAR), the received signal strength indicator (RSSI) and the communication range. The performance of the MAC sub-layer was determined by looking at its collision avoidance, channel allocation, and collective throughput capabilities. In the error control sub-layer the efficiency of several coding techniques were compared, in Matlab, to see how they would perform in harsh channel conditions. The results show that the Motes WSN have the potential to significantly enhance the user interaction in the C6 VR environment. The size of the Mica2dot makes them ideal for the use in on-body sensor interaction systems. The optimal settings for the Motes would be to set the event sensor nodes to transmit at the maximum available power level. This would decrease the lifetime of the node, but it would improve the packet delivery performance of the network. It also was determined that although it was beneficial to transmit at high packet rates, the workload of the BMAC protocol affected its channel allocation and fairness capabilities. Suggestions to maintain the performance of the MAC protocol, were for developers to interface more sensors on the event sensor nodes, or use more than one channel for each user, when they had to transmit at higher packet rates. Finally, in the error control sub-layer, it was determined that one would need to devise a low complexity scheme that was capable of correcting multiple or burst errors.
Archive Staff Only: edit this record