Dynamic timeslot allocation technique for wireless sensor network

• Main Author: Erixno, Oon
• Format: Thesis
• Language: English
• Published: 2013
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://eprints.utm.my/id/eprint/36766/5/OonErixnoMFSKSM2013.pdf; http://eprints.utm.my/id/eprint/36766/

Wireless Sensor Network (WSN) technology has been implemented in many data collection applications. Typically, these applications have a multi to one uplink traffic over multi-hop transmission in high traffic and these characteristics will introduce funnelling in WSN. Fixed traffic scheduling has been used to minimize the funnelling in these applications. However, this scheduling cannot adapt to the variation of traffic intensity and changes of topology. This research proposed a dynamic traffic scheduling using a technique named Dynamic Time slot Allocation (DTsA). The technique provided a mechanism for setting up and updating traffic scheduling dynamically according to the change of traffic intensity and topology. To evaluate the performance of DTsA, extensive experiments were performed to measure and evaluate throughput and packet loss over a variety of traffic intensities. The performance is benchmarked against MT-XLP, a cross layer protocol that applies fixed traffic scheduling. The results showed that the throughput of DTsA outperformed the throughput of MT-XLP for low and high level funnelling topologies. In the low level funnelling topology, the DTsA achieved a maximum throughput of 113 pps whereas the MT-XLP only reached 73 pps. In the high level funnelling topology, the maximum throughput of DTsA was 107 pps as compared to a drop to 28 pps for MT-XLP. These results showed that DTsA performed better than MT-XLP during the adaptation to the dynamic changes of topology. The technique maintained its throughput ratio even though there were some nodes that were disconnected and reconnected to the network. This connectivity issue cannot be handled by fixed traffic scheduling. On the contrary, the DTsA packet losses were a maximum of 2.43% in the low level funnelling topology and 4.06% in the high level funnelling topology in comparison to MT-XLP which had 2.44 % packet loss in the low level funnelling and 2.97 % in the high level funnelling. From the analysis, the high packet loss in DTsA is due to the issue of imprecision of timing for synchronizing communication between nodes. Due to this characteristic, DTsA is more suitable for connectionless application of data collection such as streaming and broadcasting.