Machine-Learning Based Channel Quality and Stability Estimation for Stream-Based Multichannel Wireless Sensor Networks
AbstractWireless sensor networks (WSNs) have become more and more diversified and are today able to also support high data rate applications, such as multimedia. In this case, per-packet channel handshaking/switching may result in inducing additional overheads, such as energy consumption, delays and, therefore, data loss. One of the solutions is to perform stream-based channel allocation where channel handshaking is performed once before transmitting the whole data stream. Deciding stream-based channel allocation is more critical in case of multichannel WSNs where channels of different quality/stability are available and the wish for high performance requires sensor nodes to switch to the best among the available channels. In this work, we will focus on devising mechanisms that perform channel quality/stability estimation in order to improve the accommodation of stream-based communication in multichannel wireless sensor networks. For performing channel quality assessment, we have formulated a composite metric, which we call channel rank measurement (CRM), that can demarcate channels into good, intermediate and bad quality on the basis of the standard deviation of the received signal strength indicator (RSSI) and the average of the link quality indicator (LQI) of the received packets. CRM is then used to generate a data set for training a supervised machine learning-based algorithm (which we call Normal Equation based Channel quality prediction (NEC) algorithm) in such a way that it may perform instantaneous channel rank estimation of any channel. Subsequently, two robust extensions of the NEC algorithm are proposed (which we call Normal Equation based Weighted Moving Average Channel quality prediction (NEWMAC) algorithm and Normal Equation based Aggregate Maturity Criteria with Beta Tracking based Channel weight prediction (NEAMCBTC) algorithm), that can perform channel quality estimation on the basis of both current and past values of channel rank estimation. In the end, simulations are made using MATLAB, and the results show that the Extended version of NEAMCBTC algorithm (Ext-NEAMCBTC) outperforms the compared techniques in terms of channel quality and stability assessment. It also minimizes channel switching overheads (in terms of switching delays and energy consumption) for accommodating stream-based communication in multichannel WSNs. View Full-Text
Scifeed alert for new publicationsNever miss any articles matching your research from any publisher
- Get alerts for new papers matching your research
- Find out the new papers from selected authors
- Updated daily for 49'000+ journals and 6000+ publishers
- Define your Scifeed now
Rehan, W.; Fischer, S.; Rehan, M. Machine-Learning Based Channel Quality and Stability Estimation for Stream-Based Multichannel Wireless Sensor Networks. Sensors 2016, 16, 1476.
Rehan W, Fischer S, Rehan M. Machine-Learning Based Channel Quality and Stability Estimation for Stream-Based Multichannel Wireless Sensor Networks. Sensors. 2016; 16(9):1476.Chicago/Turabian Style
Rehan, Waqas; Fischer, Stefan; Rehan, Maaz. 2016. "Machine-Learning Based Channel Quality and Stability Estimation for Stream-Based Multichannel Wireless Sensor Networks." Sensors 16, no. 9: 1476.
Note that from the first issue of 2016, MDPI journals use article numbers instead of page numbers. See further details here.