Το work with title Design and implementation of a solar powered wireless sensor network for autonomous inference by Nichoritis Athanasios is licensed under Creative Commons Attribution 4.0 International
Bibliographic Citation
Athanasios Nichoritis, "Design and implementation of a solar powered wireless sensor network for autonomous inference ", Diploma Work, School of Electrical and Computer Engineering, Technical University of Crete, Chania, Greece, 2021
https://doi.org/10.26233/heallink.tuc.89937
This work studies implementation of an ambiently-powered, distributed wireless sensor network (WSN) capable for in-network inference, using low-cost, low-power embedded nodes. It is found that for 10.8 dBm transmission power, at least 280 meters point-to-point communication link is possible, with packet error rate (PER) less than 2%. Furthermore a 16-node distributed message passing network is presented and compared to a Zigbee centralized network. It is found that the distributed network provides more flexibility and larger communication range, while the centralized network provides faster execution time at the expense of shorter communication range and coverage. Last but not least, Gaussian Belief Propagation and Average Consensus are implemented in the distributed message passing WSN, demonstrating why asynchrony in message passing is the key.