Abstract
Bluetooth Mesh is proven as a suitable solution for lighting and sensory
networks, where the data packets contain simple status information
and the general message workload is small. However, in other sensory
applications, the messages workload might be much higher, requiring
a better performance ( e.g. real-time monitoring capabilities). When
the sensor data sets exceed the number of bytes that a single message
transmission can carry, the data set must be transmitted into several
messages, and here Bluetooth Mesh standard meet certain challenges.
If any of these messages are lost during the transmission it results in
inconsistent data on the receiver. Thus, the goal for the project will be
to perform research work for evaluating and improving QoS metrics for
large data transfers over Bluetooth mesh network. The experiments will
be conducted in partnership with Nordic Semiconductor, in a desk setup
using nRF52840 development boards configured to emulate real-world
Bluetooth mesh network conditions. Through these experiments, the
study evaluates different chunk sizes and transmission strategies, such
as unicast targeting versus group addressing, under varying network
transmit counts (NTC) and relay retransmit counts (RRC). The results
demonstrate the impacts of these variables on the improvement of transfer
performance for the large amount of data.
The findings from this thesis are serve as historical data to contribute
to the optimization of Bluetooth Mesh networks, facilitating improved
performance in applications beyond traditional sensory networks and
potentially influencing future standards and technological advancements
in wireless communication systems