Wireless mesh networking : implementation of internet of things system
Date of Issue2016-11-10
School of Computer Engineering
Parallel and Distributed Computing Centre
Wireless sensor networks are capable of gathering large quantities of distributed multipoint data that could be used to analyse and study the behaviour of our surroundings or environmental changes. However, its capabilities are largely limited by the rationing of power supply when deployed outdoors. Processing and radio transmission power have to be reduced in order to lengthen the lifespan of the sensor nodes which are operating on batteries as much as possible. This poses difficulties with scaling up the wireless sensor network as it would mean more packet transmission hops and wireless radio power required to expand its area of coverage. Mesh network reduces this limitation by expanding the distribution of sensor networks without compromising on interconnectivity between the networks. It is a special variant of an Ad-Hoc network which supports multi-hop transmissions between nodes while unifying its local and remote networks so that clients on any node would be able to communicate with each other. Mesh networking could also further scale up a simple wireless sensor network to a more robust network system by incorporating sensor controllers and triggers that automate actions when a particular sensor data threshold has been satisfied. Such a network can otherwise be dubbed as an Internet of Things. This project aims to implement a fully self-functioning Internet of Things using 4 consumer ASUS routers installed with DD-WRT open source firmware and 4 TelosB sensor nodes. The routers would be running on the Optimised Link State Routing mesh networking protocol while the sensor nodes runs on the Collect Tree Protocol for sending data back to the sink node. The sensor data is then pushed to the router which will forward it to a remote terminal to store into a relational database. This terminal would then analyse the incoming data and determine if certain actions should be triggered. At the same time, the user can also query the database, monitor the state of the sensors from the collected data in real-time, control the sensor behaviours remotely, or even access the Internet simultaneously. While the network proved to be resilient and self-recovering, it is also prone to transmission attacks and routing loops if the mesh network increases in scale in the future. Should resources allow, the Better Approach To Mobile Ad-hoc Networking mesh protocol would be a better alternative that patches possible routing issues for large networks. Additionally, even though basic encryption and authentication measures have been implemented during the development process, security of the network should be constantly reviewed against newly discovered attacks in order to ensure the integrity of the data transmitted across the network.
Final Year Project (FYP)
Nanyang Technological University