Work in progress seminar by Tim Millward, discussing IoT networks, AI and fog computing for smart agriculture
26 February 2019
Tim Millward presented his seminar with us on the 22nd February. He began by discussing the current pestilence problem with fruit flies affecting farm land, produce and how IoT and A.I based devices could be the solution. Continents such as Asia, Africa and South America were shown to be most affected, as shown by graphs in his introduction. Following this, he then explains that current machine 2 machine (M2M) technology demonstrates that there are data transfer issues and delays in modern A.I devices. Tim reiterates that this would be the foundation for his work to build upon and by performing deep learning algorithms through fog computing, he is able to create a solution that hasn’t been attempted before.
For smart agriculture, his plan is to use drones to upload data autonomously to the cloud, the algorithm learns by itself and deploy the solution to other drones. This would be a reoccurring cycle of learning for the smart drone and would minimise the issues which currently effect M2M technology. The drones would learn from the data installed onto the chip and would follow the pattern created by the data.
Tim explains how the Nvidia Jetson provides a gateway but would not be suitable for implementation as this would only perform one task at a time. The solution to this would be to upgrade to the Nvidia Jetson Xavier, this would be able to run parallel to the drone camera, would perform inference, training, implementation and a gateway simultaneously, 7-way VLIW vision and would run 3 times more tasks than the original Nvidia Jetson.
Tim talks through the firmware will be using and which software he would be implementing with the Nvidia Jetson. By using the scope to connect the parallels, this would assist with reducing voltage, extending life and double its capacity. When discussing the architecture design, it demonstrates that the device would use RTC at night to power down, saving energy and consumption. It would also contain 1 docker connected to 20 smart fruit fly traps, this would reflect the Nvidia processing capacity and how fruit fly behaviour will determine a pattern of data.
Through Tim Millward’s seminar, we are able to gain a better understanding of his research and progress, but it also gives him the opportunity to demonstrate what he has learned and accomplished while developing his research. This seminar also gives students and staff the opportunity to benefit from the discussion of IoT networking and A.I solutions for smart agriculture, possibly incorporating the new knowledge into their own research. For individuals who may be considering a course in computing or for those who may have a general interest, these discussions provide an insight into current students’ research and where it will lead them.
If you are interested in a course in computing or for further information regarding this, click here.