Machine Learning Approach for The Enhancement of Transmission Capacity and Latency for IoT Satellite Communications Based on Resource Allocation

Internet-of-Things (IoT) has become a high-demanding and prominent solution to any aspect of our life, and yet it is still evolving. A lot of applications are using IoT technologies as their main services, such as wearable devices, healthcare, traffic monitoring, as well as hospitality. However, the escalation of Internet-of-Things applications obviously requires a network infrastructure that has the capability to deliver high-throughput, low-latency, and reliablem communication. As we have known, the existing communication systems that have those capabilities are fiber optics and cellular networks such as 5G technology. And yet, is the service available anywhere around the globe? Commonly, IoT systems will be working properly in the coverage of the internet, roughly an area that is covered by cellular communications and fiber optics as well. Meanwhile, there are also possibilities of the IoT systems demand to enhance productivity of the industries that are located in remote areas, for instance, agriculture, transportation and logistics, maritime, environment, and mining industries. For example, in the agriculture industry, the total size of palm plantations in Indonesia is roughly about 16.38 million hectares1, while all of their locations are located in remote areas. Therefore, with cellular communications or fiber optics, IoT will not be able to work since those types of communications are not available across remote areas. Other land-based industries, such as mining and power plants, also have similar problems regarding internet connectivity. This problem also occurs in transportation, logistics and maritime as well, when their fleets are cruising across an area that is not covered by cellular communications, like the ocean or sky. The primary solution to the problem in order to unleash the possibilities is by giving access to the internet. Nevertheless, it is not feasible if we should establish cellular communications or fiber optics to remote areas, since it is related to the economical aspect of the providers that will run the service. Therefore, the primary solution should be establishing a communication system that is not limited by coverage area. Satellite communications (satcom) will be the most convenient solution to the internet access problem for remote areas. Currently, several satcom providers already have the IoT services that enable the IoT system to communicate through the satellites. As users (traffic) grow, the number of data is also increasing, therefore the IoT satellites will need to facilitate a higher capacity. In addition, the IoT satellites need to be reliable as well, herewith the probability of losing a packet should be minimized, either it is obviously related to downtime or bit-error-rate. The satellites should face the existing infrastructures for IoT such as 5G, in terms of capacity and reliability.


Project Leader:
Professor Wei Xiang, La Trobe University

PhD Student:
William Lukito