Embedding Artificial Intelligence on Remote Underwater Sensor Nodes

As climate changes are becoming a global concern, the Earth's oceans can potentially offer solutions to mitigate the impact as they absorb 100 times more heat, and 50 times more carbon dioxide than atmosphere. As such, there is a strong desire to monitor subsea activity in harbours, along the North American coasts, and in target areas such as the Arctic, where there is increased shipping activity. Canada is a world leader in the development of ocean technology, since it is surrounded by three large bodies of water. To get access to the resources, it would greatly benefit from an infrastructure to monitor the activity below the ocean surface and effectively protect its assets for commercial, scientific and military applications. In this research program, a self reconfigurable sensor network is proposed and will be enabled using machine-to-machine communication between remote nodes that can be anchored to the sea bottom or embedded on autonomous underwater vessels. For this purpose, an infrastructure to enable the Internet of Underwater Things (IouT) must be considered. Acoustic propagation is the primary means to communicate between untethered nodes below the sea. However, its bandwidth is limited. With the deployment of heterogeneous sensors, bandwidth requirements are time varying. In this program, a scalable communication network will be enabled between distributed nodes underwater to ensure that a reliable end-to-end connectivity is maintained while optimizing the overall energy efficiency of the network. The network coverage depends highly on the status of the environmental conditions. Using complementary sensors that can infer the propagation conditions, the autonomous network will be self-configured depending on spontaneous demands from the network nodes. The potential for this network to produce massive data sets is significant and, as such, a multi-rate adaptive communication system will be defined to maintain connectivity. The underwater acoustic network will be deployed using custom transceiver nodes implemented on self-reconfigurable acoustic signal processing platform. Automated event identification and detection, as well as target tracking will be enabled at edge nodes, which will be responsible for preprocessing and sharing mission critical data. Finally, security features will be implemented on the sensor network to preserve data integrity and eliminate risks of intrusion. This research program will advance the complimentary fields of ocean technology and information and communication technology that are key to Canada's economy. In the long-term, a multi-node underwater communication network will be demonstrated using custom transceivers that are self-configurable. The testbed will serve to demonstrate the feasibility to deploy autonomous surveillance systems. The program will also train highly qualified personal who will promote the science internationally and who will be ready to enter the Canadian industry.