The last few decades have seen an extraordinary acceleration of biodiversity change, both in terms of severity and speed. To counter the negative impact of the negative forces that impact the marine ecosystem, it is necessary to measure the contributing factors and communicating results in real time to be able to take immediate and necessary actions. Whilst measuring various factor is feasible using the latest sensor technologies, most underwater communication networks share an underlying deficiency. It is hard if not impossible to communicate directly across two mediums, i.e. water-air due to the water-air barrier. Deeply submerged sensors are unable to interact immediately with nodes that are on the water surface. This limitation is due to inherent properties of wireless signals in different medium. Current viable solutions include utilising wires to transmit the signal to the surface to an antenna or a processing unit that includes embedded computing platforms. New emerging wireless crossmedia communication are becoming available, which have a drawback. This is a low-bandwidth and high signal attenuation. An efficient communication protocol to counteract the inherent deficiencies of these new devices. The PhD research programme focuses on developing highly-efficient Artificial Cognitive Computing devices for advanced monitoring and behavioural analysis of underwater species. Such devices will combine lossless data compression techniques with state-of-the-art Deep Learning Networks (DLNs) algorithms (e.g. YOLOv7, MobileNetv3, EfficientNetv3) running on heterogeneous edge devices (HEDs). HED are specialised embedded computing devices equipped with Processor Systems connected via very high-speed internal buses connected to parallel computing devices (such as Graphical Processor Units, Field-Programmable Gate Arrays and Tensor Processor Units). The highly efficient DLNs will be used to process, on the HED, the data being collected from a wide range of sensors (such as vision, acoustic and LIDAR) and the output will be in the form of explainable AI. Furthermore, the project will also explore the use of wireless/optical underwater communications technologies to enable a means of data exchange for underwater (IoT) sensors. The project will also utilise the IoT sensors network technology that is being developed as part of an ongoing Internet of Water (IoW) project in our research group. The IoW focuses on the creation of technology to improve sustainability through real-time Nottingham Trent University 2 monitoring and behavioural analysis of ecosystems towards the detection and visualisation of pollution agents and tracking of populations of aquatic species. This PhD research programme will transform the IoW sensors into highly efficient artificial cognitive devices.