No abstract

No abstract

The penaeid shrimp farming industry is a fast-growing sector which continues to suffer from significant feeding inefficiencies. Shrimp are slow to feed on pellets, with consumption dependent on a wide range of environmental and physiological parameters. Feed management on farms remains mainly based on feeding trays which can be difficult to observe and often result in overfeeding. While our understanding of shrimp feeding behaviour is beginning to improve under laboratory conditions, much less is known about shrimp behaviour in production ponds. Consequently, there is a growing interest within the industry to improve observations of shrimp feeding behaviour in situ, although this can be difficult due to high water turbidity and the benthic nature of shrimp. This review identifies key questions that remain unanswered in relation to shrimp feeding behaviour under commercial aquaculture conditions, and considers how they could be addressed using state-of-the-art applications based on three technologies commonly used in other areas of aquaculture. The use of passive acoustics, computer vision and telemetry are highlighted, alongside their potential to help farmers achieve better feeding efficiencies and sustainability as well as to help understand shrimp feeding behaviour in relation to various biotic and abiotic parameters.

No abstract

Existing artificial intelligence solutions typically operate in powerful platforms with high computational resources availability. However, a growing number of emerging use cases such as those based on unmanned aerial systems (UAS) require new solutions with embedded artificial intelligence on a highly mobile platform. This paper proposes an innovative UAS that explores machine learning (ML) capabilities in a smartphone‐based mobile platform for object detection and recognition applications. A new system framework tailored to this challenging use case is designed with a customized workflow specified. Furthermore, the design of the embedded ML leverages TensorFlow, a cutting‐edge open‐source ML framework. The prototype of the system integrates all the architectural components in a fully functional system, and it is suitable for real‐world operational environments such as seek and rescue use cases. Experimental results validate the design and prototyping of the system and demonstrate an overall improved performance compared with the state of the art in terms of a wide range of metrics.

No abstract

Unmanned Aerial Vehicles (UAVs) are promising technologies within many different application scenarios including human detection in search and rescue and surveillance use cases, which have received considerable attention worldwide. However, adverse conditions, such as varying altitude, overhead camera placement, changing illumination and moving platform, impose challenges for highperformance yet cost-efficient human detection. To overcome these challenges, we propose a novel combination of dilated convolutions with Path Aggregation Network (PAN) as a new deep neural network-based human detection algorithm in real time. Furthermore, we establish a comprehensive human detection dataset with varying backgrounds, illuminations, and contrast and train the proposed machine-learning model on the collected dataset. Our approach achieves both high precision (88.0% mean Average Precision (mAP)) and real time (67.0 Frames Per Second (FPS)) on a commercial off-the-shelf PC platform. In terms of accuracy, the result is comparable to the standard You Only Look Once v3 (YOLOv3). However, the speed is twice as that of the standard YOLOv3. YOLOv4 is slightly more accurate (89.8%) than our approach. However, it is slower (38.0 versus 67.0 FPS) and has more Billion Floating-Point Operations (BFLOPS). The proposed algorithm has also trained with the VisDrone2019 dataset and compared with seven studies using this dataset. The results have further validated the effectiveness of the proposed approach. Moreover, the algorithm has been evaluated on an embedded system (Jetson AGX Xavier), which demonstrates the usefulness of this method on power-constrained devices. The proposed algorithm is fast, memory efficient, and computationally less expensive to achieve high detection performance. It is expected to contribute significantly to the wider use of UAV applications including search and rescue missions to locate missing people, and surveillance particularly for applications running on resource-constrained platforms, like smartphones or tablets. This proposed system is now being used in aerial drone system of Police of Scotland to help them locate and find missing and vulnerable people. The results of the project were broadcasted by BBC Scotland.

The emerging fifth-generation (5G) mobile networks are empowered by softwarization and programmability, leading to the huge potentials of unprecedented flexibility and capability in cognitive network management such as self-reconfiguration and self-optimization. To help unlock such potentials, this paper proposes a novel framework that is able to monitor and calculate 5G network topological information in terms of advanced spatial metrics. These metrics, together with enabling and optimization algorithms, are purposely designed to address the complexity of 5G network topologies introduced by network virtualization and infrastructure sharing among operators (multi-tenancy). Consequently, this new framework, centred on a topology monitoring agent (TMA), enables on-demand 5G networks' spatial knowledge and topological awareness required by 5G cognitive network management in making smart decisions in various autonomous network management tasks including but not limited to virtual network function placement strategies. The paper describes several technical use cases enabled by the proposed framework, including proactive cache allocation, computation offloading, node overloading alerting, and load balancing. Finally, a realistic 5G testbed is deployed with the central component TMA, together with the new spatial metrics and associated algorithms, implemented. Experimental results empirically validate the proposed approach and demonstrate the scalability and performance of the TMA component. Keywords 5G networks • Topology management • Spatial network metrics • Cognitive management 1 Introduction Network management in the forthcoming fifth-generation (5G) mobile networks is notably influenced by the softwarization of network infrastructures where several hardware components are virtualized and by the multi-tenancy of the network infrastructures where hardware components are shared by different mobile operators. The main motivation of these 5G capabilities is the reduction of both capital and operational costs. 5G virtual network functions (VNFs) can now be deployed automatically and on-demand on the Edge and the Core segments of the 5G network and can be migrated between the computers that belong to the same net-Communicated by V. Loia.