Pervasive Data Science on the Edge

IEEE Internet Things M.

Motlagh

Zuniga

et al. 2021

Self Cite

Marine pollution is a growing worldwide concern, affecting the health of marine ecosystems, human health, and weather patterns. To reduce underwater pollution, it is critical to have access to accurate information about the extent of marine pollutants as otherwise appropriate countermeasures and cleaning measures cannot be chosen. Currently such information is difficult to acquire as existing monitoring solutions are highly laborious or costly, limited to specific pollutants, and have limited spatial and temporal resolution. In this article, we present a research vision of large-scale autonomous marine pollution monitoring that uses coordinated groups of autonomous underwater vehicles (AUV)s to monitor the extent and characteristics of marine pollutants. We highlight key requirements and reference technologies to establish a research roadmap for realizing this vision. We also address the feasibility of our vision, carrying out controlled experiments that address classification of pollutants and collaborative underwater processing, two key research challenges for our vision. Index Terms-AUV; Underwater drones; Marine pollution monitoring; Autonomous robots; Sensing Coordinated Movement and Processing: Besides coordinating sampling areas, AUVs should coordinate their movements to ensure they remain in each other's communication range. Coordinated movement also offers additional benefits, such as the potential for using collaborative sensing or processing strategies for improving the recognition of pollutants.

Section: B Underwater Distributed Processingmentioning

confidence: 99%

Toward Large-Scale Autonomous Marine Pollution Monitoring

IEEE Internet Things M.

Motlagh

Zuniga

et al. 2021

Self Cite

“…The shift in the paradigm even allows deep learning models to be run on smart devices in the fog [32]. A handful of smart devices can outperform a commodity desktop computer or laptop, and two to three can run deep learning-based image classification on a live 60 FPS (frames per second) video stream.…”

Section: Human Data Processing Techniquesmentioning

confidence: 99%

“…On Fog and Edge adoption: Approaches to harness the computational resources of end devices makes it possible to execute resource-intensive computations on the Fog/Edge [32,[37][38][39]. Complex Event Processing systems can benefit the most from these approaches [34] to process data by avoiding the overhead caused by communication latency.…”

Section: Related Approachesmentioning

confidence: 99%

Human Data Model

Mäkitalo

Flores-Martín

ACM Trans. Comput. Healthcare

et al. 2020

Self Cite

Today, an increasing number of systems produce, process, and store personal and intimate data. Such data has plenty of potential for entirely new types of software applications as well as for improving old applications, in particular in the domain of smart healthcare. However, utilizing this data, especially when it is continuously generated by sensors and other devices, with the current approaches is complex-data is often using proprietary formats and storage, and mixing and matching data of different origin is not easy. Furthermore, many of the systems are such that they should stimulate interactions with humans, which further complicates the systems. In this article, we introduce Human Data Model (HDM)-a new tool and a programming model for programmers as well as end-users with scripting skills that help to combine data from various sources, to perform computations, and to develop and schedule computer-human interactions. Written in JavaScript, the software implementing the model can be run in almost any computer either inside the browser or using Node.js. Its source code can be freely downloaded from GitHub, and the implementation can be used with the existing IoT platforms. As a whole, the paper is inspired by a number of interviews with professionals, and an online survey among healthcare and education professionals, where the results show that the interviewed subjects almost entirely lack ideas on how to benefit the ever-increasing amount of data measured of the humans. We believe that this is because of the missing support for programming models for accessing and handling the data, which can be satisfied with the Human Data Model. CCS Concepts: • Information systems → Data management systems; • Software and its engineering → Distributed systems organizing principles; Software design techniques; • Human-centered computing → Human computer interaction (HCI); Ubiquitous and mobile computing; • Computer systems organization → Distributed architectures; Embedded systems.

“…By attaching multiple containers into the same AUV -or different AUVs collaborating with each other -it becomes possible to form a small-scale micro cloud that is used to support different computational tasks. For example, deep learning based object recognition can be performed efficiently by using Raspberry PIs or old smartphone models as the computing units [18]. Besides image processing, these computing devices can be used to support coordination when multiple AUVs participate in monitoring [22].…”

Section: Componentsmentioning

confidence: 99%

Penguin

Proceedings of the 6th ACM Workshop on Micro Aerial Vehicle Networks, Systems, and Applications

Zuniga

Motlagh

et al. 2020

Self Cite

Underwater plastic pollution is a significant global concern, affecting everything from marine ecosystems to climate change and even human health. Currently, obtaining accurate information about aquatic plastic pollutants at high spatial and temporal resolution is difficult as existing methods are laborious (e.g., dive surveys), restricted to a subset of plastics (e.g., aerial imaging for floating debris), have limited resolution (e.g., beach surveys), or are unsuited for aquatic environments (e.g., wireless sensing or Fourier-transform infrared spectroscopy). We propose PENGUIN, a work-in-progress AUV-based solution for identifying and classifying aquatic plastic pollutants. PENGUIN has been designed as the first system that can both recognize pollutants and classify them according to specifics of the material. We present the overall design of PENGUIN, introducing the different components of the architecture, and presenting current status of development. We also present results of plastic classification experiments using optical sensing, demonstrating that simple PPG sensors provide a low-cost and energy-efficient solution for classifying different plastics. Our solution can easily monitor larger underwater areas than what current techniques offer while at the same time capturing a wider range of pollutants. CCS CONCEPTS• Computer systems organization → Robotic autonomy; Robotic components.