cogcom 15(2): e4

Research Article

Switching Brains: Cloud-based Intelligent Resources Management for the Internet of Cognitive Things

Download455 downloads
  • @ARTICLE{10.4108/cogcom.1.2.e4,
        author={R. Francisco and A.M. Arsenio},
        title={Switching Brains: Cloud-based Intelligent Resources Management for the Internet of Cognitive Things},
        journal={EAI Endorsed Transactions on Cognitive Communications},
        volume={1},
        number={2},
        publisher={ICST},
        journal_a={COGCOM},
        year={2015},
        month={5},
        keywords={Internet of Things, Finite State Machines, Resource Optimization, Wireless Sensor and Actuator Networks, Cloud Computing, Middleware},
        doi={10.4108/cogcom.1.2.e4}
    }
    
  • R. Francisco
    A.M. Arsenio
    Year: 2015
    Switching Brains: Cloud-based Intelligent Resources Management for the Internet of Cognitive Things
    COGCOM
    ICST
    DOI: 10.4108/cogcom.1.2.e4
R. Francisco1, A.M. Arsenio2,*
  • 1: YDreams Robotics and IST, Edificio A Moagem - Cidade do Engenho e da Artes, Largo da Estação, 6230-311 Fundão, Portugal
  • 2: YDreams Robotics and Universidade da Beira Interior, Edificio A Moagem - Cidade do Engenho e da Artes, Largo da Estação, 6230-311 Fundão, Portugal
*Contact email: arturarsenio@di.ubi.pt

Abstract

Cognitive technologies can bring important benefits to our everyday life, enabling connected devices to do tasks that in the past only humans could do, leading to the Cognitive Internet of Things. Wireless Sensor and Actuator Networks (WSAN) are often employed for communication between Internet objects. However, WSAN face some problems, namely sensors’ energy and CPU load consumption, which are common to other networked devices, such as mobile devices or robotic platforms. Additionally, cognitive functionalities often require large processing power, for running machine learning algorithms, computer vision processing, or behavioral and emotional architectures. Cloud massive storage capacity, large processing speeds and elasticity are appropriate to address these problems. This paper proposes a middleware that transfers flows of execution between devices and the cloud for computationally demanding applications (such as those integrating a robotic brain), to efficiently manage devices’ resources.