Ad Hoc Networks. 6th International ICST Conference, ADHOCNETS 2014, Rhodes, Greece, August 18-19, 2014, Revised Selected Papers

Research Article

Video Surveillance Applications Based on Ultra-Low Power Sensors

Download65 downloads
  • @INPROCEEDINGS{10.1007/978-3-319-13329-4_21,
        author={Valeria Loscr\^{\i} and Michele Magno and Rosario Surace},
        title={Video Surveillance Applications Based on Ultra-Low Power Sensors},
        proceedings={Ad Hoc Networks. 6th International ICST Conference, ADHOCNETS 2014, Rhodes, Greece, August 18-19, 2014, Revised Selected Papers},
        proceedings_a={ADHOCNETS},
        year={2014},
        month={11},
        keywords={Surveillance PIR sensors Neural/genetic algorithm Coverage Connectivity},
        doi={10.1007/978-3-319-13329-4_21}
    }
    
  • Valeria Loscrí
    Michele Magno
    Rosario Surace
    Year: 2014
    Video Surveillance Applications Based on Ultra-Low Power Sensors
    ADHOCNETS
    Springer
    DOI: 10.1007/978-3-319-13329-4_21
Valeria Loscrí1,*, Michele Magno, Rosario Surace2
  • 1: Inria Lille
  • 2: University of Calabria
*Contact email: valeria.loscri@inria.fr

Abstract

Power consumption is an important goal for many applications, expecially when the power can be wasted doing nothing. Video surveillance is one of this application where the camera can be on for long period without “see” nothing. For this reason several power management techniques were carried out in order to reduce the activities of the camera when it is not needed. In this work we focus on surveillance applications performed through Video Surveillance Camera (VSC) that are not permanently active, but need to be properly “woken-up”, by specific ultra Low Power wireless Sensor Nodes (LPSN) able to monitor continuously the area. named. The LPSN are equipped by Piezoelectric “Passive” Infrared (PIR) sensors to detect the movement, thus they have a specific transmission range (to wirelessly send the “wake-up” messages to the camera sensor device) and a sensing range to detect events of interest (i.e. a man that crosses a specific area). Different deployments may highly impact not only in terms of events detectable, but also in terms of the number of VDS that can be woken-up. In this work, we propose a neural/genetic algorithm, that tries to compute the best deployment of the LPSN, based on two weight factors that “prioritize” the first objective, that is the number of VSC that can be woken-up or the second objective, namely the events detectable. The two objectives can be opposite and based on the different values assigned to the weight factors, different deployments can be obtained. The performance evaluation is realized through a simulation tool and we will show the effectiveness of our approach to reach very effective deployments in different scenarios.