Standard Integration of Sensing and Opportunistic Diffusion for Urban Monitoring in Vehicular Sensor Networks: the MobEyes Architecture

Abstract

The emerging industrial relevance of vehicular sensor networks pushes towards their adoption for large-scale applications, from traffic routing and relief to environmental monitoring and distributed surveillance. With homeland security issues in mind, we have developed MobEyes, a fully distributed opportunistic harvesting system for urban monitoring. In MobEyes, regular vehicles equipped with sensors collect and locally store monitoring data while moving on the streets. Sensors may generate a sheer data amount, especially in the case of audio/video recording, thus making traditional reporting unfeasible. MobEyes originally adopts the guidelines of locally generating summaries of sensed data and of taking advantage of vehicle mobility and opportunistic one-hop communications to pump summaries towards mobile collectors, with minimal overhead, reasonable completeness, and limited latency. To that purpose, it carefully considers standard specifications to portably integrate with heterogeneous sensors, in particular by exploiting the Java Media Framework to interwork with cameras, the JSR179 Location API to interface with heterogeneous localization systems, and the Java Communications API to access lower-layer environmental sensors.