Towards a smarter system for Human Sensor Web

Abstract

Human sensing is a notion of crowd-sourcing whereby ICT devices are utilized for data collection. Human Sensor Web (HSW) is a network of people who interact with their devices in order to forward their observations to a designated receiving server in the form of messages (such as SMS and USSD). It capitalizes on the accessibility of ICT tools (such as mobile phones) by non-experts to use them as sensory nodes in order to generate useful data regarding various location-oriented phenomena - such as the status of public service facilities. We presumed and tested that with the controlled use of its context, the smartness of the HSW can be boosted. The smart system uses context to make intelligent analysis such as credibility assessment of user-generated data and context-aware retrieval of geo-information, which would have been impossible otherwise. In this paper, we proposed and tested a software architecture to achieve this goal. We deployed the design solution on a mobile reporting system for functionality of water points in rural Tanzania under the SEMA project. The context-enabled smart system has been developed based on latest semantic technologies and linked data principles.