Usability Of Mobile Based Participatory Sensing Application Using Intelligent Agents

Abstract

Citizens can act as sensors in crises management and volunteers for Geographic information and their surroundings. The proliferation of mobile phone in the developing countries sets a stage for the users to become participants in sensing their environment. A feat that was previously a preserve of specialized and hard coded sensors. Today’s mobile phone is equipped with numerous sensors ranging from the camera, the gyroscope, GPS, thermometer, barometer, accelerometer and the microphone. Whenever participants are engaged to crowd source data, they start with zeal but later drop out before sufficient data is gathered for accurate analysis. This project carries the objective of designing and developing an enhanced participatory sensing (PS) tool that will expand usability and entrench subtle incentive mechanism through porting intelligent agents. Experimental research design was used to ensure collaboration with volunteering participants of the tool in its development. The resulting tool is used to crowd source data form volunteering participants, which is analyzed and compared with data from non-enhanced tool. Data from the enhanced PS tool has accuracy levels of 86.6% as compared to 50% from non-enhanced tool under the SMART framework. However, very consistent participants even with the non-enhanced tool, showed an 85% data accuracy levels. This indicated the leveraging power of mediation of intelligent agents in PS tools.