Performance analysis of a user-centric crowd-sensing water quality assessment system

The water is an important commodity for the existence and development of human society. The placing, maintenance, and distribution of water, drinking or not, is laborious. The current physical sampling-based monitoring with periodic collection has been proven ineffective. The need of real-time monitoring motivated the development of the contamination warning systems (CWSs). Various approaches of CWSs that employ sensor technology have been developed to monitor water quality. By placing the humans in the loop, the assessment of the water quality and the provision of feedback can be enhanced. To evaluate this hypothesis, we developed the QoWater, a novel user-centric crowd-sourcing system that relies on measurements collected from sensors and user feedback to detect changes in the water quality. This system enables smartphone users to evaluate the quality of water based on its chemical properties, taste, odor, and color. It also informs users for potential contamination events, that have been detected close to their location. This paper analyses the performance of the QoWater system, in terms of power consumption, response delay, scalability and shows that the power consumption and response delay are relatively low. Furthermore, it presents a Trickle-based algorithm that extends the QoWater sensor battery lifetime by approximately 4 times.