Estimating Missing Environmental Information by Contextual Data Cooperation

Abstract

The quality of life of users and energy consumption could be optimized by a complex network of sensors. Nevertheless, smart environments depend on their size, so it is expensive to provide enough sensors at low cost to monitor each part of the environment. We propose a cooperative multi-agent solution to estimate missing environmental information in smart environment when no ad-hoc sensors are available. We evaluated our proposal on a real dataset and compared the results to standard state-of-the-art solutions.

Edge-Based Missing Data Imputation in Large-Scale Environments

“…With respect to the state of the art, our proposal is based on the MAS approach that allows distributing the computation among agents operating locally. The local computation by agents allows reducing network traffic and deploying the solution in a large-scale environment where a high number of devices can be present [16]. This enables the development of edge computing applications.…”

A Cooperative Multi-Agent System for Crowd Sensing Based Estimation in Smart Cities

“…On the other hand, too much information makes the ACWs no specific of the time interval they describe. The mechanism to determine ACWs of variable size allows overcoming this problem; agents autonomously determine the size that each ACW must have so that it is useful for estimating missing information in the presence of similar information dynamics [11].…”

“…Our previous works are part of the neOCampus project [10] and focus on the estimation of missing values in the smart city to reduce the number of sensors deploy and ensure data availability to both citizens and experts. In [11], we propose a MAS to estimate missing information by exploiting the historical values perceived by individual sensors; the proposed solution performs an endogenous estimation. In [12], we propose a MAS to estimate missing information through a mechanism of cooperation between agents associated with devices perceiving homogeneous information (of the same type) and in a local environment.…”