Design of a Wireless Underground Sensor Network for Precision Agriculture

Résumé

During the past few years, Wireless Underground Sensor Networks (WUSNs) become widely used due to their large amount of applications. These applications are classified into mine detection, landslide activities, ecology monitoring, or precision agriculture. in this latter, the buried nodes have to check the good growth of plants by verifying the water content, the temperature, and the presence of nutriment. Thus, the user is able to decide to water or to add fertilizers in a particular area, therefore, efficient use of the resources is performed. However, since the ground which is denser than the air is the communication channel, the electromagnetic (EM) waves used for wireless communications are widely attenuated due to soil properties that may change over time. Thus, a sensor node must waste its energy by sending its sensed data to a destination node without being received by this latter due to signal loss in soil. A WUSN requires beforehand to allow reliable communication between buried sensor nodes. This thesis aims at allowing reliable and energy-efficient communication in WUSN for real-time application of precision agriculture. To achieve it, we proposed the Wireless Underground Sensor Network Path Loss Model called WUSN-PLM for the prediction of the signal loss in precision agriculture. In order to validate the proposed WUSN-PLM, intensive measurements have been conducted in a real agricultural field of onions culture at the Botanic Garden of the University Cheikh Anta Diop of Dakar. Over the 140 measurements, WUSN-PLM outperforms the existing models with 87:13% precision. Furthermore, for a real-time prediction of the packet loss, we proposed a link channel optimization for reliable communications in WUSNs based on the Sugeno Fuzzy Inference System (FIS). The proposed FIS consists of 04 inputs, one output, and 36 rules. The inputs give information related to the buried depth of transmitter and receiver nodes, the average value of the soil moisture proportion, and the linear distance between nodes. The output of the FIS gives the reception probability of a packet sent by the source node according to the previous parameters. The evaluation of the proposed approach obtains a higher accuracy and precision than WUSN-PLM (91:429% and 87:129% respectively) The Fuzzy Logic-based approach has been integrated within real sensor nodes made up of ARDUINO boards for practical use.

Citer ce document

Accès au document

Ce document est hébergé sur son dépôt institutionnel d'origine.

Auteur(s)

Statistiques

Consultations : 2

Téléchargements : 0