Indoorplant: um modelo computacional de serviços inteligentes baseados em históricos de contextos voltados à agricultura indoor

Abstract

The application of ubiquitous computing has increased in recent years, especially due to the development of technologies such as mobile furniture, more accurate sensors and applicable protocols for IoT. One of the trends in this area of research is the use of context sensitivity. In agriculture, the context can be related to the environment, for example, as conditions found inside a greenhouse. Recently, a series of studies has proposed the use of sensors to monitor production and/or the use of cameras to obtain information on cultivation, providing data, reminders and alerts to farmers. This dissertation applies a computational model for Agriculture indoor called IndoorPlant, which uses an analysis of historical contexts to provide intelligent services, such as previewing, indicating the problems that cultivation may suffer, giving suggestions for performance in greenhouse tests , among others. IndoorPlant was tested in 3 scenarios of the daily life of farmers with hydroponic production data that were used during 7 months of cultivation of radite, surface and arugula. These 3 display modes allow the ability of the defined model of a bot developed in Telegram Messenger for those who are farmers, communicate and receive it as model information. Finally, the results obtained through intelligent services using context histories are presented. The scenarios used services to recommend improvements in cultivation, profiles and finally, the prediction of the cultivation time of radite, lettuce and arugula occurred through the use of the PLS technique. The results of the predictions were relevant, since the radite obtained the values (R2 ) of 0.964, (RMSE) of 1.06 and (RMSECV) of 1.94. Lettuce obtained values of (R2 ) of 0.947, (RMSE) of 1.37 and (RMSECV) of 3.31 and arugula obtained values of(R2 ) of 0.932, (RMSE) of 1.10 and (RMSECV) of 1.89. Finally, a TAM form was applied to analyze the utility and ease perceived by farmers, obtaining approval of 92% for the utility of the model and 98% for the ease of use of the proposed model.