DEVELOPMENT OF SMART FARMING IOT PLATFORM BASED ON SERVICE ORIENTED ARCHITECTURE

• Main Author: Andrianto, Heri
• Format: Dissertations
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/69727
• Institution: Institut Teknologi Bandung

The development of Internet of Things (IoT) technology has provided benefits for agricultural organizations to manage agricultural business processes such as monitoring, controlling, logistics and prediction. The development of IoT technology has the potential to create new information technology (IT) services in agriculture that can meet the needs and demands of the agricultural business. One of the factors that affect crop yields is the condition of nutrient deficiency in plants. Fertilization is an important step to increase soil nutrients and improve plant growth conditions. However, fertilization in plants is often done only based on habits, without clear information about the number of nutrients needed by plants. Therefore, we need a device and IT services that can determine the condition of nutrient deficiency and nutritional needs in plants. A chlorophyll meter is a device used to determine deficiency conditions and nutritional needs, especially Nitrogen (N). However, chlorophyll meters on the market today have limited data storage. In addition, the chlorophyll meter does not have a data transmission facility to the service system platform, therefore the data from the measurement of chlorophyll content cannot be monitored and analyzed further to provide fertilization recommendation services. The purpose of this research is to develop a smart farming IoT platform which includes an IoT-based chlorophyll meter device, a remote sensing system and a SOA-based smart farming IoT platform software that can be used to monitor nutrient deficiency conditions and provide fertilizer recommendations on plants based on chlorophyll content and Normalized Difference Vegetation Index (NDVI). In this study, an IoT-based chlorophyll meter has been developed and the device has worked well, namely being able to measure plant chlorophyll content, get position measurements, store data in memory modules, and send data to IoT smart farming platform software. The performance of the IoT-based chlorophyll meter has been compared with the performance of a commercial chlorophyll meter (SPAD-502) in measuring the chlorophyll index of plant leaves (Maniltoa grandiflora and Oryza sativa), with a coefficient of determination (R2 ) 0.9631 (Maniltoa grandiflora) and 0.7171 (Oryza sativa), this shows a significant correlation. A remote sensing system has also been developed using the Mapir Survey3 RGN multispectral camera mounted on the DJI Mavic 2 Pro. The calculation of the NDVI value uses the proposed NDVI algorithm. NDVI values have been compared with SPAD-502 values. The test results show the coefficient of determination (R2 ) between NDVI values, and SPAD-502 values is 0.81 (Oryza sativa), this indicates a significant correlation. The smart farming IoT platform software has been developed based on a service computing system platform reference model that is supported by using the service computing system engineering (SCSE) framework as an engineering methodology. The results of the smart farming IoT platform software design have met four SOA principles, namely a coupling factor of 0.00645 which indicates a loose coupling condition between services, a cohesion factor of 0.538 which indicates a strong relationship between services, a complexity factor of 0.012 which indicates a low level of complexity and reusability of 5.167 which indicates that the services can be reused quite well. The performance of the smart farming IoT platform has been tested using the Jmeter software. The results of the smart farming IoT platform performance test show the values of reliability (0.9999), availability (0.9997), integrity (0.9990), maintainability (0.9629) and safety (0.9090). The value of dependability measured from the five variables shows a value of 0.97 which is a very good level of system confidence not to fail in providing services to users under normal operating conditions with the possibility of failure that can still be tolerated. The main outputs of this research are IoT-based chlorophyll meter devices, algorithms to provide fertilization recommendations based on chlorophyll content values, modified NDVI algorithms, algorithms to provide fertilization recommendations based on NDVI values, and SOA-based smart farming IoT platform software for services of nutrient deficiency monitoring and fertilizer recommendations on plants. The main contribution of this research is to increase knowledge about the development of smart farming IoT platforms using a system engineering approach and SCSE framework, evaluation methods for smart farming IoT platforms using four SOA principles, performance evaluation methods for smart farming IoT platforms using dependability, and algorithm to provide fertilizer recommendations on plants based on chlorophyll index and NDVI values using linear regression and K-Means clustering correlated with leaf color chart (LCC).