Specific Scientific and Technological Objectives

The goal of this objective is to design a system-wide solution that can model diverse agricultural practices and operations in a comprehensive way, relying on the cooperation between humans, environment, and smart hardware solutions (Cyber Physical Systems (CPSs) with edge-processing capabilities and Sensors). In CyCROP we aim at creating a new comprehensive, modelbased paradigm that is extended during the operational phase of the project and is flexible enough to support redesign, deployment, commissioning and decommissioning of hardware and software components. The central target of this objective is to develop methods and models that accurately depict and connect agricultural practices, such as, e.g., water and fertilizer usage, while including in the model all the hardware and software solutions developed within CyCROP. For these models to be coherent and to be able to be used for the deduction of useful results, they additionally need to adequately incorporate plant biophysical conditions, crop development timeline, and finally data regarding environmental conditions, impactful or extreme weather phenomena and relevant soil parameters. The final goal is to use this set of methods and models to develop a comprehensive Decision Support System (DSS) which will provide the end user with the relevant information to guide their actions through a user-friendly interface.

This objective focuses on the development of sophisticated hardware and software tools which as a whole will provide fast and accurate gathering of relevant agricultural and environmental data. Furthermore, specific software tools will be designed and deployed for data visualization, sensor data classification, clustering, and fusion.

Hardware: the solution will include hardware (and the required software) for the deployment of specialized Unmanned Aerial Systems (UASs) supported by 5G connectivity and long-range Wireless Sensor Networks (WSNs), targeting real-time crop monitoring and remote sensing applications. Furthermore, the use of edge processing on embedded devices will be considered and evaluated as a fast pre-processing stage for high bandwidth applications such as hyperspectral imaging.

Software: the proposed framework will take advantage of a continuous data feed taken from hyperspectral UAS imaging and a network of distributed, Internet of Things (IoT) enabled sensors. Both structured and unstructured data streams will be gathered, fused and managed either locally or in cooperation with other edge nodes and the cloud. The data management tools will support the collection, synchronization, and processing of data in order to update the current crop management models. For this purpose, current data (UAS spectral imaging, IoT sensors) and historical data (e.g., weather, crop disease models) will be collected and fused in a dynamic data stream. Finally, the software solution will include user feedback (expert input) in order to further tune the accuracy of the models taking into account the human experience.

Attention will be given to every step of the design and development process in order to ensure that the system can be easily replicated and scaled according to the deployment needs of different geospatial locations and crop types. This aspect will strengthen CyCROP’s proposition as a viable solution and exploitable service.