Designing augmented human-plant interactions
The complex interactive system of embedded sensors in and around plants growing in closed environments scans the entire crop canopy and measures the plant temperature along with the chlorophyll fluorescence. The data collected on the rate that plants absorb energy, which reflect their current state of health are interpreted into high-level context information and communicated through specially developed wireless transmitters. This mixed society of plants and digital artefacts was regarded within the PLANTS project as a distributed system, which could globally manage its resources, functions and interactions with the environment. Plants were transformed into 'ePlants' with enhanced computational and communication abilities, which could be subsequently organised into groups in a virtual computing space. This hierarchical structure enabled the distribution of communication loads as well as its power and memory resources, but more importantly it facilitated distributed decision-making. At the heart of the system's layered, modular architecture, designed within the PLANTS project lies the management software. Both 'ePlantOS' for the 'ePlants' and 'eGadgetOS' for the digital artefacts adhere to the same basic concepts, and communicate by using common protocols and message structures. The functionalities of this middleware, developed at the Computer Technology Institute in Greece were further extended to include machine learning mechanisms and fulfil the evolving requirements of the mixed society. An interface between minute biosensors and actuators is provided, while an enhanced, plant-specific ontology is maintained by the ePlantOS middleware. The input/output (I/O) unit and connectivity layers administer the communication intricacies as plants become more than an information source. Local and global decision-making, in addition to distributed resource management can be supported as plants have the means to impact growing strategies in domestic among other closed environments. Moving this research effort one step further, towards to a more autonomous system with self-adaptation characteristics, the project partners will explore ways of incorporating Artificial Intelligence features into the system.
