CPS: MEDIUM: CONTROLLED WASTEWATER-HYDROPONIC SYSTEMS FOR ENHANCED NUTRIENT AND WATER EFFICIENCY USING REAL-TIME SENSING AND DATA-DRIVEN TECHNOLOGIES

Thrust 2: Hydroponic vegetable growing system sensing, modeling, and simulation.In this thrust, we will gather robotics imaging data by plant growth monitoring, and TENG-waterdrop sensor data for nutrient monitoring. Similar to Thrust 1, two ML models will be established for lettuce growing status identification from the imaging data, and nutrient profile identification from the triboelectric response signals, respectively. This will allow real-time monitoring of plant and water conditions for RL-based control and optimization. A BPINN model will be developed for the simulation of the plant growth nutrient uptake process.Task 2.1: Experimental data acquisition and quantification for hydroponic lettuce growing systemOur robotics imaging system serves to capture comprehensive, plant-level research plot images. These images facilitate the generation of precise 3D reconstructions, offering crucial structural insights such as volume measurements for individual lettuce crops. Complementary 2D imagery, capturing the canopy cover, supplements this dataset. Validation of the imagery involves laboratory analyses that establish ground-truth data, including metrics like true mass and volume. The accuracy and reliability of both 2D and 3D image parameters and analysis techniques are verified against hand-measured metrics. The cable robot's operational procedure involves positioning the camera above the target crop and rotating around it to capture a series of 64 images. These images serve as the basis for 3D point-cloud reconstruction, enabling direct inference of morphology, volume, and ultimately, mass estimation.?