Research highlight #2: AI and robot driven autonomous irrigation

This research explores the integration of real-time plant health monitoring through non-invasive 3D-printed electrophysiological (EP) sensors and artificial intelligence (AI) to enhance smart agriculture. 

The developed system consists of a mobile robotic platform equipped with a conformable 3D EP sensor and a portable Faraday cage for data acquisition, alongside a customized AI-powered convolution neural network (CNN) to analyze the collected data. This innovative autonomous system is capable of monitoring tomato plants' EP signals under various irrigation levels, providing crucial insights into optimizing water usage. Notably, the 3D EP sensor showcases improved contact reliability and sensitivity compared to flat thin-film sensors, enabling accurate detection of subtle physiological changes associated with different irrigation conditions.

The results highlight significant findings: the 3D EP sensor can classify irrigation levels with an impressive accuracy of 86.91%, comparable to traditional RGB image-based methods, which achieved 86.37%. The study also employs scalogram images generated by wavelet transforms to analyze changes in frequency ranges and signal intensities across various irrigation levels. The successful identification of these parameters suggests that the EP monitoring approach offers a robust data analysis tool for understanding plant physiological responses to irrigation. Ultimately, this research lays the foundation for an advanced, automated irrigation management system, which could enhance crop yields, conserve water resources, and contribute to sustainable agricultural practices.