Digital Twin of a dual-arm autonomous seeding robot for hydroponic greenhouses

The technological transfer from industrial automation to agriculture in the last decade has accelerated the adoption of advanced tools for improving efficiency, productivity, and sustainability. Among these innovations, Digital Twin (DT) technology has emerged as a promising paradigm for modeling, monitoring, and analyzing complex agri-robotic processes. This article presents the design, implementation, and experimental validation of a DT for a dual-arm autonomous robot dedicated to the seeding process in hydroponic greenhouses. The system integrates a physical hydroponic environment with a virtual replica built in NVIDIA Isaac Sim, enabling simulation-based task testing, trajectory assessment, and hardware-in-the-loop evaluation under controlled geometric and operational conditions. The DT architecture is structured around IoT connectivity using FIWARE components, supporting bidirectional communication between the physical and virtual layers. The dual-arm robot is capable of perception-driven seeding tasks, including planter hole detection, cube positioning, seed collection, and deposition into rock-wool substrates, with collision-free manipulation strategies. The results demonstrate that the DT provides an effective framework for testing control algorithms, refining perception–manipulation strategies, and supporting robust operation under the evaluated greenhouse conditions. The proposed approach represents a step forward in applying DT technology to controlled environment agriculture, offering a basis for future scalable, AI-enhanced robotic systems that can optimize crop establishment and contribute to sustainable food production.