Unmanned tobacco curing using a novel adaptive decision support system

• An adaptive decision support system was developed for unmanned curing, comprised perception, control, execution and software units. • A completely new control logic was developed and predefined in the adaptive control unit. • Temperature-humidity commands were automatically transmitted and executed based on real-time tobacco leaf status. • Eight conducted curing experiments reduced curing losses, improved curing quality and saved labor costs. Unmanned curing represents a key solution for addressing labor shortages and improving the quality and uniformity of tobacco leaves. This study developed an adaptive decision-support system specifically designed for practical application in curing barns, comprising perception, control, execution, and software units. The system enables automated perception, intelligent recognition, and real-time analysis of tobacco leaf status within the barn, along with autonomous transmission and execution of temperature and humidity control commands. Real-time monitoring of in-barn temperature and humidity, as well as RGB images of the leaves was achieved through self-developed sensors. A tobacco leaf status identification algorithm was employed to dynamically characterize the curing process, enabling the generation and implementation of adaptive optimization commands based on real-time leaf conditions to precisely regulate environmental parameters. Eight curing verification trials were conducted across Henan and Guangdong provinces in China. Compared with traditional curing methods, the unmanned system demonstrated enhanced precise and real-time capability in monitoring barn environments. Furthermore, it reduced curing losses by an average of 4.2 percentage points and generated annual cost savings of approximately 6500 CNY per curing barn. These improvements in operational efficiency and economic performance provide a solid foundation for large-scale deployment and indicate strong commercial viability.