Detecting stresses in crops early by collaborative robot-sensors-human system automation

Stress in crops indicates a disease will soon infect the plants. Detecting the stress condition in the earliest stage is critical to preventing diseases and loss of yield. A new protocol is developed in this project to solve key execution planning and control problems: When, how, and where to optimally monitor and detect. The solution aims to significantly prevent disease propagation in large crop production areas, under time, accuracy, and cost constraints. Collaborative control theory (CCT) is utilized to design and construct the system, which synchronizes humans, a mobile robot, and sensors to effectively perform the task. By using greenhouse as a case study model, the protocol will route a robot to logically visit the sampled locations and searching guidance can be improved by scientifically known characteristics of how stressed crops can spread the disease to other plants. Simulation experiments have been performed and results indicate that the routing algorithm increases the detection rate of stressed plants by 37.74%. The adaptive search algorithm improves the number of detected stressed plants by 71.88%. This presentation describes the human-robot-sensors system, the CCT protocol and algorithms, and experiments' findings, including how the new protocol yields the highest overall robotic efficiency and accuracy.