Field Robotics in Agriculture and Life Science
Multiple Intelligent Autonomous Spraying Robots
Typical sprayers spray a certain amount of pesticides in all directions without consideration to the surrounding environment. As a result, various problems occur, such as the acceleration of soil acidiﬁcation and economic losses due to the waste of agricultural chemicals. To overcome the above mentioned problems, we propose a deep learning-based intelligent spraying system that only sprays pesticides on only the fruit trees. The RGB image obtained from the camera is divided into images of 4-by-4 areas to determine if there is any fruit tree in the area, and each signals the mapped nozzle for on/off control. Also, we are applying SLAM to our intelligent spraying system so that our spraying can self-driving in orchards. Besides, task allocation studies are underway to apply this system to multiple robots to increase the efficiency of spraying in large orchards. When the system is completed, field tests will be performing at the real orchard.
Autonomous Fruit-Vegetable Harvesting Robot
As a solution of labor shortages in agriculture, especially, in fruit and vegetable harvesting, due to various reasons (e.g., increase of labor age) there is a huge attention for the development of an autonomous harvesting robot. From this perspective, we are focusing to develop an end-effector for harvesting fruit-vegetable in controlled horticulture. Firstly, we developed a four-finger type end-effector for sweet pepper harvesting, which can easily grasp the crop much less dependent on its size and shape. We also, secondly, developed a soft-robotics based end-effector for harvesting cherry tomato which has weaker surface than sweet pepper. Instead of a rigid grasping finger in the end-effector of sweet pepper, a soft material such as fabric was used to minimize the damage of surface. The developed end-effector will be integrated a harvesting robot and its performance also will be evaluated in a controlled horticulture.
Autonomous UAV-based Active Tracking and Mapping of Small Insect
Tracking micro-sized insects is one of the challenges of protecting ecosystems and biodiversity. In this study, we propose an approach for the localization and autonomous tracking of micro-sized radio-tagged flying insects, and develop an unmanned aerial vehicle (UAV)-based robotic system. The Kalman filter is applied to the received signal strength emitted from radio telemetry to estimate the position while reducing the measurement error and noise. The autonomous tracking strategy is a method in which the UAV rotates at one point to measure the signal strength and control its position in the strongest direction of the signal. We also design a system architecture comprising a tracking sensor system and a UAV system for radio-tagged micro-sized insects. The estimation and autonomous tracking of the target position by the proposed system are verified and evaluated through dynamic simulation. Furthermore, field tests are performed to demonstrate the feasibility and scalability of our approach. Therefore, in this study, we propose and validate a UAV-based tracking system for micro-sized flying insects, which has not been proposed in studies conducted thus far.