We are developing a system to estimate the position of rice seedlings throughout a paddy field as part of a control system for a rice-duck farming robot. In this paper, we propose an inexpensive and easy method for mapping rice seedling positions in an entire paddy field using a video taken from the ground with a monocular camera, and verify the accuracy of rice seedling position estimation using this method. Although some problems were observed in the validation of the method, the estimated positions of rice seedlings generally reflected the actual arrangement of rice seedlings, indicating that the method was capable of estimating rice seedling positions.

Rescue robots have been developed to collect information and perform tasks in place of humans at disaster sites, and there is a demand for autonomous rescue robots. To make robots autonomous, map generation based on advanced environmental recognition is necessary. However, the reflectance characteristics of the laser or infrared light emitted by the distance sensor used for map generation are affected by the material of objects in the environment. So there was a problem that the accuracy of the map deteriorated depending on the materials of objects in the environment. Therefore, we have proposed a method to classify the materials of objects in the environment using data from a distance sensor. As the first step of this research, we verified whether wood, metal, and concrete could be classified using a 3D-LiDAR and an RGB-D camera in our previous studies. We compared three types of machine learning methods (decision tree, AdaBoost, and neural network), and obtained an accuracy rate of 97% for the 3D-LiDAR and 83% for the RGB-D camera using decision tree. In this study, we aim to increase the accuracy rate of the RGB-D camera to 95% or more. For the machine learning method, we use the decision tree, which was the best in previous studies. In order to improve accuracy, three types of five explanatory variables are added. In addition, we optimize the hyper parameters of the decision tree using the existing hyperparameter optimization tool Optuna. We considered the contribution of each explanatory variable to the classification with the SHAP. And the essential explanatory variables were identified. As a result, we were able to increase the accuracy rate of material classification using the RGB-D camera to 96.5% and the Macro-F1 to 96.6%. In addition, we discussed the reasons why these explanatory variables.

Overhead-trellis environment is an environment where it is difficult to use existing position estimation methods such as satellite positioning systems. To automate farm work in the trellis environment, we proposed an autonomous mobile robot system using UWB as a position estimation method. In order to use UWB positioning system, it is necessary to determine coordinates of UWB devices placed in the trellis environment in advance. In related works, methods were proposed to automatically determine coordinates of UWB devices. In a method of related works, there is a problem that errors in coordinates of UWB devices that are far from the origin accumulate. Therefore, proposed method splits the UWB positioning area into multiple coordinate systems. The effectiveness of proposed method was verified through the evaluation of splitting coordinate system and autonomous moving of mobile robot.

This paper presents a feedback gain determination method for shaping the controllability Gramian with manifold optimization. We define a controllability Gramian that depends on the feedback gain, enabling it to be used in both stable and unstable systems. The controller design is then formulated as shaping this Gramian to achieve a desired input-state relationship. This study shows that the proposed method can be formulated as an optimization on a manifold with equality constraints. We also compare the Gramian shaping and the disturbance rejection control, such as H₂ control, and discuss the similarities and differences between the proposed method and the H₂ control approach. Finally, two examples demonstrate the effectiveness of the proposed method.