We developed a drive mechanism and control system to enable low-cost vehicle speed interlocking control for rice planting devices. A flat permanent magnet synchronous motor and pair of spur gears were installed between two rows of planting devices. The control system was implemented on two microcontrollers connected by CAN communication. Only Hall-effect and current sensors were used for motor control. The system showed a 0.2s tracking response for a 5.0Hz ramp input. Under constant velocity control, the system showed a good tracking response up to 7.5Hz. Non-constant velocity control assuming sparse planting was also possible; a speed ratio of up to 30% was realized at an average planting speed of 5.0Hz.

This paper proposes a corner harvesting control for reliable turning at headlands during automatic harvesting operations using a head-feeding combine robot. On the basis of field experiments with a 4-row head-feeding combine robot, the proposed control methodology enabled the robot to make turns without trampling crop down or leaving any crops uncut. Regarding the accuracy of following the target path, the maximal lateral error of corner harvesting was 0.102m, and the maximal lateral error was 0.064m, and the angle error at the next target path after turning was 8.10°.

A multi-operation system was developed to improve the productivity and working hours of an intelligent greenhouse. The multi-operation robot is composed of a running unit, a growth diagnosis unit, a pest detection unit, a pollination unit, and a harvesting unit. In this study, we propose a system for monitoring the pest population by capturing an image of a trap sheet and recognizing it automatically. For the experimental data set, trap sheet images were collected in an intelligent greenhouse and the image processing engine that recognizes the number of trapped pests was tested. A standard deviation of 9.26 in the difference between the image processing result and the correct answer data was obtained from processing the images and from their subsequent analysis. The proposed system made it practical to collect the detailed mapping data of pest populations in and around a field.

An autonomous, electric-powered, unmanned airboat was developed for the efficient measurement of depth and water quality of inland water bodies. Compared with a gasoline engine, an electrical brushless motor was lighter, cleaner, quieter, less vibrating, and more controllable. The maximum distance and maximum operation period were estimated by measuring the vessel speed and power consumption with respect to the rotational speeds of the propellers. The accuracies of staying controlled on fixed points and in line following control were evaluated as the basic performance test of the boat. The results showed that the accuracies were sufficient enough for creating a map of the entire lake. Field trials were done in Churui Bay, Lake Akan, and Lake Onneto in Hokkaido, Japan. Depth maps and various water quality maps were successfully drawn for these lakes using the prototype.

Field trials of variable rate nitrogen fertilization with a　tractor-mounted optical real-time sensor were conducted in eight fields over 6 years for winter wheat. The algorithm determining the nitrogen application rate was based on the following procedure. First, a standard application rate was set to an average sensor value in a field measured by pre-sensing. Sensor readings were then translated into nitorogen uptake differences from an average point and converted to an application rate in real-time application. The experimental results showed that the average yield with variable rate application increased by 4.0% compared with uniform application, and that this effect was well-defined in panicle formation-stage fertilization. Protein content was more equalized in flag-leaf stage application. Furthermore, nitrogen harvest index with variable rate application was higher than that with uniform application. Therefore, variable rate application is an effective approach for winter wheat production.