For a sustainable agriculture, reliable measurements of soil properties and its interpretation are of highest relevance. Until today most of the measurements are carried out manually or by integrating off-line laboratories. Moreover, the number and density of measurement points is always an important aspect with respect to the statistical significance of the results. In this work a fully automatic measurement system has been developed and applied for the first time with free selectable measurement points.

As first measurement examples – in particular for soil compaction interpretations – a penetrometer module has been designed, where the penetration can be measured up to 80 cm below ground level. If the resistance (as for example caused by a stone) measured with a pressure sensor at the probe exceeds a certain level the measurement is automatically stopped via the vertical linear actuator control. New trials can be performed close to these measurements by programming the implemented horizontal linear motor for positioning the measuring rod perpendicular to the driving direction (within a range of 60 cm). Moreover, surface moisture and temperature are measured in parallel. The complete measurement module is attached to a multi-purpose field robot platform (“BoniRob”) which can navigate autonomously or via remote control on the field. The module can be considered as an application module (“App”) of the robot with defined mechanical, energetic and digital interfaces. A RTK-DGPS attached to the robot allows a precise positioning and GIS-based documentation of the measurement data. By combining the GPS navigation with the option of the horizontal linear positioning actuator user-defined positioning and statistics can be predefined.

 

Robot Operating System (ROS) has been chosen as the open framework software for the application module and the field robot. Within ROS a simulation of the application module and the field robot allows a realistic development environment as well as a transfer of the software to the real devices. The navigation in a stop-and-go mode for the robot is simulated as well as the app-controlled soil measurements. For validating the system, the measurement data has been compared to the data of a commercial penetrologger. The results match within the standard deviation of the data. At first the system has been tested independent of the robot, afterwards field measurements have been performed by using the robot in two modes: a “manual mode”, where the user controls the system via a remote control panel and a “semiautomatic mode” where the robot acts completely automatic, however, safety is assigned to the user. As a result GIS maps of sensor data have been generated as a basis for further interpretation and agricultural processing. The field robot based soil sensor system has a high potential for further applications, for example by including additional sensors.