Intelligent sensor networks in agricultural machinery In this project, solutions are developed to optimize the efficiency of the harvesting process of agricultural machinery. Through an intelligent sensor network, processes in the machine will be monitored and its data interpreted by AI. This requires novel concepts to determine process variables in different processing stages within self-propelled harvesters. This will increase the efficiency of the harvester while optimizing crop quality. In addition, the machine operator is significantly relieved.

Innovative sensor technology for monitoring combine harvesters How can the harvesting process of a combine be better monitored? How can the driver's workload be further reduced? What exactly happens in the combine harvester? These are the questions that CLAAS and the University of Bielefeld are addressing in the InSensEPro project. We are supported in the area of sensor technology by the company Müller-Elektronik.

Development of a modular sensor network for monitoring the harvesting process To develop the sensor network, a modular sensor platform will be designed in the project, on the basis of which both various physical measurement principles and technical interfaces for data transfer will be evaluated. On the basis of the measurement data acquired with the sensor network, an evaluation is developed using AI methods, which records the respective state of the process deep inside the harvester. The evaluation of the data from the individual sensor nodes is adaptively adjusted to the respective state, thus bundling the knowledge on the harvester and relieving the machine operator.

Comprehensive sensor network for process monitoring in harvesting machines Up to now, sensor technology has been very selective. Intermediate processes over the process are not yet detected by the sensor system. These must be estimated, but this can lead to incorrect estimates depending on the field conditions. For this reason, the operator must perform regular manual checks and calibrations. However, these are very time-consuming, which places an additional burden on the driver. The goal in InSensEPro is to develop the most comprehensive process monitoring possible by using a sensor network distributed throughout the harvester. To this end, a wide variety of techniques from the fields of machine learning and artificial intelligence will be investigated, ranging from simple logistic regression approaches to clustering methods and the use of neural networks. In doing so, we want to test and compare the capabilities of each method. Finally, we want to determine the optimal measurement positions for the ideal methods, which will enable the realization of the most efficient sensor network possible. In view of the scope of the sensor network and sensor positions that may be difficult to access with cables, another objective of InSensEPro is to investigate wireless communication methods for transmitting sensor data within the combine in the form of a feasibility study. The aim is to find out to what extent wireless communication is possible at all and how efficient it is in terms of transmission speed, signal strength, susceptibility to errors and energy requirements.