Automated Agricultural Robot and Sensor Data Collection and Analysis through a Biomass Feedstock Production Information System

The increasing environmental pollution resulting from the use of non-renewable fossil fuels as well as the development of economic dependencies among countries because of the lack of such types of fuels underline the intense need for the use of sustainable forms of energy. Biomass derived biofuels provide such an alternative. The main tasks of biomass feedstock production are planting and cultivation, harvest, storage, and transportation. A number of complex decisions characterize each of these tasks. These decisions are related to the monitoring of crop health, the improvement of crop productivity using innovative technologies, and the examination of limitations in existing processes and technologies associated with biomass feedstock production. Other critical issues are the development of sustainable methods for the delivery of the biomass while maintaining product quality. There is the need for the development of an automated integrated research tool based on resilience and sustainability which will allow the coordination of different research fields but also perform research on its own. The specific tool should aim in the optimization of different parameters which specify the research done and in the case of biomass feedstock production; such parameters are the transportation of biomass from the field to the biorefinery, the equipment used, and the biomass storage conditions. This optimization would enhance decision making in the field of bioenergy production. Based on the need for such an automated integrated research tool, this paper presents an information system that provides automated functionalities for better decision making in the bioenergy production field based on the collection and analysis of agricultural robot and sensor data.

An End-to-End Learning-Based Row-Following System for an Agricultural Robot in Structured Apple Orchards

A row-following system based on end-to-end learning for an agricultural robot in an apple orchard was developed in this study. Instead of dividing the navigation into multiple traditional subtasks, the designed end-to-end learning method maps images from the camera directly to driving commands, which reduces the complexity of the navigation system. A sample collection method for network training was also proposed, by which the robot could automatically drive and collect data without an operator or remote control. No hand labeling of training samples is required. To improve the network generalization, methods such as batch normalization, dropout, data augmentation, and 10-fold cross-validation were adopted. In addition, internal representations of the network were analyzed, and row-following tests were carried out. Test results showed that the visual navigation system based on end-to-end learning could guide the robot by adjusting its posture according to different scenarios and successfully passing through the tree rows.