This paper introduces an open-source platform called ISOBlue HD for acquisition of context-rich data from agricultural machinery. We call these datasets context-rich, because they enable the identification of machine status and farming logistics by properly labeling, fusing, and processing the data. The system includes a single board computer, a cellular modem, local storage, and power-over-ethernet switch to sensors. The system allows remote diagnostics and access, automatic startup/shut down with vehicle operations, and uses Apache Kafka to enable robust data exchange. ISOBlue HD was deployed in a combine harvester during a 2019 wheat harvest for simultaneously capturing 69 million CAN frames, 230,000 GPS points, and 437 GB of video data, focusing on header status and operator actions over 84 h of harvest time. Analyses of the collected data demonstrate that contextual knowledge can be inferred on harvest logistics (paths, speeds, header status, material transfer) and sensor data semantics.

For the past decade, estimating and predicting the geographical position of agricultural machines have become one of the primary research interests in autonomous agricultural vehicles, mapping and precision agriculture. The GPS data usually contain noisy and erroneous position estimates of the machine. Several experiments have devised kinematic models for agricultural machines and applied Kalman filtering technique for correcting the error and improving the accuracy of GPS data (Gomez-Gil, Alonso-Garcia, Gómez-Gil, & Stombaugh, 2011; Gomez-Gil, Ruiz-Gonzalez, Alonso-Garcia, & Gomez-Gil, 2013; Shufeng Han, Qin Zhang, & Hyun Kwon Noh, 2001). However, the existing Kalman filters are not sufficient in distinguishing different machine movements (moving in constant speed, turning or stationary) from the data. As a result, these Kalman filters unintentionally introduce some noisier estimates of machine location. In addition, the applications of these filters offer few insights in understanding and determining machine‘s motion. In this work, the Interaction Multiple Model with Extended Kalman filter (IMM-EKF) algorithm was introduced to compare against the single-model based Kalman filer algorithm with the simulated and collected GPS data set for a combine in a harvesting scenario. The kinematic models for modeling combine motions in harvest as well as the basics of the Kalman filter and IMM algorithm were presented. After the applications of the filters, results from both simulation and collected data suggested that the IMM-EKF algorithm not only offered improvement in estimates over the Kalman filter algorithm but also demonstrated an intuitive understanding of combine motions.