An adaptive control system for path tracking of crawler combine harvester based on paddy ground conditions identification

• Published in: Computers and electronics in agriculture Vol. 210; p. 107948
• Main Authors: He, Yongqiang, Zhou, Jun, Sun, Jingwei, Jia, Hongbo, Liang, Zian, Awuah, Emmanuel
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.07.2023
• Subjects: Adaptive navigation control; Crawler combine harvester; Ground condition identification; Path tracking; PSO-SVM; Path tracking; Adaptive navigation control; PSO-SVM; Ground condition identification; Crawler combine harvester
• ISSN: 0168-1699; 1872-7107
• DOI: 10.1016/j.compag.2023.107948

•Identifying the ground conditions of paddy.•A steering control model of the crawler combine harvester was established.•Local tracking path adaptive planning.•High navigation accuracy and stable control in various paddy ground conditions. The navigation technology of crawler combine harvester has a pivotal role in unmanned rice harvesting and has attracted much interest in recent years due to its labor-saving and effective operation. However, the complex and variable paddy ground conditions (PGCs) significantly impact the harvester's navigation tracking accuracy, and it is not easy to obtain stable and excellent performance. In this study, the cone index (CI) was used to classify the PGC and constructed a feature vector for characterizing the PGC using the vibration acceleration of the harvester. The Particle Swarm Optimization Support Vector Machine (PSO-SVM) algorithm was used to identify the PGC classes, and the identification accuracy was 91.73% and 91.67% for the training and the test sets, respectively. The field tests showed that the identification accuracy of the PSO-SVM algorithm for PGCs of each class is 85.0%, 83.0% and 90.0%, respectively. An orthogonal test method developed a steering control model adapted to PGC of three classes to enhance the path tracking accuracy, and the target path was tracked using a novel adaptive tracking control strategy. The field path tracking tests showed that the standard deviations of the lateral deviation of the harvester in the paddy fields of each PGC class are 0.053 m, 0.039 m and 0.045 m, respectively, and the standard deviations of the heading deviation are 1.120°, 0.895° and 0.877°, respectively. The proposed algorithm improves the performance of straight-line path tracking over the conventional pure tracking algorithm. Accordingly, the method proposed in this study can enable the crawler combine harvester to operate stably with high path tracking accuracy in paddy fields with various ground conditions.