Search

Path tracking is critical for agricultural vehicles to achieve autonomous operation and to improve operational efficiency and accuracy. This study aims to address the high-precision path tracking requirements for the tracked vehicle GY-8. An improved pure pursuit path tracking control method is proposed to enhance the performance of path tracking. For the GY-8 vehicle's single-sided braking and steering approach, a dual-wheel differential kinematic model is established. A smooth steering method using PWM (Pulse Width Modulation) is designed to reduce the likelihood of deviation from the predetermined path due to PWM braking steering. A method based on the theory of circular arc similarity is introduced to determine the path curvature and segmentation. The NSGA-II optimisation algorithm is employed to optimise and obtain the optimal look-ahead distance for different curvature segments, thereby enhancing the accuracy of path tracking. The improved algorithm was experimentally validated for path tracking on paved surfaces. In the experiments, the improved algorithm demonstrated average error, maximum error, error standard deviation, and Fréchet distance of 0.0266 m, 0.0973 m, 0.0195 m, and 0.0891 m, respectively. This represents a 15.6%, 25.8%, 4.9%, and 27.6% improvement over the pure tracking algorithm. When applying the improved pure tracking algorithm to path tracking in agricultural orchard soil environments, the results indicated maximum error, average error, and error standard deviation of 0.1272 m, 0.0351 m, and 0.0215 m, respectively. The overall findings suggest that the improved method significantly enhances the accuracy of path tracking, providing theoretical support for advancing navigation technology in tracked vehicles. • An improved path tracking control approach proposed. • Circular arc similarity method determines curvature of each path segment. • NSGA-II optimizes look-ahead distance for varying curvatures. • Experimental results show that the algorithm performs satisfactorily. [ABSTRACT FROM AUTHOR]

The fully convolutional siamese network based trackers achieve great progress recently. Most of these methods focus on improving the capability of siamese network to represent the target. In this paper, we propose our model which focuses on estimating the state of the target with our proposed novel IoU (intersection over union) loss function which is named AIoU. Our model consists of a siamese subnetwork for feature extraction and a target estimation subnetwork for state representation. The target estimation subnetwork contains a classification head for classifying background and foreground and a regression head for estimating target. In order to regress better bounding boxes, we further study the loss function utilized in the regression head and propose a powerful IoU loss function. Our tracker achieves competitive performance on OTB2015, VOT2018, and VOT2019 benchmarks with a speed of 180 FPS, which proves the effectiveness of our method. • Powerful state estimation approaches promote Siamese trackers' performance. • Visual tracking can be decomposed to classification and regression task. • IoU loss with appropriate penalty terms leads to better regression results. [ABSTRACT FROM AUTHOR]