Journal of Terramechanics最新文献

Scene information plays a crucial role in motion control, attitude perception, and path planning for wheeled planetary rovers (WPRs). Terrain recognition is the fundamental component of scene recognition. Due to the rich information, visual sensors are usually used in terrain classification. However, teleoperation delay prevents WPRs from using visual information efficiently. End-to-end learning method of deep learning (DL) that does not need complex image preprocessing was proposed to deal with this issue. This paper first built a terrain dataset (consists of loose sand, bedrock, small rock, large rock, and outcrop) using real Mars images to directly support You Only Look Once (YOLOv5) to test its performance on terrain classification. Because the capability of end-to-end training scheme is positively correlated with dataset, the performance of YOLOv5 can be significantly improved by exploiting orders of magnitude more data. The best combination of hyperparameters and models was achieved by slightly tuning YOLOv5, and data augmentation was also applied to optimize its accuracy. Furthermore, its performance was compared with two other end-to-end network architectures. Deep learning algorithms can be used in the future planetary exploration missions, such as WPRs autonomy improvement, traversability analysis, and avoiding getting trapped.

Discrete element modelling (DEM) is widely used to estimate soil-tool interaction and tillage forces. To run an accurate simulation, it is essential to determine the appropriate DEM contact model and parameters. Although previous work has been introduced to determine the DEM contact model and parameters, the accuracy of numerical simulation is not high because of the soil differences when tillage tools operate in cohesion and adhesion soil in the middle and lower reaches of the Yangtze River in China. In this paper a Hertz-Mindlin with JKR Cohesion contact model and Linear Cohesion contact model were used to predict soil disturbance area and draft forces. The DEM parameters were determined using cone penetration, and uniaxial unconfined compression as an assisted test. The field experiment using sweep tool was used to validate the simulation results. A good agreement has been showed between simulation results and experiment results. Using verified model, the relative error for the predicted soil disturbance area at speeds of 0.50, 0.75 and 1.00 m/s were 5.3, 3.6 and 7.1 %, respectively. The maximum and average relative errors between simulated and measured draft forces were 6.98 and 3.91%, respectively. The effect of tillage depth and speed at soil disturbance area and draft forces were found which can provide some guidance for the selection of parameters during actual operation.

It is a well-known phenomenon that the color of sand changes due to moisture. As the moisture content increases, the sand will typically become darker. In our research, we are looking for the answer to the exact function according to which this darkening process changes and what relationship there may be between the different, initial (dry) soil colors and how this relationship can influence the evolution of the darkening process. Using the suitably chosen range of the color spectrum, we created a color parameter to characterize the color of the measured sandy soil, and then, in addition to determining the moisture content, we also looked for other uses of the color spectrum for sandy soil. Thus, based on the color spectrum and other input parameters, not only the moisture content can be calculated, but in the case of two components, the grain composition can also be determined.

A comprehensive plate-sinkage equation is necessary for the description of the load bearing capacity of soils. In the last century, several improvements to the existing equations were attempted but with limited success. The main aim of this paper is to verify, evaluate and develop a load bearing capacity theory of finite half space soil. Agricultural soils may be regarded as a finite half space in which the tilled soil layer is comparable to the loading diameter. Harder soil is found below the tilled soil layer and this hard soil can be considered as a rigid layer. A new consideration is the compacted cone-shaped zone developing under a loading device and its possible interaction with the rigid bottom surface. Theoretical and experimental investigations reported in this paper have shown that these approaches have facilitated deriving new relationships valid for finite half space. These include two independent variables and developing a dimensionless load bearing number. This paper introduces a new dimensionless plate-sinkage equation describing soil deformation in a general form.

This article aims to give a background about the soil shear strength and its measurement methods based on scientific articles and the work of researchers. A brief introduction is given about terramechanics science and the loads acting at the interaction zone between the tractive element (wheel/track) and the terrain. The most important loads exciting the terrain from the machine’s tractive element are the normal and the tangential loads. The tangential load will shear the terrain/soil and might lead to slippage, thus it is important to study the shear strength of the soil. In the review the soil terrain behaviour as an elastic and a plastic region is discussed. The conventional methods for measuring the soil strength used by scientists in terramechanics studies are reviewed. The influence of moisture content on soil strength is also taken into consideration. New ideas created by terramechanics scientists that emulate a real wheel/track - terrain interaction case for measuring the soil shear strength and are not civil or geotechnical engineering methods are mentioned. Finally, the shear strength results of loam sand soil obtained using the direct shear test conducted at the Hungarian University of Agriculture and Life Sciences (MATE) are presented.

In this study, an adaptive variable posture bionic mechanical foot is designed, which enables the transformation of different postures during the touchdown period. At the same time, the bolts at the joints are tightened to enable the non-variable configuration function of the bionic mechanical foot. A test rig was used to test the travelling and traction performance of the bionic mechanical foot at different speeds on sandy and hard surfaces. The results show that on sandy surfaces, at both high and low speeds, the variable posture mechanical foot outperforms the non-variable posture mechanical foot, especially at high speeds, indicating that the variable mechanical foot is suitable for movement at higher speeds on sandy ground. On hard ground, the traction and pedaling forces generated by the variable posture mechanical foot are essentially the same as those generated by the non-variable posture mechanical foot at low and high speeds, indicating that the travelling and traction performance of both mechanical feet on hard ground is the same. The variable posture mechanical foot is suitable for high-speed movement on sandy ground, providing a theoretical and technical basis for the design of future legged robots for efficient movement on desert surfaces and deep space soft surface environments.

Pressure imposed on an arable farmland by farm machinery can lead to severe soil compaction. A model was developed with Discrete Element Method (DEM) to simulate soil – tire interaction. Virtual dead weight method was performed for the purpose of model calibration. Simulated soil pressure data were obtained from the topsoil layer under varying number of tractor tire passes (1P, 2P, 3P, 4P, 5P, 6P, 7P and 8P). Simulation results were validated with maximum soil pressure data from a field experiment in which soil pressure was measured at 0.1 m depth in sandy loam soil. Model results of maximum soil pressure increased from 137.7 to 242.5 kPa when the number of passes increased from 1P to 8P. Prediction of the maximum soil pressure was reasonably accurate for 1P and 2P with Relative Mean Errors (R.M.E) less than 9%. Predictions for 3P to 8P had higher R.M.E. In terms of model application, soil sinkage and rolling resistance ranged from 0.07 to 0.14 m and 225.3 to 517.8 N respectively between one to eight passes. The model developed in this study can be used in the simulation of soil pressure distribution and deformation in the topsoil layer induced by heavy farm machinery.

The field of terramechanics focuses largely on two types of simulation approaches. First, the classical semi-empirical methods that rely on empirically determined soil parameters and equations to calculate the soil reaction forces acting on a wheel, track or tool. One major drawback to these methods is that they are only valid under steady-state conditions. The more flexible modelling approaches are discrete or finite element methods (DEM, FEM) that discretize the soil into elements. These computationally demanding approaches do away with the steady state assumption at the cost of including more model parameters that can be difficult to accurately tune. Model-free approaches in which machine learning algorithms are used to predict soil reaction forces have been explored in the past, but the use of these models comes at the cost of the valuable insight that the semi-empirical models provide. In this work, we presume that in a dynamic simulation, the soil reaction forces can be divided into a steady state component that can be captured using semi-empirical models and a dynamic component that cannot. We propose an augmented modelling approach in which a neural network is trained to predict the dynamic component of the reaction forces. We explore how this theory can be applied to the simulation of a soil-cutting blade using the Fundamental Earthmoving Equation and of a wheel driving over soft soil using the Bekker wheel-soil model.