ABSTRACT. Some organizations in Japan have planned planetary explorations using a small rover. However, the rovers with cylindrical wheels have a risk of failing to move in the loose soil area, such as the surface of the lunar/planet. The movement using the supporting force of the locked wheel, for example, a wheeled walking, or push-pull locomotion, can reduce the risk. Previous studies have mainly developed relatively large rovers over a mass of 10 kg. Meanwhile, our group has developed a small and lightweight (under a mass of a few kg) push-rolling rover with minimal configuration, and our previous studies indicated that the rover could climb steep slopes over 30 degrees with low slip. Planetary exploration rovers are required to move in an arbitrary direction; therefore, the developed rover also needs to have a function of a turning motion. This paper proposes the turning motion while inching for the push-rolling rover with minimal configuration and evaluates its performance by the traveling experiments at different conditions (driving conditions, and wheel shapes). The experimental results indicated that the proposed turning motion using high wheel slip and lug’s wheel realized turning with low slip on loose surfaces with steep slopes. Additionally, the experimental results also showed that the slip amount of the rotational center is lower than that of a normal skid-steering.

ABSTRACT. Recently, legged robots have attracted considerable attention as highly mobile rovers for planetary exploration. However, the surfaces of celestial bodies such as Mars and the Moon are primarily loose, causing slippage due to surface deformation from the rover’s leg movements. To mitigate this, we proposed a walking method aimed at preventing slippage. In our previous study, we evaluated the effectiveness of this method using a legged testbed on sloped, loose ground. The results demonstrated improved mobility performance of the legged rover. It is crucial to investigate the increase in bearing capacity by imparting vibration in a realistic space environment to validate the effectiveness of the proposed method. This study examines the change in bearing capacity due to vibration under low atmospheric pressure, a condition found on Mars and the Moon. This condition affects the characteristics of ground sand due to decreased air resistance. Our findings provide valuable insights into using vibration in planetary exploration to enhance the performance of legged rovers. The experimental results indicate that the bearing capacity under low atmospheric pressure is nearly identical to that under standard atmospheric pressure, suggesting that atmospheric pressure conditions have minimal impact on the supporting force when vibration is applied.

ABSTRACT. In recent years, many organizations have developed lunar rovers for traveling over uneven terrain. Our research team has studied a rover using independently extending/retracting left and right wheelbases. In the case of four wheels, this method allows three wheels to remain stationary while moving. The independent extending/retracting locomotion of the wheel enables the traction to increase. This method provides a greater support force and improves the climbing performance on a loose slope. However, increased load on a sloped surface causes sideslipping, and the rover tilts its posture in downward. Our previous study confirmed that intentionally increased sinkage by the large wheel slip increased side forces from the soil. However, the lateral force is still insufficient and cause sideslipping on steep slopes. To increase the side force, this study focuses on the resistance between the wheel and the ground, along with the direction of wheel rotation. Specifically, we propose reverse rotation of the upper side wheel of the rover. This method increases dynamic sinkage by increased resistance from the soil during wheel driving. To confirm the effectiveness of method, this study conducts experiments that traversing loose slope with the rover. The results suggest that the proposed method greatly reduces sideslipping. In addition, the results indicated that the wheel sinkage increased. Therefore, the reverse rotation could increase the normal force of the wheels on the upper side. In summary, proposed method increased side forces and suppressed sideslipping. Hence, the reverse rotation prevents tilting the rover’s posture downward on the slope.

ABSTRACT. There has been an interest in the exploration of planetary bodies for more than half a century. Scientific research has been focused on understanding the origins of life, potential for habitability in extraterrestrial environments, and discovery of rare minerals and energy sources. Extraterrestrial rovers have been used previously by space agencies for the exploration of Lunar and Martian surfaces. Reduced gravity on such surfaces influences the performance and mobility of extraterrestrial vehicles. Gravity can change how the soil behaves under the wheel as well as the traction force and sinkage developed by the wheel. The sinkage rate and depth of a wheel and motion resistance are also impacted by gravity. As the rovers are operated by interplanetary telemetry, autonomous real-time control is employed in these systems due to communication delays. Optimal online control with a high-fidelity terramechanics model may not be feasible and it necessitates the development of simplified models which consider the effect of reduced gravity on dynamics of a vehicle. The goal of this work is to study the effect of reduced gravity on Lunar terrain and its effect on vehicle performance of rovers through analytical modeling of tire-terrain interaction. We have developed an optimal control algorithms to improve vehicle stability and trajectory tracking performance in standard maneuvers. The simplified terrain models can also be extended for hardware-in-loop simulations.

ABSTRACT. The Moon's surface is covered with a layer of loose, weathered rock (fine regolith) several meters thick. This type of surface can create problems for the movement of small lunar rovers, especially in sloping areas. Legged robots are able to perform complex operations, but their design requires appropriate sensors to maintain balance, on the other hand, in the case of wheeled robots there is a risk of slipping. In order to combine the advantages of both, it was decided to create a hybrid robot. The main purpose of the article is to present the design and testing of a prototype lunar rover. The design of a hybrid robot, which combines the characteristics of wheeled and legged robots, is shown in the article. The center of gravity may be dynamically adjusted thanks to the use of two robot arms with wheels attached. Furthermore, the robot can "crawl" on sloping areas where a wheeled vehicle could slip owing to loose ground. Crawling is possible through the use of a mechanism for swinging the blades, which are mounted on the arms. The article describes the design of the rover, describes the results of research regarding the performance of crawling mechanism, and presents the limitations of the current design of the rover and presents possible improvements and further development work.

ABSTRACT. Tracked robots, which have flippers on the front, back, left, and right sides, are expected to be used for disaster investigation because of their high performance to overcome obstacles such as debris and bumps. However, it is difficult for the operator to control the robot because of its high degree of freedom. Therefore, the system that automatically controls the flippers and crawlers is required. In this study, we examined the acquired motion of a tracked robot with flippers to climb a step without using environment information by reinforcement learning. The learned motions are targeted to climb a step efficiently with a small amount of motion and to prevent the large impact when landing on a step. In order to reduce the amount of information required for the decision of the motion, only the information obtained from the internal sensors is used without the information of the surrounding environment. The agent Learned in a simulation environment using multi-body dynamics. First, the robot was trained to climb a step from the front, and the effectiveness of the acquired motion was confirmed from the results of a step climbing simulation and experiments using the trained agent. Then, the motion of the robot for climbing a step from an angle acquired by randomly changing the robot's initial orientation was clarified.

ABSTRACT. The vehicle dynamics of off-road vehicles tend to be non-linear concerning suspension dynamics, tyre-terrain interactions, and subsystem kinematics. These non-linearities limit the performance of control due to the need for linearization to enable real-time application. Implementing real-time non-linear controllers and prediction methods on embedded systems, such as the dSPACE MicroAutoBox 2 (MABX2), is challenging due it being computationally intensive. This challenge is particularly pronounced for controllers utilizing optimization schemes with non-linear solvers to solve optimal control problems. In this paper, we explore the possibility of embedding non-linear optimal controllers and predictors on a MABX2 for real-time applications. The implementation uses CasADi to formulate and code-generate the problems, and then executes this code using S-Functions in Simulink or off-loading the execution to a powerful external computer. We discuss the limitations and special considerations for these real-time implementations for off-road vehicle applications.

ABSTRACT. An integral part of autonomous forestry is the ability of the vehicles, e.g., forwarders and harvesters, to perceive their environment. At Luleå University of Technology, object detectors have previously been developed, allowing forestry vehicles to detect and position important objects in forestry, such as tree stumps, stones, and logs. These detectors have been developed by training on physical manually annotated data, which is both time-consuming and costly. Training on virtual data allows for significant time- and cost reductions. Since the ground truth in the virtual model is known, the training data can be auto-annotated, allowing for the creation of larger training datasets, at a lower cost. In this work, a virtual environment in Unity is used in co-simulation with a real-time digital twin of a physical forestry vehicle, to generate auto-annotated training data, as captured by an onboard stereo camera. A detailed emulation of the stereo camera is used to achieve realistic results. First, a log detector trained on physical manually annotated data, is evaluated on virtually created data. It is shown that the log detector trained on physical data can detect logs in the virtual environment. Second, new detectors are trained, using different shares of physical and virtual data. It is shown that a detector trained using only virtual data, can learn to detect logs in the physical world. Moreover, virtual pre-training is shown to improve the performance of physically trained and tested detectors, both at low availability of physical training data, and in terms of domain generalization. Furthermore, the real-time capable virtual models also enable future machine learning tasks utilizing different levels of Hardware-in-the-Loop.

ABSTRACT. Accurate modelling of the interaction between a tyre and the terrain is crucial for successful vehicle dynamics simulation, particularly in challenging off-road conditions. This study introduces a novel approach for obtaining highly accurate three-dimensional terrain models for vehicle simulation. Utilizing an unmanned aerial vehicle (UAV) equipped with a high-resolution camera, we developed three-dimensional terrain models of an undulating test track with two approaches. To validate the accuracy, these models were compared with measurements obtained from a traditional mechanical road profilometer. The results demonstrate a strong correlation between the two measurement approaches. The primary advantages of the UAV method lie in its speed of data acquisition without compromising accuracy and that the measurements are immune to the terrain macro roughness. In contrast to the labour-intensive measurements required by a mechanical road profilometer and extensive subsequent post-processing, the UAV approach requires minimal time and effort. Additionally, since the UAV method is not coupled to the terrain roughness, it eliminates the integration drift associated with other ground-based approaches which are subjected to terrain excitation. This opens the door to the possibility of modelling very rough terrains and using these terrain models for vehicle dynamics simulations in extreme off-road environments.

ABSTRACT. During off-road operations, mobile robotic platforms often encounter objects that influence the platform’s route. As determining the outcome of an interaction with an object (e.g., overriding) is difficult, many robotic planners are designed to avoid interactions with all environmental objects. Yet, this object-adverse planning behavior is not reflected in the actions of expert human operators, who may interact with objects in order to find a viable path towards their goal. This work intends to emulate that human operator intuition. Our objective is to improve the performance of robotic traversals in off-road terrains through the development of a planning paradigm that allows certain safe contact with environmental objects. Specifically, we design a two-level hierarchical planning and control system which couples a contact-informed regional motion planner with contact-constrained local nonlinear trajectory optimization techniques. The approach is demonstrated for classes of vegetative objects which are characterized by existing parameterized collision models from the terramechanics community. The top level of the hierarchy combines sampling-based planning techniques with a set of override (velocity) constraints derived from these collision models during the search for a minimum-time trajectory towards the platform’s goal. This minimum-time trajectory is then passed to the lower-level of the architecture, which utilizes direct collocation and model predictive control techniques to ensure that the velocity constraints are enforced during trajectory execution. The capabilities of the architecture are shown both in simulation and onboard a robotic platform, where vegetation is reasoned about and then overridden depending on the environment, platform, and object geometry.

ABSTRACT. DEM is computationally intensive for granular dynamics simulation, leading to a need for efficient strategies. This study explores using local particle refinement, scaling particle size based on expected spatial resolution needs, inspired by adaptive mesh refinement in FEM. Finer particles are used where intense interaction occurs, and coarser particles further away.

We hypothesize this method can maintain good accuracy while reducing particle count and computational effort. Fine particles are used on the soil bed's top, with coarser particles at greater depth, creating a particle size gradient. By adjusting the gradient we introduce a “scaling aggressiveness”, allowing control over the trade-off between efficiency and accuracy.

We use triaxial tests to verify that the method is scale invariant. Pressure-sinkage and shear-displacement tests are then used to evaluate the method's effectiveness and accuracy in terramechanics applications.

All beds were compared to a reference bed with homogenous particle size, where the mean static sinkage was 1.25 mm for a 50 kPa load. The dynamic sinkage was 73 mm for the full simulation time.

For quasi-2d simulations, mild scaling aggressiveness reduced the particle count by 2-4 times with relative error up to 4% for dynamic sinkage (11% for static sinkage). For medium aggressiveness, 4-6 times reduction with relative error of 4% (19% static). For highest aggressiveness, 6-8 times reduction with relative error of 7% (29% static). The internal friction proved to be very resistant to gradient changes, with errors within 1%.

When extending the model to full 3D, we estimate up to a reduction in particle count of up to a factor 25.

ABSTRACT. Modern militaries are exploring the teaming of military vehicles with smaller uninhabited ground vehicles (UGVs), to improve the success of operations in the off-road terrains. The UGVs can be used to perform initial mobility testing on soft soils, to predict the go/no-go performance of vehicles. Because of the variation in the sizes of the UGV and military vehicle, it is imperative whether the scalability of tyre-soil interaction exists or not. The scalability assumes that similar systems behave in a similar manner at different dimensional scales. Dimensional analysis is carried out to determine similarity between the systems and identify design parameters affecting scalability. In this study, the lightweight vehicles (FED Alpha) are considered as the full-scale systems (as upper boundary) and UGVs (Husky or Warthog) as scaled system. The 335/65R22.5 tyre with operational range of loading for full scale vehicle is considered. The smaller UGV tyres (0.7, 0.5 and 0.25 scale) represent scaled system. The 2NS and fine-grain sands were modelled using the DEM (EpAM contact model). The direct shear and pressure-sinkage tests were simulated to calibrate the soil model (cone index from 14.79-149 kPa). Validated simulations of tyre-soil interaction, show that 'drawbar-pull vs slip' and 'tractive-efficiency vs slip' are scalable, within given size and loading conditions. However, the prediction is dependent on soil parameters and size of the scaled systems (0.7 and 0.5 scale demonstrated the scalability clearly). The prediction was better in 2NS sand due to higher cone index. Up to 0.5 scale-system can predict the full- scale system’s mobility performance on sandy soils. This finding can be used to develop lighter UGVs to support full-scale vehicles in the off-road terrains.

ABSTRACT. Tracked vehicles are extensively employed in unstructured environments, traversing on uneven and deformable terrains. The mobility facilitated by the track unit which is composed of interconnected metal plates, is pivotal for enhancing vehicle mobility and safety. In the context of construction machinery, there is a paramount need for stable traversability across sand imbued with moisture. Despite this, the existing research on evaluating tracked vehicle performance in wet sand conditions is significantly limited in contrast to the extensive research conducted on dry sand. This study aims to develop a DEM-SPH-based simulation methodology for assessing tracked vehicle traversability on water-laden soil. The DEM, or Distinct Element Method, is a proven technique for soil-machinery interaction analysis, but it inaccurately represents water-infused sand dynamics. Incorporating the Smoothed Particle Hydrodynamics (SPH) addresses these difficulties, allowing it to improve the simulation accuracy of moist soil behavior. Our approach involves calibrating and validating DEM-SPH parameters through cone penetration tests across various soil moisture levels and resistance force variations. Subsequent cross-validation of the calibration was performed by examining the deposition angles of differently moistened sand in a cylindrical container. By calibrating DEM-SPH parameters, our simulations of track unit vehicles on moist sand considering various slip ratios of the track provide insights into the complex dynamics of vehicle-soil interactions. This research highlights the potential of DEM-SPH in delivering precise analyses of tracked vehicle performance on moist sand.

ABSTRACT. Biomimetic robots that adopt structure properties of desert animals have shown improved traversing abilities on granular soils over wheeled vehicles. To conveniently assess traversing abilities of these biomimetic robots, simulations for predicting robot behaviors on soil terrains can be used. Nonetheless, available simulation methods have not yet capable of predicting interactions of between robot foot and soil subjected to arbitrary motions. This work develops EDEM-Recurdyn-Simulink co-simulation method to simulate robot behaviors when traversing granular soil terrain. Using soil simulation parameters determined by experiment tests with a type of Mars soil analog, the traversing behaviors for robot to move forward and turn are simulated. Meanwhile, dynamic forces of robot foots are numerically estimated. The simulated robot behaviors and dynamics are in agreement with experimental tests. Therefore, this paper introduces an effective simulation method to assess traversing abilities of biomimetic robots on granular soil terrains.

ABSTRACT. Real-time wheel-soil models in terramechanics primarily rely on traditional semi-empirical terramechanics models as a foundation. Because of the steady-state assumption that these models are formulated with, their accuracy can suffer in dynamic simulations. Methods such as the Finite Element Method (FEM) and the Discrete Element Method (DEM) can capture transient effects that traditional semi-empirical models cannot, but their computational cost is prohibitive for real-time applications. Using a machine-learning (ML) approach in combination with the semi-empirical models, the authors have previously shown that it is possible to create a hybrid model that can run in real time while capturing transient wheel-soil behaviour. This was achieved by generating training data in the form of DEM simulations and using the resulting forces to compute the difference between the force prediction of the semi-empirical and DEM models. A neural network was trained to predict the difference between these forces. This work builds off the authors’ previous work to expand the scope of this modelling approach so that it can be used in a wider range of practical applications. This is achieved by studying the ideal network input and output parameters and creating DEM simulation scenarios that result in high quality training data. Notably, the idea of having the neural network predict a force per unit width of the wheel is explored as strategy for saving computational resources and creating neural networks that work for a wider range of problems. Detailed description of the creation of training data and network training procedure as well as various examples of the trained network implemented as part of a hybrid wheel-soil model in real-time dynamic simulations will be presented.

ABSTRACT. The distribution of normal stress within soil plays an important role in determining the tire-soil interaction dynamics, affecting tire traction performance, soil compaction, and overall vehicle stability. A comprehensive analysis of the normal stress patterns will allow researchers to refine tire designs, tread characteristics, and inflation pressures for enhanced off-road performance across diverse terrains. This study presents a novel soil instrumentation device designed to measure normal stress distribution in soil following a single pass of an off-road tire. The developed device employs an array of piezoresistive vertical load sensors for providing real-time, accurate data of normal stress distribution beneath the tire during operation. Several tire-soil interaction tests were performed at the Virginia Tech Terramechanics Test Rig to verify the output of the load sensors. The sensor array was embedded in GRC-1 Lunar Soil Simulant at a depth of 150 mm while an off-road tire loaded with a vertical force of 2000 N traversed over the surface of the terrain. The results obtained showed significant differences in pressure patterns under different toe angles, indicating a good correlation between the actual normal stress distribution and the readings from the sensor array.

ABSTRACT. This paper proposes a spatio-temporal analysis of sand-density distribution beneath a traveling wheel based on a particle image velocimetry (PIV) method. With the advancement of image processing technology and the higher resolution of commercially available digital cameras, the PIV methods have been widely used to directly visualize the sand flow field without tracer particles. They can output the numerous sand velocity vectors at a moment in time. By continuously connecting sand flow fields obtained, we can surmise their spatio-temporal changes, though, each flow field is obtained as a spatio-temporally independent feature. Thus, we attempt to propose a PIV-based spatio-temporal analysis accumulating the sand flow vectors, and thereby show a spatio-temporal change of the sand-density distributions beneath the traveling wheel, where the sand density is represented by the relative density. Although the method requires some assumptions, it can represent the spatio-temporal behavior of the sandy terrain, which has not been shown before.

ABSTRACT. This paper proposes a new approach to understanding the wheel-soil interaction, which is an indirect estimation method of soil stress distributions beneath a traveling wheel soil using a photoelastic method. Thus far, in the conventional studies applying the photoelastic method to the wheel-soil terramechanics, the terrain has been emulated by photoelastic disks or plates, which enable visualization of internal stresses in the simulated terrain. In particular, the photoelastic disks have performed visualization of two-dimensional dynamic stress distributions of the simulated granular terrain. With this method, we have visualized and analyzed the dynamic force chain structure of the terrain under different wheel slip conditions. Still, it is difficult for the previous configuration to simulate the dynamic behaviors of natural soil, e.g., compaction, failure, or wheel ruts. Accordingly, achieving both the stress visualization and the dynamic behaviors of soil is a significant challenge to make the photoelastic method more practical. To cope with this challenging issue, we have developed a novel experimental setup consisting of a photoelastic wheel (top layer), soil (middle layer), and a photoelastic plate (bottom layer). By vertically sandwiching the soil between the photoelastic wheel and plate, the soil stresses can be indirectly estimated to satisfy the boundary stress conditions. To achieve this approach, we have conducted calibration tests of the photoelastic wheel and plate, and then identified the force vector and contact patch corresponding to the visualized stresses. In this paper, we demonstrate that it is possible to indirectly estimate how the stress propagates and attenuates in the soil by the proposed method.

ABSTRACT. Tyre models used in soft soil simulation analysis requires tyre parameters in the form of stiffness in multiple directions. These parameters are obtained from measurements on hard terrain as these parameters are a function of the tyre carcass construction. Vehicles used on soft terrain are also used on hard terrain. Many off-road vehicles used in construction, mining, agriculture and forestry use large tyres operating under heavy loads. Testing of large tyres is not a trivial or inexpensive exercise and outdoor testing has limitations on repeatability and load application. This is the case for testing on soft terrain/soil and hard terrain, thus it is preferred to conduct laboratory tests when characterizing tyres as higher loads can be applied and conditions can be controlled. This study investigates the lateral tyre characteristics measured during dynamic/rolling and static tyre tests. Tests are conducted on the actual concrete surface of interest, typically used in field tests with the use of a Dynamic Tyre Test Trailer. Static tests are conducted on the same tyre over representative surfaces in a laboratory with the use of the Static Tyre Test Rig. Multiple tyres are tested and measured tyre characteristics compared. The data can be used to parameterize tyre models of special, large off-road tyres.

ABSTRACT. Most forces on a ground vehicle go through the tyre contact patch. Because the tyre interfaces with the road surface, it is not possible to directly measure the contact patch. This presents a significant challenge in assessing soil characteristics. Predictions of the tread deformation allow estimates of total soil deformations. This can be used to determine elastic and plastic deformation of soil under moving vehicles, giving insight into soil parameters. Soil deformations and parameters are beneficial for vehicle control as well as evaluating the environmental impact. Previous studies utilized digital image correlation techniques to measure tyre deformation on the inner surface of the tyre. This paper investigates the feasibility of developing a measurement system that uses deformations on the inner surface of a tyre to predict the deformation on the outside. The proposed method involves offsetting the inner surface along its normal directions by the tread thickness to obtain the outer surface of the tyre. A 2D proof of concept was developed, showing the ability to predict tread deformation based on inner surface measurements. Reasonable accuracy was achieved, thus confirming the feasibility of predicting tyre tread deformations from inner surface measurements by a geometry offset. The identified errors are deemed acceptable for the given problem, with the errors due to the simplification of the problem proving significantly smaller than measurement-induced errors. A full 3D model was consequently developed to predict tread deformation over the full contact patch region. The findings pave the way for the development of a real-time system to predict soil volume displacement, providing crucial insights for vehicle control and environmental impacts in offroad scenarios.

ABSTRACT. Mobility in the Arctic involves more than operating on only snow and ice surfaces. Vehicles must be able to navigate the during the summer, fall and spring which include both frozen and unfrozen soil surfaces and the transition periods where the ground is freezing in the fall and thawing in the spring. Each of these ground states present a unique challenge to vehicle mobility and must be studied to see fully understand the impacts on mobility. They either get covered by a blanket of snow which insulates the surface slowing and sometimes preventing freezing, or the experimenting team arrives to the test site a day late finding the surface already completely thawed. To solve this problem U.S. Army Corps of Engineers researchers have constructed three frost susceptible soil surfaces for full scale vehicle testing at the Cold Regions Research Laboratory in Hanover, NH. Inside its Frost Effects Research Facility, CRREL researchers may freeze and thaw the soil as needed without relying on the difficult timing of mother nature. The facility allows the user to directly control the freeze and thaw cycles allowing multiple vehicles to be tested on different surfaces over the course of a single winter. This speeds up the research on frost susceptible soils during these difficult transition periods and provides much needed mobility data for predictive modeling and planning purposes. Multiple soil conditions and freeze thaw states have already been studied and this paper will discuss the results for light tracked and wheeled vehicles.

ABSTRACT. During our previous tests, we determined the change in colour due to moisture content in the case of sandy soil of three colours that can be easily distinguished with the naked eye. To characterize the colour, we used the average of the 600 and 700 nm range of the reflectance curve, because in this range the curves are more linear and the points are better separated, i.e. their resolution is better. Also, this wavelength range corresponds to yellow and red colours, which are better absorbed by water molecules, compared to the 400-600 nm range, which includes violet, blue, green and yellow colours. We found that the parameters characteristic of the colour of the examined sandy soils react more sensitively to the influence of soil moisture up to a moisture content of ~4-5%. The created function used the coordinates of this breaking point to determine the moisture content based on reflectance. In our current article, we continue the further processing of the measurement results, as well as carry out new measurements to clarify the functional relationships. We examine the effect of soil porosity and grain density on the reflectance value. This means that in the future, reflectance is given as a function of saturation. The results predict that reflectance is more highly correlated with saturation than with moisture content alone. Another goal is to determine the effect of the initial colour of the sandy soil on the moisture content. As a reference point, we use a characteristic of the reflectance curve for 0% moisture content. We are looking for an answer, how the water content of wet soil can be determined based on the data of dry soil and the reflectance of wet soil.

ABSTRACT. Military maneuvers performed with different types of vehicles often happen outside of the road network. Essentially, areas trafficked by tracked and wheeled machinery can be divided as arable land and natural areas, depending on the impact of human activity. The latter type typically includes natural grasslands, forest and swamp areas. From the perspective of soil strength, the main difference between these land types lies in the compaction state. Compared to natural grasslands, cultivated fields are in a precompacted state. The most well-known mobility performance parameter for military vehicles is the Vehicle Cone Index (VCI). In short, this entails the minimum soil strength required for a successful passage, whereas the soil’s shear strength is determined with a cone penetrometer. There are many reports and scientific papers describing trafficability experiments and validation of VCI as a Go/No-Go indicator for different soils and vehicles. However, the majority of the test results concern natural grassland areas while the applicability of the VCI calculation concept for agriculturally used areas is lacking. The aim of this study is to carry out trafficability tests with wheeled military trucks in order to validate VCI as a mobility performance parameter for cultivated fields. Moreover, changes in soil strength and moisture conditions were monitored for selected field parcels throughout the warmer half of the year. The 70 kN and 125 kN military trucks were used as test vehicles. The soils under observation included sandy loam and clay textures as well as highly organic soils. The paper provides an overview of the experiment’s results and a discussion about trafficability conditions on cultivated areas.

ABSTRACT. Soft soils provide mobility challenges, even for vehicles designed with superior off-road capabilities. When numerous vehicles travel the same path, permanent deformation of the soil can result in rut depths that exceed vehicle ground clearance. These challenges can be overcome by modifying ground conditions to improve bearing capacity or spreading wheel loads over a greater area. Researchers at the U.S. Army Engineer Research and Development Center conducted field testing to quantify performance benefits from using a ground matting system comprised of connected fiberglass panels and designed to improve soft soil vehicle mobility. Soil conditions included soft sand, silt, and peat/sand mixtures with varying soil strength. Test vehicles included wheeled trucks with gross weights of approximately 14,000 lbs. per axle. Performance of the matting system was assessed by the number of allowable vehicle crossings with and without matting present. Results from testing showed that allowable number of vehicles could be increased by a factor of ten on the weakest soils. Data presented herein includes geotechnical site characterization, soil deformation as a function of traffic, and material characteristics for the fiberglass matting system.

ABSTRACT. The mobility of off-road vehicles, especially military vehicles, integrated rescue system vehicles and civilian vehicles in an open terrain where it is necessary to cross rivers, is dependent on the current hydrological conditions and technical parameters of the vehicles. In addition, we can also calculate the ability of drivers to maneuver the vehicle while crossing water courses. The article describes the methodology for determining the parameters of changes in the depth and speed of the water flow and also includes the results of testing the ability to wade through water bodies or overcome them by swimming amphibious vehicles. The mentioned classification of changes in the characteristics of water courses is important for a better understanding of the influence of watercourses on the mobility of vehicles in open terrain outside the road network.

ABSTRACT. Rapid assessment of low volume road surfaces remains a challenge when attempting to forecast allowable vehicle crossings. Variations in moisture content of the soil can greatly affect trafficability, and predictive equations for soil deformation under vehicle loads often have reduced reliability for low-strength materials. Portable tools to characterize soil stiffness and corresponding relationships to load-induced deformation are needed. In this effort, researchers performed comparative testing of multiple rapid assessment tools as potential devices for giving estimations of vehicle trafficability. The test devices included a Clegg hammer and lightweight deflectometer as instruments that measure response from impulse loading. A dynamic cone penetrometer was used as a basis for comparison. Silty sand with and without chemical stabilizers at varying moisture content were used for testing. These soil conditions represented very weak conditions capable of supporting fewer than 50 vehicle passes to moderate strength conditions capable of supporting several thousand vehicle passes. Data from full-scale tests were used to correlate allowable traffic with data obtained from the rapid assessment tools. Recommendations from the effort include ranges of response data to categorize low-volume road surfaces based on their ability to handle ranges of vehicle loadings.

ABSTRACT. The capacity of off-road vehicles to navigate unprepared terrain is determined by the forces imparted by the terrain. The strength of the tire-soil interface is typically lesser than the soil's internal strength. In particular, the interface friction characteristics between the tire and terrain influence the drawbar pull performance and are often overlooked in studies. As a result, accurate experimental determination of the interface friction and its influence on physics-based modeling is the focus of this paper. The tire interaction with fully saturated low plasticity clay (CL) in the plastic state is studied. A large-scale direct shear test between the off-road lugged tire and clay at different contact pressures is conducted to determine the interface failure envelope. As a novel contribution, the tire is sheared in both the longitudinal and transverse orientations. The reliability in the experimental findings is increased by conducting more small-scale direct shear tests with tire rubber samples. The results of the tests are imported into the numerical model consisting of an advanced FE tire interacting with the FE clay. The sensitivity of the drawbar pull and sinkage to friction is studied. Comparative analysis employing different terrain strengths (firm and soft ground) and tire designs (lugged and smooth) are conducted. The advanced FE tire consists of 11 parts including Mooney-Rivlin rubber and orthotropic elastic carcass plies. The research highlights the scenarios where the interface friction influences the drawbar pull, contrasting with situations where the clay deformations dominate. As a conclusion, the proposed methodology shows the importance of accurate interface modeling, providing insights for better correlation between modeling and experimentation.

ABSTRACT. Military and agriculture vehicles alike often experience uncertainties regarding sinkage and traction while traversing non-linear deformable soils. The off-road tires of these vehicles are equipped with stiffer sidewall characteristics to reduce puncture risk, increase load bearing capacity, and improve stability and traction. Since physically testing tires is time-consuming and costly, enhancing the virtual physics-based modeling capabilities can be beneficial for design decisions. Identification of crucial design parameters is essential to ensure that modelers prioritize these parameters to improve the model's accuracy. As a result, a tire sensitivity analysis is proposed to evaluate the tire model's robustness.

The tire used for this study is a lugged bias-ply tire with improved self-cleaning characteristics for muddy terrains. An advanced FE tire model is developed with detailed modeling of the distinct tire parts, including the bead, apex, inner layer, 2-ply carcass, sidewalls, under tread, and lugs. Two types of parameters are of focus: (a) tire operational parameters and (b) tire design parameters. The influence of tire operational parameters, i.e., tire normal load, inflation pressure, and velocity are first conducted and verified with existing studies in the literature. As a novel contribution, the tire design parameters, i.e., tire width, tire diameter, number of lugs, orientation of lugs, and lug size, are also considered. The advanced FE tire is made to negotiate on the rigid ground first and then on low plasticity CL clay to analyze the trends in sensitivities. The results of this sensitivity analysis provide valuable insights into the reliability of the FE tire, enabling more informed conclusions to be drawn from simulated data.

ABSTRACT. Offroad mobility of extraterrestrial exploration rovers can be significantly degraded by vehicle slippage on soft terrain covered with fine powdery sand known as regolith. Comprehensive design analysis and verification of the rover’s mobility have been widely investigated using numerical and experimental approaches. Numerical simulations for the mobility analysis require accurate modeling of the contact forces between the wheel and sand. While several wheel-sand interaction models have been proposed based on the Bekker-Wong-Reece theory, the Dynamic Resistive Force Theory (DRFT) proposed in the early 2020s is a notable approach for its low computational cost and adaptability to high-speed motion. The scaling factor used in DRFT is the only parameter for representing soil-dependent parameters, and its value needs to be empirically tuned to achieve accurate and reliable mobility analysis. Therefore, the authors have integrated a single-wheel test bed into a closed-loop Hardware-In-the-Loop Simulation (HILS) associated with DRFT. This HILS-DRFT experimentally observes the characteristics of the wheel sinking phenomenon in the wheel test bed, and then, the value of the scaling factor is tuned in real-time based on the difference between the observed wheel sinkage and the DRFT simulation. This real-time parameter tuning method accurately reproduces wheel mobility, particularly in its transient states, which were previously unachievable with DRFT alone. The proposed HILS-DRFT with the real-time tuning method will contribute to the efficient and reliable development of rovers.

ABSTRACT. The analysis of tire dynamics is essential in the simulation of vehicle behaviours. The forces exerted on a tire depend on the tire and terrain interaction, and the tire structure directly influences this interaction. Complex models with a high number of degrees of freedom, such as lumped parameter models or finite element models, are typically required to represent tire flexibility appropriately. These models, however, can lead to high computational costs that can be a significant challenge for real-time simulation. In this work, we developed a reduced model for the flexible tire that can represent the effects of tire flexibility with low computational costs in the simulation. By calculating the effective stiffness of a flexible tire model (base model) and augmenting it with a model representing the rigid body motion of the wheel, we could represent the flexibility of the tire more efficiently. The tire deformation affects the tire contact patch size, influencing the traction forces acting on the tire. By having the effective stiffness at hand, we are able to calculate the contact patch size at each instant of time. The traction forces acting on the wheel depend on the size of the contact patch since a larger contact patch would be capable of carrying more tangential load. In order to observe the effect of the contact patch size on the traction forces, we scaled the friction coefficient based on the size of the contact patch. This way, we could take the effect of tire deformation on traction forces into account. Our simulations show efficient real-time performance while maintaining accuracy. By integrating effective stiffness and adjusting friction, we effectively capture tire dynamics, making them a practical solution for diverse vehicle simulation applications.

ABSTRACT. Based on the pioneering work by Bekker, Wong, and Reece, terramechanics models have been used to evaluate the traveling performance of off-road vehicles of a wide range of scales, from small robots to mining dump trucks. Recently, it has also been used to study the performance of lunar and planetary exploration rovers. Meanwhile, multi-body dynamics analysis implementing the terramechanics model is a typical method to study the performance of off-road vehicles on soft ground. This approach can be a powerful tool not only for evaluating the overall vehicle behavior, but also for evaluating safe work plans. The quality of a multi-body dynamics analysis of an off-road vehicle depends on the terramechanics model that describes the interaction between the driving parts and the ground. In this study, the effectiveness of an extended terramechanics model (xTerramechanics model) is demonstrated, which considers soil deformation actions based on cellular automata, for the evaluation of traveling performance of a rigid wheel. First, the results of single-wheel traveling analysis are compared with experimental results under the forced-slip condition, and it is shown that drawbar-pull and sinkage are represented with good accuracy. We then apply the xTerramechanics model under the self-propelled traveling condition at a constant towing load and slope climbing. The model successfully reproduced the well-known “difference in traveling performance depending on traveling conditions.”

ABSTRACT. To promote lunar and planetary exploration missions, the development and operation of exploration rovers is essential. However, it is difficult to evaluate prototypes of wheels and vehicles under environments corresponding to extraterrestrial gravity and atmosphere conditions. Therefore, systematic evaluation of vehicle traveling characteristics using a numerical simulation is required. Under these circumstances, an extended terramechanics model based on cellular automaton that considers terrain surface deformation was proposed by Yokohama National University group, and its validity was confirmed through comparison with the results of traveling experiments using a rigid wheel. In this study, we propose a model that can handle various traveling modes to improve the versatility of numerical analysis simulations that implement extended terramechanics theory. First, we conduct single-wheel experiments using a rigid wheels and obtain data for model verification of both straight traveling and braking. Then, we extend the extended terramechanics model by implementing braking logic, etc., and establish a simulation model for a single rigid wheel using the commercial software package Simscape. In addition, systematic simulations are performed under forced-slip condition in which the angular velocity is fixed, while the translational velocity is varied. Here, the wheel specification and traveling conditions are the same as those of the experiment equipment. Finally, we demonstrate the effectiveness of the proposed model by comparing it with the results of experiment under various conditions.

ABSTRACT. From the terramechanics viewpoint, assessing sinkage and penetration resistance in a deformable terrain is crucial for on-site terrain characterization. Existing devices often encounter limitations when characterizing compacted snow, as documented in the literature. The Clegg impact hammer and the Russian snow penetrometer (utilizing the Rammsonde principle) are extensively employed tools for gauging the penetration resistance of compacted snow. Despite the widespread use of the Clegg impact hammer, there are relatively few studies, primarily conducted by CRREL, focusing on its application in snowy conditions. This study aims to propose enhancements to classical literature formulations of two devices that have been used in compacted snow conditions to evaluate snow characteristics. One such device viz. the Clegg impact hammer has outputs correlated to the evaluation of sinkage and Young’s modulus. Modifications to the classical methodologies have been proposed in this work. The in-house device yielded consistent measurements even when the CTI gauge exhibited variations on a commercial test track used for winter tire evaluations. In this case, formulations have been developed that contribute to a more analytical assessment of the resistance pressure. Additionally, these formulations may aid in evaluating the compacted zone forming in front of the cone as it penetrates deeper into the terrain. The comparison results offer insights into the potential variation in calculating these parameters, with a subsequent discussion of the findings from a physics-based perspective. Future work in this field could entail evaluating the resistance pressure across various terrain types to validate the proposed hypotheses or offer correction factors based on field testing results.

ABSTRACT. In relation to the interaction of the earth's surface with machines and organisms and its engineering applications, there has been a recent increase in interest in the penetration resistive force into granular materials at shallow depths. Recent studies have proposed the model in which penetration resistive forces into dry granular materials have a coefficient dependent on the angle of repose and increase in proportion to the penetration volume. In these studies, the model has been validated for various geometries such as cylinders, cones and spheres. However, for cones, the model has only been validated under conditions of a tip angle close to the angle of repose. In this study, the effect of cone tip angle on penetration resistive force is investigated under several conditions with different angles of repose. This study carries out cone penetration simulations using the discrete element method. For the cone geometry, this study prepares five tip angles ranging from sharp to blunt. The results show that the penetration resistive force for cones with blunt tip angles is much higher than that computed by the model proposed in previous studies. To solve the discrepancy between the model and simulation results, this study modifies the penetration volume by assuming that the stagnant zone formed in front of the cone penetrating the granular material behaves as an effective cone. Thereby, the proposed model can calculate penetration resistive forces more accurately for cones with a wider range of tip angles than in the previous model.

ABSTRACT. Mobility is solved in almost all areas of our lives. The importance of transportation of people and material is increasing every day. Mobility has many subareas, but the authors of this article focus on the mobility of wheeled vehicles in the terrain. Off-road mobility is necessary not only in crisis situations such as road damage, floods, but also in everyday life in agriculture, forestry and it is an essential part of the military sphere as well. In order to avoid getting stuck, it is important to be able to evaluate whether the given route is trafficable or not. Penetrometers are used for such evaluation; they measure the resistance of penetration to the ground - the bearing capacity of the soil. In the Czech army, a telescopic penetrometer is used for this evaluation. Based on many years of field measurements, this instrument measures reliably but it has unreliable evaluation system. Therefore, the authors decided to propose a new evaluation manual. To develop it, a comparison of existing evaluation methods was made. Next, the authors analysed the parameters influencing trafficability of wheeled vehicles and decided which ones to include in the procedure. Based on their previous research, they also decided to include a new driver parameter. New forms and evaluation tables were created. The whole manual was verified by the field measurements. The contribution of the work is the creation of a reliable manual for the evaluation of the trafficability for wheeled vehicles by the telescopic penetrometer, which is already established in the Army of Czech Republic.

ABSTRACT. The paper concerns the performance of ground-based aircraft operating from grass airports, in particular the take-off and landing distances. These performances are strictly dependent on the conditions on the runway, and in the case of a grass runway, they depend on the influence of weather factors. Year-long measurements of the soil cone index were carried out on the runways of 5 grass airports, which are characterized by different ground soils. The measurements were carried out in the period from July 2019 to June 2020. The paper presents a statistical analysis of the measurement results. Generally, it was found that weather factors have a significant impact on the values of the cone index and changes in the CI value reach up to 400% (the highest relative to the lowest). Moreover, it was observed that the type of soil underlying the grass on the runway had a significant impact on the CI index values. The most sensitive ground was marl, with a high chalk content, which constitutes the foundation of one of the tested grass airfields. The results will be used to parameterize and verify the model describing the road wheel and ground performance of the aircraft.

ABSTRACT. When employing semi-empirical terramechanics models, dynamic simulations typically rely on the assumption that all terrain, even rough terrain, can be approximated as a plane in any given simulation time step. This assumption is made necessary by the fact that traditional semi-empirical terramechanics models are formulated to compute wheel-soil interaction forces for a wheel travelling over a flat plane, and adapting these models to compute the reaction forces for a wheel travelling on a sloped plane is straightforward. In many cases, approximating the contact between a wheel and a complex terrain mesh as contact between a wheel and a sloped plane - whose normal direction is determined by the terrain nodes that intersect with the collision geometry of the wheel - can give a reasonable approximation for the contact forces. However, there are conceivable scenarios where important terrain features, such as gaps in the wheel-terrain contact patch, can be overlooked when using this approach. The authors have previously proposed an alternative method for adapting the traditional semi-empirical models for rough terrain without the need to simplify the contact problem to a wheel on a sloped plane. This is accomplished by treating the terrain as a height field and integrating the contact stresses along the wheel's rim by computing the sinkage at each point along the rim based on the un-deformed height field height at that position in space rather than based on the wheel's position relative to an approximate contact plane. In this work, the previous 2-D implementation of the proposed approach is extended to three dimensions to illustrate how it can be used in full-scale dynamic simulations.

ABSTRACT. In off-road wheel/tracked vehicle operations, shearing and compaction actions of- ten occur simultaneously, with both contributing to the total sinkage of the vehicle. The sinkage induced by the vehicle’s compaction is typically referred to as static sinkage. This type of sinkage can be accurately quantified using pressure-sinkage relations such as the Bekker equation or the Wong and Reece equation. The sinkage resulting from the shearing behavior of the wheel and track is referred to as slip-sinkage (or dynamic sinkage in some literature). Compared with the first part, the slip sinkage is usually neglected in simulation. However, neglecting slip-sinkage can lead to inaccuracies in predicting motion resistance and traction, both of which are closely tied to vehicle performance. In past literature, slip sinkage has typically been calculated based on the slip ratio and static sinkage. In this research, the Discrete Element Method (DEM) virtual experiments were conducted to investigate the physical causes of wheel slip sinkage and identify the main contributing factors. The experimental results demonstrated that, in steady-state conditions, slip-sinkage increased with higher slip ratios, while also being influenced by the shear velocity of the wheel. In transient states, the primary contributor appears to be shear displacement rather than the slip ratio. The DEM virtual experimental results were utilized to develop an empirical equation for calculating slip sinkage, incorporating shear displacement, shear velocity, and slip ratio through symbolic regression. The normal force calculation based on the pressure sinkage equation was also modified to better accommodate slip-sinkage, thus improving the accuracy and stability of the simulation.

ABSTRACT. The ability to determine whether a vehicle can traverse a particular winter surface is crucial for planning and executing successful off-road operations in cold regions environments. Various works have detailed strategies and improvements to vehicle mobility models snow and ice surfaces, and others have proposed initial methods to estimate the forces required for vehicles to travel through deep snow exceeding the height of the vehicle bumper or undercarriage These methods demonstrate a strong correlation between their approach and measurements examined in the field. However, they lack sufficient field-based measurements in varied snow conditions to fully capture real-world force estimates.

In this paper, we build upon prior work to develop a preliminary model with the ability to estimate the resistive forces acting on a vehicle traveling through deep snow. We draw upon a wide range of disciplines, discussing the strength of a model based on the present knowledge of snow compression against a methodology based on fluid dynamics techniques. Both models exhibit promising results. However, these results demonstrate a stronger positive correlation using the snow compression model when compared with field-based measurements taken over a variety of winter seasons. These measurements were obtained using a purpose-built plowing apparatus, which simulated a range of bumper ride heights and profiles.

ABSTRACT. A high-fidelity simulation model of tracked vehicles is required for proper prediction of the mobility of tracked vehicles traveling over soft soils. The vehicle components can be modelled using standard tools of multi-body programs. A track model was developed and successfully worked together with Altair's MotionSolve multi-body program. The model based on classic soil mechanics equations. The grousers, which are a significant part of many types of track-links, are taking into the account. The plasticity and viscosity properties of the soil are considered in the model. Verification tests were conducted in an agricultural field in the Jezreel Valley. The tests were run over soils with varying mechanical prop-erties, achieved through irrigation and tillage. The chosen test vehicle was an M113 armoured carrier. Several drawbar pull loads were applied on each soil condition. Reasonable correlation between the tests and simulation results were achieved. However, this model is not efficient and require consumption of large amount of CPU time. On the other hand, the more efficient models are based on simplifying assumptions with lack of accuracy. This work proposes a method for creating an efficient high-fidelity model. This can be implemented by representing the interaction between the track and the ground according to the previous work. The solution for the track-links will be done independently to the solu-tion of the entire vehicle. The solution of the track link will be obtained through a solver that will be developed for this purpose. The solution of the entire vehicle will be done by the solver of the multibody program (Altair MotionSolve). Good correlation between the efficient model and the previous model were achieved.

ABSTRACT. Mobility on terrain is one of the key factors of off road vehicles. In conventional vehicle technology, an experienced driver can assess the obstacle negotiation capability of their vehicle, hence mobility models for manned vehicles usually have smaller resolution and are less detailed. However, unmanned vehicles can only rely on an algorithm both for global path planning and local vehicle control, thus such an algorithm for mobility mapping needs to be more detailed and have higher resolution than convnetional mobility models. In this paper, a method is shown to create a mobility map for a ClearPath A200 small scale test vehicle based on a drone-borne photogrammetry survey of the test area. The vehicle specific mobility model examines the vehicle-terrain interaction from the perspective of obstacle negotiation capability, while also taking into account the soil deformation and the stability of the vehicle. The vehicle-wheel deformations are characterized by a parallel element rheological model. Apart from the perspectives of off road mobility, the computational capacity requirement of the model is also examinded, which is a key factor for real time applications of unmanned vehicles. At the end, a method and test setup is proposed for the validation of the theoretical mobility model via field measuremets.