Design and development of a new transformable wheel used in amphibious all-terrain vehicles (A-ATV)

Conventional ground-wheeled vehicles usually have poor trafficability, low efficiency, a large amount of energy consumption and possible failure when driving on soft terrain. To solve this problem, this paper presents a new design of transformable wheels for use in an amphibious all-terrain vehicle. The wheel has two extreme working statuses: unfolded walking-wheel and folded rigid wheel. Furthermore, the kinematic characteristics of the transformable wheel were studied using a kinematic method. When the wheel is unfolded at walking-wheel status, the displacement, velocity and acceleration of the wheel with different slip rates were analyzed. The stress condition is studied by using a classic soil mechanics method when the transformable wheel is driven on soft terrain. The relationship among wheel traction, wheel parameters and soil deformation under the stress were obtained. The results show that both the wheel traction and trafficability can be improved by using the proposed transformable wheel. Finally, a finite element model is established based on the vehicle terramechanics, and the interaction result between the transformable wheel and elastic–plastic soil is simulated when the transformable wheel is driven at different unfold angles. The simulation results are consistent with the theoretical analysis, which verifies the applicability and effectiveness of the transformable wheel developed in this paper.     

Experimental study of a tracked mobile robot’s mobility performance

This paper proposes an experimental method of predicting the traction performance of a small tracked mobile robot. Firstly, a track-terrain interaction model based on terramechanics is built. Then, an experimental platform of the tracked robot is established, on which the measurement methods of the parameters that influencing the accuracy of the prediction model are introduced and the data post-processing are improved, including drawbar pull, slip ratio, sinkage, track deformation and so on. Based on the experimental data, several key terrain parameters are identified. With the tracked robot platform, the drawbar pull-slip ratio relationship is tested, and the effects on drawbar pull considering different kinds of terrain and the influence of the grousers are analyzed as well. The research results provide a reference for the experimental study on the traction performance of small tracked robots.     

Terrain classification using intelligent tire

A wheeled ground robot was designed and built for better understanding of the challenges involved in utilization of accelerometer-based intelligent tires for mobility improvements. Since robot traction forces depend on the surface type and the friction associated with the tire-road interaction, the measured acceleration signals were used for terrain classification and surface characterization. To accomplish this, the robot was instrumented with appropriate sensors (a tri-axial accelerometer attached to the tire innerliner, a single axis accelerometer attached to the robot chassis and wheel speed sensors) and a data acquisition system. Wheel slip was measured accurately using encoders attached to driven and non-driven wheels. A fuzzy logic algorithm was developed and used for terrain classification. This algorithm uses the power of the acceleration signal and wheel slip ratio as inputs and classifies all different surfaces into four main categories; asphalt, concrete, grass, and sand. The performance of the algorithm was evaluated using experimental data and good agreements were observed between the surface types and estimated ones.     

The effect of velocity on rigid wheel performance

A simulation model to predict the effect of velocity on rigid-wheel performance for off-road terrain was examined. The soil–wheel simulation model is based on determining the forces acting on a wheel in steady state conditions. The stress distribution at the interface was analyzed from the instantaneous equilibrium between wheel and soil elements. The soil was presented by its reaction to penetration and shear. The simulation model describes the effect of wheel velocity on the soil–wheel interaction performances such as: wheel sinkage, wheel slip, net tractive ratio, gross traction ratio, tractive efficiency and motion resistance ratio. Simulation results from several soil-wheel configurations corroborate that the effect of velocity should be considered. It was found that wheel performance such as net tractive ratio and tractive efficiency, increases with increasing velocity. Both, relative wheel sinkage and relative free rolling wheel force ratio, decrease as velocity increases. The suggested model improves the performance prediction of off-road operating vehicles and can be used for applications such as controlling and improving off-road vehicle performance.     

Machine learning in planetary rovers: A survey of learning versus classical estimation methods in terramechanics for in situ exploration

For the design of space missions in the Moon and planets, analysis of mobility in robots is crucial and poor planning has led to abortion of missions in the past. To mitigate the risk of mission failure, improved algorithms relying intrinsically on fusing visual odometry with other sensory inputs are developed for slip detection and navigation. However, these approaches are significantly expensive computationally and difficult to meet for future space exploration robots. Hence, today the central question in the field is how to develop a novel framework for in situ estimation of rover mobility with available space hardware and low-computational demanding terramechanics predictors. Ranging from pure simulations up to experimentally validated studies, this paper surveys dozens of existing methodologies for detection of vehicle motion performance (wheel forces and torques), surface hazards (slip-sinkage) and other parameters (soil strenght constants) using classical terramechanics maps, and compare them with novel approaches introduced by machine learning, allowing to establish future directions of research towards distributed exteroceptive and proprioceptive sensing for visionless exploration in dynamic environments. To avoid making it challenging to collect all relevant studies expeditiously, we propose a global classification of terramechanics according most common practices in the field, allowing to form an structured framework that condense most works in the domain within three estimator categories (direct/forward or inverse terramechanics, and slip estimators). Likewise, from the experiences collected in previous MER (Mars Exploration Rover) missions, five overlooked problems are documented that will need to be addressed in next generation of planetary vehicles, along three research questions and few hypothesis that will pave the road towards future applications of machine learning-based terramechanics.     

Tractive performance analysis of a lugged wheel by open-source 3D DEM software

Open-source software (OSS) is free to use and has accessible source codes, thus, it can be modified by various users. By using OSS, it is possible to easily and economically develop a target program for interaction studies in terramechanics. Yet Another Dynamic Engine (YADE) is an OSS for the 3D discrete element method (DEM), but its applicability to various contact interaction problems in terramechanics is not well-known. To investigate the applicability of YADE in terramechanics, the tractive performance of a lugged wheel was analyzed in this study. An idea of a proportional-integral-differential control model was applied to realize the constant rotation of the wheel in YADE. Our previous experiments on the locomotion of a small lugged wheel on a lunar-soil simulant were analyzed by YADE, and the results were found to be qualitatively similar to the obtained experimental results when considering the effects of the lug height, lug thickness, lug number, and wheel diameter. By applying a quasi-2D analysis with the same soil bin width and wheel width, the computational load of 3D DEM by YADE can be reduced up to 36.8% with similar net traction behavior against the wheel slip in a 3D analysis.     

Multi-objective traction optimization of vehicles in loose dry sand using the generalized reduced gradient method

A work optimization strategy is combined with algorithms within the vehicle-terrain interface (VTI) model to maximize the traction of a four-wheel vehicle operating on loose dry sand. The optimization model distributes traction among the steered and non-steered wheels with the work optimum coefficient (WOC) of each wheel treated as an independent design objective. Drawbar pull (DBP), motion resistance (MR), longitudinal traction coefficient (LTC), lateral force coefficient (LFC), tire deflection, and wheel slip are key parameters that appear in the VTI model for traction performance analysis. The analysis includes wheels of different diameters, widths, heights, and inflation pressures, under variable wheel slips. A multi-objective optimization problem is formulated over a thirteen-dimensional search space bounded by eight design constraints. The generalized reduced gradient method is used to predict optimal values of the design variables as well as ground and traction parameters such as DBP, MR, LTC, and LFC for maximum slope climbing efficiency. The WOCs are maximized for lateral slip angles between 0° and 24° to find a set of Pareto optimal solutions over a wide range of weight factors. A method to apply the optimization results for predicting vehicle performance and traction control on dry sand is presented and discussed.     

Stability and Bifurcation of Longitudinal Vehicle Braking

The longitudinal braking dynamics of a two-wheel vehicle model on an incline are considered using techniques from nonlinear dynamics. The model is planar and incorporates the coupled dynamics of two independently braked wheels and the vehicle body, and takes into account the slip dynamics of each wheel. By using the wheel slip values and the vehicle speed as dynamic states, it is shown that the qualitative behavior of the system can be completely captured by studying a relatively simple phase plane problem described in terms of the slip values. A systematic bifurcation analysis is carried out in which the brake torques of the two wheels are varied, and it is shown how the system transitions from stable braking to the possibility of lockup in one or both wheels, to guaranteed lockup in both wheels. In this manner a quite complete picture of the dynamic behavior is obtained as a function of the two brake torques, including regions with multiple possible steady braking outcomes, depending on the initial conditions. This analysis provides new insights into the dynamics of vehicle braking, and it provides a correction to the standard result for the critical values of the brake torques at which the wheels undergo lockup. This approach may also prove useful for evaluating brake proportioning schedules, or for investigating anti-lock braking systems and other methods of traction control.     

Wheel slip measurement in 2WD tractor

A microcontroller-based slip sensor was developed for a 2WD tractor to indicate slip values during on-farm use. The ‘zero condition’ considered for the development of slip sensor was – tractor supplied with a driving torque to propel any device across a tarmacadam surface while delivering zero net traction (self-propelled condition). This sensor comprised of four components: power supply; sensing of throttle position, gear position, and wheel rpm; processing of collected data; and display unit. Power was taken from the tractor battery. Rotary potentiometer and proximity switches were installed on the tractor to measure throttle position and wheel revolution, respectively. The performance of developed slip sensor was evaluated both on tarmacadam surface as well as in the field. The variations between indicated and actual slip were found to be within 0–5% for both the surfaces, thus indicating the accuracy of slip measurement by the developed slip sensor.     

Development of in-wheel sensor system for accurate measurement of wheel terrain interaction characteristics

Planetary rovers need high mobility on a rough terrain such as sandy soil, because such a terrain often impedes the rover mobility and causes significant wheel slip. Therefore, the accurate estimation of wheel soil interaction characteristics is an important issue. Recent studies related to wheel soil interaction mechanics have revealed that the classical wheel model has not adequately addressed the actual interaction characteristics observed through experiments. This article proposes an in-wheel sensor system equipped with two sensory devices on the wheel surface: force sensors that directly measure the force distribution between the wheel and soil and light sensors that accurately detect the wheel soil surface boundary line. This sensor design enables the accurate measurement of wheel terrain interaction characteristics such as wheel force distribution, wheel–soil contact angles, and wheel sinkage when the powered wheel runs on loose sand. In this article, the development of the in-wheel sensor system is introduced along with its system diagram and sensor modules. The usefulness of the in-wheel sensor system is then experimentally evaluated via a single wheel test bench. The experimental results confirm that explicit differences can be observed between the classical wheel model and practical data measured by the in-wheel sensor system.     

Discrete element method-based studies on dynamic interactions of a lugged wheel with granular media

Terramechanics plays an important role in determining the design and control of autonomous robots and other vehicles that move on granular surfaces. Traction capabilities, slippage, and sinkage of a robot are governed by the interaction of a robot’s appendage with the operating terrain. It is important to understand how the terrain flows under this appendage during such an interaction. In this work, dynamics of soil performance and locomotion performance of a lugged wheel travelling on soft soil are numerically investigated. Studies are conducted with a two-dimensional model by using the discrete element method to analyze the interactions between a lugged wheel and the soil. The soil performance is studied by examining the force distribution and evolution of force networks during the course of the wheel travel. For two different control modes, namely, slip-based wheel control and angular velocity-based wheel control, the performance parameters such as, sinkage, traction, traction efficiency, and power consumption of the wheel are compared for various wheel configurations. The findings of this work are expected to be useful for optimal design and control of the lugged wheel travelling on deformable surfaces.     

Review of terramechanics models and their applicability to real-time applications

ISTVS embarked on a project in 2016 that aims at updating the current ISTVS standards related to nomenclature, definitions, and measurement techniques for modelling, parameterizing, and, respectively, testing and validation of soft soil parameters and vehicle running gear-terrain interaction. As part of this project, a comprehensive literature review was conducted on the parameterization of fundamental terramechanics models. Soil parameters of the empirical models to assess off-road vehicle mobility, and parameters of the models to characterize the response of the terrain interacting with running gears or plates from the existing terramechanics literature and other researchers’ reports were identified. This review documents and summarizes the modelling approaches that may be applicable to real-time applications of terramechanics in simulation, as well as in controller design.     

Longitudinal skid model for wheels of planetary rovers based on improved wheel sinkage considering soil bulldozing effect

To successfully deploy a wheeled mobile robot on deformable rough terrains, the wheel-terrain interaction mechanics should be considered. Skid terramechanics is an essential part of the wheel terramechanics and has been studied by the authors based on the wheel sinkage obtained using a linear displacement sensor that does not consider soil bulldozing effect. The sinkage measured by a newly developed wheel via detecting the entrance angle is about 2 times of that measured by the linear displacement sensor. On the basis of the wheel sinkage that takes the soil bulldozing effect into account, a linear function is proposed to the sinkage exponent. Soil flow in the rear region of wheel-soil interface is considered in the calculation of soil shear displacement, and its average velocity is assumed to be equal to the tangential velocity component of the transition point of shear stress. To compute the normal stress in the rear region directly, the connection of the entrance and leaving points is supposed as the reference of wheel sinkage. The wheel performance can be accurately estimated using the proposed model by comparing the simulation results against the experimental data obtained using two wheels and on two types of sands.     

Slip-based experimental studies of a vehicle interacting with natural snowy terrain

As longitudinal slip affects vehicle–pavement interactions on roads and hard surfaces, so too does it play an important role in interactions between vehicles and soft terrains, including snow. Although many slip-based models have been developed recently for tire–snow interactions (e.g., [1] and references cited therein), these models have only been partially validated, due to a lack of relevant experimental data. This paper presents comprehensive data from tests that were performed using a newly-developed test vehicle traversing natural snowy terrain, over a wide range of values for longitudinal slip, vertical load and torque via an effective accelerate/brake maneuver. Drawbar pull, motion resistance, wheel states and tire stiffness were presented as a function of slip; tire sinkage was obtained using a laser profilometer; strength and depth of snow were found using a snow micropenetrometer. The effects of the rear tire going over snow compacted by the front tire were also studied. The maximum traction force normalized by the vertical load is found to be ≈0.47, maximum motion resistance normalized by the vertical load is ≈0.4. Comparison of the trend and order-of-magnitude of test results with those from existing slip-based numerical model [1] shows good comparison in motion resistance, tire sinkage, and longitudinal stiffness, but indicates that a better traction model is needed to improve the comparison.     

Shape effects of wheel grousers on traction performance on sandy terrain

This paper presents the effects of different wheel grouser shapes on the traction performance of a grouser wheel traveling on sandy terrain. Grouser wheels are locomotion gears that allow small and lightweight exploration rovers to traverse on the loose sand on extraterrestrial surfaces. Although various grouser shapes have been analyzed by some research groups, a more synthetic and direct comparison of possible grousers is required for practical applications. In this study, we developed a single wheel testbed and experimentally investigated the effects of four grouser shapes (parallel, slanted, V-shaped, and offset V-shaped) on the traction performance of linear movement on flat sand. The wheel slip, sinkage, traction and side force acting on the wheel axle, the wheel driving torque, and the efficiency of each wheel were examined. Thereafter, the effects on the lateral slope traversability of a small and lightweight four-wheeled rover with different grouser shapes were also examined. The traversability experiment demonstrated the vehicle mobility performance in order to contribute to the design optimization of rover systems. These experimental results and their comparisons suggested that, of the shapes studies herein, the slanted shape was the optimal grouser design for use in wheeled rovers on lunar and planetary soil.     

Torque distribution influence on tractive efficiency and mobility of off-road wheeled vehicles

Off-road vehicle performance is strongly influenced by the tire-terrain interaction mechanism. Soft soil reduces traction and significantly modifies vehicle handling; therefore tire dynamics plays a strong role in off-road mobility evaluation and needs to be addressed with ad-hoc models. Starting from a semi-empirical tire model based on Bekker–Wong theory, this paper, analyzes the performance of a large four wheeled vehicle driving on deformable terrain. A 14 degree of freedom vehicle model is implemented in order to investigate the influence of torque distribution on tractive efficiency through the simulation of front, rear, and all wheel drive configuration. Results show that optimal performance, regardless vertical load distribution, is achieved when torque is biased toward the rear axle. This suggests that it is possible to improve tractive efficiency without sacrificing traction and mobility. Vehicle motion is simulated over dry sand, moist loam, flat terrain and inclined terrain.     

Analysis of Mars Exploration Rover wheel mobility processes and the limitations of classical terramechanics models using discrete element method simulations

A previous three-dimensional discrete element method (DEM) model of Mars Exploration Rovers (MERs) wheel mobility demonstrated agreement with test data for wheel drawbar pull and sinkage for wheel slips from 0.0 to 0.7. Here, results from the previous model are compared with wheel mobility data for non-MER wheels that cover the range of wheel slip from 0.0 to 1.0. Wheel slips near 1.0 are of interest for assessing rover mobility hazards. DEM MER wheel model predictions show close agreement with weight-normalized wheel drawbar pull data from 0.0 to 0.99 wheel slip and show a similar trend for wheel sinkage. The nonlinear increase in MER wheel drawbar pull and sinkage for wheel slips greater that 0.7 is caused by development of a tailings pile behind the wheel as it digs into the regolith.Classical terramechanics wheel mobility equations used in the ARTEMIS MER mobility model are inaccurate above wheel slips of 0.6 as they do not account for the regolith tailings pile behind the wheel. To improve ARTEMIS accuracy at wheel slips greater that 0.6 a lookup table of drawbar pull, wheel torque, and sinkage derived from DEM mobility simulations can be substituted for terramechanics equation calculations.     

Next-generation NATO reference mobility model complex terramechanics – Part 1: Definition and literature review

The US army along with NATO member and partner nations’ militaries need an accurate software tool for predicting ground vehicle mobility (such as speed-made-good and fuel-consumption) on world-wide terrains where military vehicles may be required to operate. Currently, the NATO Reference Mobility Model (NRMM) is the only NATO recognized tool for assessing ground vehicle mobility. NRMM was developed from the 1960s to the 1980s and relies on steady-state empirical formulas which may not be accurate for new military ground vehicles. A NATO research task group (RTG-248) was established from 2016 to 2018 to develop the NG-NRMM (next-generation NRMM) software tool requirements and an NG-NRMM prototype which uses high-fidelity “simple” or “complex” terramechanics models for the terrain/soil along with modern 3D multibody dynamics software tools for modeling the vehicle. NG-NRMM Complex Terramechanics (CT) models are those that utilize full 3D soil models capable of predicting the 3D soil reaction forces on the vehicle surfaces (including tires, tracks, legs, and under body) and the 3D flow and deformation of the soil including both elastic and plastic deformation under any 3D loading condition. In Part 1 of this paper, an overview of the full spectrum of terramechanics models from the highest fidelity to the lowest fidelity is presented along with a literature review of CT ground vehicle mobility models.     

Stochastic modeling of tire–snow interaction using a polynomial chaos approach

Developing accurate models to simulate the interaction between pneumatic tires and unprepared terrain is a demanding task. Such tire–terrain contact models are often used to analyze the mobility of a wheeled vehicle on a given type of soil, or to predict the vehicle performance under specified operational conditions (as related to the vehicle and tires, as well as to the running support). Due to the complex nature of the interaction between a tire and off-road environment, one usually needs to make simplifying assumptions when modeling such an interaction. It is often assumed that the tire–terrain interaction can be captured using a deterministic approach, which means that one assumes fixed values for several vehicle or tire parameters, and expects exact responses from the system. While this is rarely the case in real life, it is nevertheless a necessary step in the modeling process of a deterministic framework. In reality, the external excitations affecting the system, as well as the values of the vehicle and terrain parameters, do not have fixed values, but vary in time or space. Thus, although a deterministic model may capture the response of the system given one set of deterministic values for the system parameters, inputs, etc., this is in fact only one possible realization of the multitude of responses that could occur in reality. The goal of our study is to develop a mathematically sound methodology to improve the prediction of the tire–snow interaction by considering the variability of snow depth and snow density, which will lead to a significantly better understanding and a more realistic representation of tire–snow interaction. We constructed stochastic snow models using a polynomial chaos approach developed at Virginia Tech, to account for the variability of snow depth and of snow density. The stochastic tire–snow models developed are based on the extension of two representative deterministic tire–snow interaction models developed at the University of Alaska, including the pressure–stress deterministic model and the hybrid (on-road extended for off-road) deterministic model. Case studies of a select combination of uncertainties were conducted to quantify the uncertainties of the interfacial forces, sinkage, entry angle, and the friction ellipses as a function of wheel load, longitudinal slip, and slip angle. The simulation results of the stochastic pressure–stress model and the stochastic hybrid model are compared and analyzed to identify the most convenient tire design stage for which they are more suitable. The computational efficiency of the two models is also discussed.     

The problems of multi-axle vehicle drives

The multi-axle drive provides for maximum traction of the vehicle. However, this mode of driving, apart from the complicated design of the driving mechanism, is connected with major problems due to the so-called kinematic discrepancy encountered by a vehicle under operation. In certain conditions, the kinematic discrepancy results in the phenomenon of so-called circulating power which causes, first of all, additional loading on the drive system, then increased energy loss and tyre wear and, in the case of an articulated frame-steer vehicle, the substantial growth of steering resistance.The paper reviews some examples of results of the studies on wheel slips, load on the drive system, and on the 4-wheel-drive articulated frame-steer vehicle's steering gear. The author also discusses a new concept for determining kinematic discrepancy. The suggested equations, when applying the present mathematical models which express the dependence of a tyre's slip rate on longitudinal forces, permit to solve the statically undeterminable problem concerning the distribution of the wheel's longitudinal forces in any multi-axle vehicle.