Terrain Traversability Research Articles

Semantic Terrain Traversability Analysis Based on Deep Learning Aimed at Planetary Rover Navigation

Semantic Terrain Traversability Analysis Based on Deep Learning Aimed at Planetary Rover Navigation

Learning Terrain Traversability for a Mobile Robot based on Information Fusion

In this work, we propose an approach to determine terrain traversability for a car-like robot. Our approach has two main modules: a neural network classifier that makes use of sensors' readings to assign traversability levels to control inputs of the robot, and a second neural network that, based on the outputs of the first network, mimics the control selection performed by a human driver. The approach incorporates sensor fusion from a variety of sources to enhance the traversability estimation, and it is trained employing a semi-supervised learning scheme with examples resulting from the interaction of the car with the environment. This semi-supervised scheme avoids exhausting manual labeling and is built on the premise that there is a correlation between the terrain traversability and the required and observed behaviors of the vehicle. The method is validated with data obtained from a physical electric car.

Autonomous Vehicles Traversability Mapping Fusing Semantic–Geometric in Off-Road Navigation

This paper proposes an evaluating and mapping methodology of terrain traversability for off-road navigation of autonomous vehicles in unstructured environments. Terrain features are extracted from RGB images and 3D point clouds to create a traversal cost map. The cost map is then employed to plan safe trajectories. Bayesian generalized kernel inference is employed to assess unknown grid attributes due to the sparse raw point cloud data. A Kalman filter also creates density local elevation maps in real time by fusing multiframe information. Consequently, the terrain semantic mapping procedure considers the uncertainty of semantic segmentation and the impact of sensor noise. A Bayesian filter is used to update the surface semantic information in a probabilistic manner. Ultimately, the elevation map is utilized to extract geometric characteristics, which are then integrated with the probabilistic semantic map. This combined map is then used in conjunction with the extended motion primitive planner to plan the most effective trajectory. The experimental results demonstrate that the autonomous vehicles obtain a success rate enhancement ranging from 4.4% to 13.6% and a decrease in trajectory roughness ranging from 5.1% to 35.8% when compared with the most developed outdoor navigation algorithms. Additionally, the autonomous vehicles maintain a terrain surface selection accuracy of over 85% during the navigation process.

Modeling the Enceladus dust plume based on in situ measurements performed with the Cassini Cosmic Dust Analyzer

We analyzed data recorded by the Cosmic Dust Analyzer on board the Cassini spacecraft during Enceladus dust plume traversals. Our focus was on profiles of relative abundances of grains of different compositional types derived from mass spectra recorded with the Dust Analyzer subsystem during the Cassini flybys E5 and E17. The E5 profile, corresponding to a steep and fast traversal of the plume, has already been analyzed. In this paper, we included a second profile from the E17 flyby involving a nearly horizontal traversal of the south polar terrain at a significantly lower velocity. Additionally, we incorporated dust detection rates from the High Rate Detector subsystem during flybys E7 and E21. We derived grain size ranges in the different observational data sets and used these data to constrain parameters for a new dust plume model. This model was constructed using a mathematical description of dust ejection implemented in the software package DUDI. Further constraints included published velocities of gas ejection, positions of gas and dust jets, and the mass production rate of the plume. Our model employs two different types of sources: diffuse sources of dust ejected with a lower velocity and jets with a faster and more colimated emission. From our model, we derived dust mass production rates for different compositional grain types, amounting to at least 28 kg/s. Previously, salt-rich dust was believed to dominate the plume mass based on E5 data alone. The E17 profile shows a dominance of organic-enriched grains over the south polar terrain, a region not well constrained by E5 data. By including both E5 and E17 profiles, we find the salt-rich dust contribution to be at most 1<!PCT!> by mass. This revision also results from an improved understanding of grain masses of various compositional types that implies smaller sizes for salt-rich grains. Our new model can predict grain numbers and masses for future mission detectors during plume traversals.

An Integrated Route and Path Planning Strategy for Skid–Steer Mobile Robots in Assisted Harvesting Tasks with Terrain Traversability Constraints

This article presents a combined route and path planning strategy to guide Skid–Steer Mobile Robots (SSMRs) in scheduled harvest tasks within expansive crop rows with complex terrain conditions. The proposed strategy integrates: (i) a global planning algorithm based on the Traveling Salesman Problem under the Capacitated Vehicle Routing approach and Optimization Routing (OR-tools from Google) to prioritize harvesting positions by minimum path length, unexplored harvest points, and vehicle payload capacity; and (ii) a local planning strategy using Informed Rapidly-exploring Random Tree (IRRT*) to coordinate scheduled harvesting points while avoiding low-traction terrain obstacles. The global approach generates an ordered queue of harvesting locations, maximizing the crop yield in a workspace map. In the second stage, the IRRT* planner avoids potential obstacles, including farm layout and slippery terrain. The path planning scheme incorporates a traversability model and a motion model of SSMRs to meet kinematic constraints. Experimental results in a generic fruit orchard demonstrate the effectiveness of the proposed strategy. In particular, the IRRT* algorithm outperformed RRT and RRT* with 96.1% and 97.6% smoother paths, respectively. The IRRT* also showed improved navigation efficiency, avoiding obstacles and slippage zones, making it suitable for precision agriculture.

Terrain traversability prediction through self-supervised learning and unsupervised domain adaptation on synthetic data

Terrain traversability estimation is a fundamental task for supporting robot navigation on uneven surfaces. Recent learning-based approaches for predicting traversability from RGB images have shown promising results, but require manual annotation of a large number of images for training. To address this limitation, we present a method for traversability estimation on unlabeled videos that combines dataset synthesis, self-supervision and unsupervised domain adaptation. We pose the traversability estimation as a vector regression task over vertical bands of the observed frame. The model is pre-trained through self-supervision to reduce the distribution shift between synthetic and real data and encourage shared feature learning. Then, supervised training on synthetic videos is carried out, while employing an unsupervised domain adaptation loss to improve its generalization capabilities on real scenes. Experimental results show that our approach is on par with standard supervised training, and effectively supports robot navigation without the need of manual annotations. Training code and synthetic dataset will be publicly released at: https://github.com/perceivelab/traversability-synth.

Terrain-Shape-Adaptive Coverage Path Planning With Traversability Analysis

Coverage path planning (CPP) is in great demand with applications in agriculture, mining, manufacturing, etc. Most research in this area focused on 2D CPP problems solving the coverage problem with irregular 2D maps. Comparatively, CPP on uneven terrains is not fully solved. When there are many slopy areas in the working field, it is necessary to adjust the path shape and make it adapt to the 3D terrain surface to save energy consumption. This article proposes a terrain-shape-adaptive CPP method with three significant features. First, the paths grow by themselves according to the local terrain surface shapes. Second, the growth rule utilizes the 3D terrain traversability analysis, which makes them automatically avoid entering hazardous zones. Third, the irregularly distributed paths are connected under an optimal sequence with an improved genetic algorithm. As a result, the method can provide an autonomously growing terrain-adaptive coverage path with high energy efficiency and coverage rate compared to previous research works. It is demonstrated on various maps and is proven to be robust to terrain conditions.

A Survey on Path Planning for Autonomous Ground Vehicles in Unstructured Environments

Autonomous driving in unstructured environments is crucial for various applications, including agriculture, military, and mining. However, research in unstructured environments significantly lags behind that in structured environments, mainly due to the challenges posed by harsh environmental conditions and the intricate interactions between vehicles and terrains. This article first categorizes unstructured path planning into hierarchical and end-to-end approaches and then the special parts compared to structured path planning are emphatically reviewed, such as terrain traversability analysis, cost estimation, and terrain-dependent constraints. This article offers a comprehensive review of the relevant factors, vehicle–terrain interactions, and methods of terrain traversability analysis. The estimation methods of safety cost, energy cost, and comfort cost are also emphatically summarized. Moreover, the constraints caused by the limits of terrains and vehicles are discussed. The applications of algorithms in recent articles for path planners are reviewed. Finally, crucial areas requiring further research are analyzed in unstructured path planning.

Deep Semantic Segmentation for Identifying Traversable Terrain in Off-Road Autonomous Driving

Deep Semantic Segmentation for Identifying Traversable Terrain in Off-Road Autonomous Driving

Dependence of Tractor Traction Properties on Agricultural Machinery Connections

Tractor traction properties depend on many factors including engine power, operating weight and tractor weight distribution across axles, dimensions, profile, gears range and gear count in the gearbox. Moreover, the tractor technical condition plays a vital role, encompassing aspects from the fuel system affecting the engine power output, to the chassis, influencing terrain traversability, maneuverability, and soil deformation. (Research purpose) The paper addresses several research problems. To investigate the impact of the agricultural machinery connections with the tractor on the tractor traction properties, and determine the impact of agricultural machinery weight, and grip width on the main technical and economic parameters. To analyze how geometrical parameters of tractors and agricultural machinery correlate with soil reactions under the machine support wheels as well as the front and rear wheels of the tractor. To determine the influence of the design characteristics of cultivators with varying working widths on the main operational properties of the machine-tractor unit. (Materials and methods) The study involves examining the operation of the Belarus 82.1 tractor with cultivators KPS-4 and KPS-5 from different manufacturers. Various methods were employed, including calculation, calculation-construction, calculation-graphic analysis, economic-mathematical techniques, along with comparison, measurement, and description. (Results and discussion) This study identified indicators impacting the load redistribution transmitted to the tractor from mounted cultivators. Based on these indicators, alterations in soil reactions under cultivator support wheels and the tractor’s rear wheels were deduced. (Conclusions) The calculations on the effect of mounted cultivators with different working widths (different operating weight) on the redistribution of soil reactions between axles, enabled the determination of modifications in tractor traction force, machine-tractor unit productivity, and specific fuel consumption.

TTA-GEP: Terrain Traversability Analysis with Geometry and Environmental Perception for the Path Planning of Planetary Rovers

Terrain traversability analysis (TTA), the key to the navigation of planetary rovers, is significant to the safety of the rover. Therefore, owing to its complexity, the Martian terrain is worth analysing comprehensively based on the terrain variability and hazard level. In this work, we propose a novel method for terrain traversability analysis for the path planning of planetary rovers by integrating Martian terrain geometry features with terrain semantic information, which includes geometry and environmental perception (GEP). Specifically, we deploy semantic segmentation to classify common terrain types, such as rocks, bedrocks, and sand, obtaining semantic information as one part of terrain traversability analysis at the same time. Simultaneously, the point cloud is generated by using binocular images from the planetary rover navigation camera (Navcam) to construct a 2.5D elevation map of the environment to analyse the geometric characteristics of the terrain. Besides, we implement path planning based on the results of TTA-GEP. Overall, our proposed method improves the performance of the terrain traversability analysis and reduces the risk of planetary rovers while detecting in an unstructured environment.

Eurasian woodcock (Scolopax rusticola) in intensively managed Central European forests use large home ranges with diverse habitats

Modern management of Europe's forests should strive for a delicate balance between efficient production, which is their primary purpose, and forest management practices that conserve and support biodiversity. This requires detailed knowledge of the ecology of forest-dwelling species, including their site-specific behavior and habitat preferences. One such species is the Eurasian woodcock (Scolopax rusticola), an important game bird that breeds throughout the Palearctic region. Despite its importance, our knowledge of woodcock ecology is limited, and collecting reliable data is challenging due to the species' cryptic behavior. Here, we used high-precision GPS-GSM telemetry to determine home range size and to investigate habitat selection in male Eurasian woodcocks inhabiting Central European forests. Using Bayesian mixed-effects models corrected for spatial autocorrelation, we compared several habitat characteristics (including dominant tree and herb species) of sites where woodcocks occurred with randomly selected sites within their home ranges. Our results show that woodcock in Central Europe move over much larger home ranges during the breeding season (median 7.26 km2) than previously reported in other parts of their breeding range, although they regularly use only small part of this large area (median 1.09 km2). Within their home range, woodcocks used a wide range of habitats, including intensively managed conifer monocultures, but preferred deciduous forests, particularly those dominated by European white birch (Betula pendula). Specifically, 24 % of the plots with woodcock were dominated by birch, compared to only 12 % of the randomly selected plots. Importantly, regardless of dominant tree species, woodcocks preferred sites with dense vegetation cover, but good terrain traversability. During the day, woodcocks preferred sites with dense canopy cover, while at night they preferred more open areas.However, in contrast to previously published results elsewhere, these open areas were more likely to be sparse forest (73%) than grasslands or non-forest cropland (24%), and were characterized by less continuous canopy cover, with a median of only 40%.Our study highlights the importance of specific microhabitats, such as sites with dominant representation of commercially non-target broadleaved species (birch), or sites with sparse canopy cover but denser understorey vegetation. Appropriate management practices that promote such specific microhabitats would have a positive impact on populations of woodcock and other forest species inhabiting commercial forests in Central Europe.

Learning Off-Road Terrain Traversability With Self-Supervisions Only

Estimating the traversability of terrain should be reliable and accurate in diverse conditions for autonomous driving in off-road environments. However, learning-based approaches often yield unreliable results when confronted with unfamiliar contexts, and it is challenging to obtain manual annotations frequently for new circumstances. In this paper, we introduce a method for learning traversability from images that utilizes only self-supervision and no manual labels, enabling it to easily learn traversability in new circumstances. To this end, we first generate self-supervised traversability labels from past driving trajectories by labeling regions traversed by the vehicle as highly traversable. Using the self-supervised labels, we then train a neural network that identifies terrains that are safe to traverse from an image using a one-class classification algorithm. Additionally, we supplement the limitations of self-supervised labels by incorporating methods of self-supervised learning of visual representations. To conduct a comprehensive evaluation, we collect data in a variety of driving environments and perceptual conditions and show that our method produces reliable estimations in various environments. In addition, the experimental results validate that our method outperforms other self-supervised traversability estimation methods and achieves comparable performances with supervised learning methods trained on manually labeled data.

TNES: terrain traversability mapping, navigation and excavation system for autonomous excavators on worksite

TNES: terrain traversability mapping, navigation and excavation system for autonomous excavators on worksite

Terrain Characterization via Machine vs. Deep Learning Using Remote Sensing

Terrain traversability is critical for developing Go/No-Go maps for ground vehicles, which significantly impact a mission’s success. To predict the mobility of terrain, one must understand the soil characteristics. In-situ measurements performed in the field are the current method of collecting this information, which is time-consuming, costly, and can be lethal for military operations. This paper investigates an alternative approach using thermal, multispectral, and hyperspectral remote sensing from an unmanned aerial vehicle (UAV) platform. Remotely sensed data combined with machine learning (linear, ridge, lasso, partial least squares (PLS), support vector machines (SVM), and k nearest neighbors (KNN)) and deep learning (multi-layer perceptron (MLP) and convolutional neural network (CNN)) are used to perform a comparative study to estimate the soil properties, such as the soil moisture and terrain strength, used to generate prediction maps of these terrain characteristics. This study found that deep learning outperformed machine learning. Specifically, a multi-layer perceptron performed the best for predicting the percent moisture content (R2/RMSE = 0.97/1.55) and the soil strength (in PSI), as measured by a cone penetrometer for the averaged 0–6” (CP06) (R2/RMSE = 0.95/67) and 0–12” depth (CP12) (R2/RMSE = 0.92/94). A Polaris MRZR vehicle was used to test the application of these prediction maps for mobility purposes, and correlations were observed between the CP06 and the rear wheel slip and the CP12 and the vehicle speed. Thus, this study demonstrates the potential of a more rapid, cost-efficient, and safer approach to predict terrain properties for mobility mapping using remote sensing data with machine and deep learning algorithms.

Kumlbúa þáttr, an Eschatological Event: The Traumatic Ontology of Being in Time in Medieval Iceland

<i>Kumlbúa þáttr</i>, an Eschatological Event: The Traumatic Ontology of Being in Time in Medieval Iceland

Geotechnical assessment of terrain strength properties on Mars using the Perseverance rover’s abrading bit

An instrument for measuring the geotechnical properties of the Martian soil provides high-value science opportunities and high-priority mission-support capabilities that serve reconnaissance, science, and engineering. Such instrument can be used to characterize the terrain for drilling, landing, trafficability, and other surface operations. However, to date, instruments for measuring geotechnical properties have been absent from Mars exploration missions. This paper examines the use of the Mars 2020 Perseverance rover’s abrading bit tool to perform regolith strength measurements. The Perseverance rover’s abrading bit has a similar form factor to a Bevameter, a tool commonly used to collect engineering geotechnical data to assess terrain traversability. To demonstrate a methodology by which the abrading bit can be employed to characterize soil strength, a portable testbed that mimics the flight system is designed and built using commercial off-the-shelf components. A data-processing pipeline is developed to convert raw measurements to soil strength. Tests were performed on three characterized simulants with different mechanical properties. The results show that the abrading bit is capable of making shear and bearing strength measurements with quantified uncertainties and demonstrate, for the first time, the ability to perform controlled geotechnical analysis using a standard science instrument.

DEM-AIA: Asymmetric inclination-aware trajectory planner for off-road vehicles with digital elevation models

Planning safe and effective trajectories for off-road unmanned ground vehicles (UGV) is a critical Artificial Intelligence (AI) challenge that can benefit from recent advances in digital elevation models (DEM) for readily capturing accurate terrain geometry. Considering path slopes is crucial to preserve stability and assess terrain traversability at feasible speeds to optimize travel time, which is highly dependent on direction (i.e., pitch and roll). In this article, we propose a new DEM-based asymmetric inclination-aware (DEM-AIA) trajectory planner for ground vehicles. The planner is an any-angle variant of the A⋆ algorithm that computes pitch and roll estimations for each segment crossing cell triangles in the line-of-sight. Furthermore, we define a non-linear velocity constraints function that integrates information about tip-over safety limitations, maximum uphill and downhill slopes for the vehicle, and asymmetric modulation of nominal flat-ground velocity for all pitch and roll combinations. The planner produces a time sub-optimal trajectory with feasible speed references for each segment crossing a cell triangle. Moreover, we provide an extensive experimental analysis of inclination-aware performance on simulated and real-world DEMs as well as a comparison with state-of-the-art path planners adapted to travel-time optimization. An executable version of the planner with parameterizable variations is publicly available.

ScaTE: A Scalable Framework for Self- Supervised Traversability Estimation in Unstructured Environments

For the safe and successful navigation of autonomous vehicles in unstructured environments, the traversability of terrain should vary based on the driving capabilities of the vehicles. Actual driving experience can be utilized in a self-supervised fashion to learn vehicle-specific traversability. However, existing methods for learning self-supervised traversability are not highly scalable for learning the traversability of various vehicles. In this work, we introduce a scalable framework for learning self-supervised traversability, which can learn the traversability directly from vehicle-terrain interaction without any human supervision. We train a neural network that predicts the proprioceptive experience that a vehicle would undergo from 3D point clouds. Using a novel PU learning method, the network simultaneously identifies non-traversable regions where estimations can be overconfident. With driving data of various vehicles gathered from simulation and the real world, we show that our framework is capable of learning the self-supervised traversability of various vehicles. By integrating our framework with a model predictive controller, we demonstrate that estimated traversability results in effective navigation that enables distinct maneuvers based on the driving characteristics of the vehicles. In addition, experimental results validate the ability of our method to identify and avoid non-traversable regions.

End-to-End Learning for Visual Navigation of Forest Environments

Off-road navigation in forest environments is a challenging problem in field robotics. Rovers are required to infer their traversability over a priori unknown and dynamically changing forest terrain using noisy onboard navigation sensors. The problem is compounded for small-sized rovers, such as that of a swarm. Their size-proportional low-viewpoint affords them a restricted view for navigation, which may be partially occluded by forest vegetation. Hand-crafted features, typically employed for terrain traversability analysis, are often brittle and may fail to discriminate obstacles in varying lighting and weather conditions. We design a low-cost navigation system tailored for small-sized forest rovers using self-learned features. The MobileNet-V1 and MobileNet-V2 models, trained following an end-to-end learning approach, are deployed to steer a mobile platform, with a human-in-the-loop, towards traversable paths while avoiding obstacles. Receiving a 128 × 96 pixel RGB image from a monocular camera as input, the algorithm running on a Raspberry Pi 4, exhibited robustness to motion blur, low lighting, shadows and high-contrast lighting conditions. It was able to successfully navigate a total of over 3 km of real-world forest terrain comprising shrubs, dense bushes, tall grass, fallen branches, fallen tree trunks, and standing trees, in over five different weather conditions and four different times of day.