Off-road Environments Research Articles

Advanced scene aware navigation for the heavy duty off-road vehicle Unimog

This paper describes a novel approach for scene aware navigation in rough off-road environments using behavior-based control, graph-like world descriptions, and ontologies to determine correlations of surrounding objects and environmental properties. A geometry-based obstacle segmentation and convolutional neural networks for classification determine available scene items integrated into low and higher-level scene representations that are processed by a semantic control. Additionally, the surface's shape, vehicle kinematics, and dynamics and object properties are considered for the planning of low-level trajectories that are hierarchically integrated into the framework. The approach was implemented using behavior-based control and was evaluated using a Unimog U5023 off-road truck in simulation and real-world scenarios.

Adaptive road detection method combining lane line and obstacle boundary

In order to realise autonomous navigation of unmanned platforms in urban or off-road environments, it is crucial to study accurate, versatile and real-time road detection methods. This study proposes an adaptive road detection method that combines lane lines and obstacle boundaries, which can be applied to a variety of driving environments. Combining multi-channel threshold processing, it is robust to lane feature detection under various complex situations. Obstacle information extracted from the grid image constructed by 3D LIDAR point cloud is used for lane feature selection to avoid interference from pedestrians and vehicles. The proposed method makes use of adaptive sliding window for feature selection, and piecewise least squares method for road line fitting. Experimental results on dataset and in real-world environments show that the proposed method can overcome illumination changes, shadow occlusion, pedestrian, vehicle interference and so on in a variety of scenes. The proposed method has good enough efficiency, robustness and real-time performance.

LIDAR Scan Matching in Off-Road Environments

Accurately matching the LIDAR scans is a critical step for an Autonomous Land Vehicle (ALV). Whilst most previous works have focused on the urban environment, this paper focuses on the off-road environment. Due to the lack of a publicly available dataset for algorithm comparison, a dataset containing LIDAR pairs with varying amounts of offsets in off-road environments is firstly constructed. Several popular scan matching approaches are then evaluated using this dataset. Results indicate that global approaches, such as Correlative Scan Matching (CSM), perform best on large offset datasets, whilst local scan matching approaches perform better on small offset datasets. To combine the merits of both approaches, a two-stage fusion algorithm is designed. In the first stage, several transformation candidates are sampled from the score map of the CSM algorithm. Local scan matching approaches then start from these transformation candidates to obtain the final results. Four performance indicators are also designed to select the best transformation. Experiments on a real-world dataset prove the effectiveness of the proposed approach.

A survey on multi-sensor fusion based obstacle detection for intelligent ground vehicles in off-road environments

With the development of sensor fusion technologies, there has been a lot of research on intelligent ground vehicles, where obstacle detection is one of the key aspects of vehicle driving. Obstacle detection is a complicated task, which involves the diversity of obstacles, sensor characteristics, and environmental conditions. While the on-road driver assistance system or autonomous driving system has been well researched, the methods developed for the structured road of city scenes may fail in an off-road environment because of its uncertainty and diversity. A single type of sensor finds it hard to satisfy the needs of obstacle detection because of the sensing limitations in range, signal features, and working conditions of detection, and this motivates researchers and engineers to develop multi–sensor fusion and system integration methodology. This survey aims at summarizing the main considerations for the onboard multi-sensor configuration of intelligent ground vehicles in the off-road environments and providing users with a guideline for selecting sensors based on their performance requirements and application environments. State-of-the-art multi-sensor fusion methods and system prototypes are reviewed and associated to the corresponding heterogeneous sensor configurations. Finally, emerging technologies and challenges are discussed for future study.

LiDAR Data Segmentation in Off-Road Environment Using Convolutional Neural Networks (CNN)

<div class="section abstract"><div class="htmlview paragraph">Recent developments in the area of autonomous vehicle navigation have emphasized algorithm development for the characterization of LiDAR 3D point-cloud data. The LiDAR sensor data provides a detailed understanding of the environment surrounding the vehicle for safe navigation. However, LiDAR point cloud datasets need point-level labels which require a significant amount of annotation effort. We present a framework which generates simulated labeled point cloud data. The simulated LiDAR data was generated by a physics-based platform, the Mississippi State University Autonomous Vehicle Simulator (MAVS). In this work, we use the simulation framework and labeled LiDAR data to develop and test algorithms for autonomous ground vehicle off-road navigation. The MAVS framework generates 3D point clouds for off-road environments that include trails and trees. The important first step in off-road autonomous navigation is the accurate segmentation of 3D point cloud data to identify the potential obstacles in the vehicle path. We use simulated LiDAR data to segment and detect obstacles using convolutional neural networks (CNN). Our analysis is based on SqueezeSeg, a CNN-based model for point cloud segmentation. The CNN was trained with a labelled dataset of off-road imagery generated by MAVS and evaluated on the simulated dataset. The segmentation of the LiDAR data is done by point-wise classification and the results show excellent accuracy in identifying different objects and obstacles in the vehicle path. In this paper, we evaluated the segmentation performance at different LiDAR vertical resolutions: the 8-beam and 16-beam. The results showed that there is about 5% increase in accuracy with 16-beam sensors compared with the 8-beam sensors in detecting obstacles and trees. However, the 8-beam LiDAR performance is comparable with the 16-beam sensor in segmenting vegetation, trail-road and ground.</div></div>

Multi-camera visual SLAM for off-road navigation

With the rapid development of computer vision, vision-based simultaneous localization and mapping (vSLAM) plays an increasingly important role in the field of unmanned driving. However, traditional SLAM methods based on a monocular camera only perform well in simple indoor environments or urban environments with obvious structural features. In off-road environments, the situation that SLAM encounters could be complicated by problems such as direct sunlight, leaf occlusion, rough roads, sensor failure, sparsity of stably trackable texture. Traditional methods are highly susceptible to these factors, which lead to compromised stability and reliability. To counter such problems, we propose a panoramic vision SLAM method based on multi-camera collaboration, aiming at utilizing the characters of panoramic vision and stereo perception to improve the localization precision in off-road environments. At the same time, the independence and information sharing of each camera in multi-camera system can improve its ability to resist bumps, illumination, occlusion and sparse texture in an off-road environment, and enable our method to recover the metric scale. These characters ensure unmanned ground vehicles (UGVs) to locate and navigate safely and reliably in complex off-road environments.

Power Prediction for Heterogeneous Ground Robots Through Spatial Mapping and Sharing of Terrain Data

Ground robot power consumption often varies significantly throughout an off-road environment, depending on the terrain. Previous traversals over the environment can inform future predictions through spatial mapping of collected power data. However, such predictions are particular to a single robot. We develop a framework based on multi-task Gaussian process regression (MTGP) to share terrain information across multiple robots. The framework allows for the efficient selection of hyperparameters and the prediction of power consumption along paths through nearest neighbor data selection. The scaling of this approach is demonstrated in simulation. Furthermore, experimental testing with two robots shows a significant reduction in power prediction error when an environment, mapped by one robot, informs the power predictions of another robot.

Automation Kit for Dual-Mode Military Unmanned Ground Vehicle for Surveillance Missions

The field of autonomous vehicles is an area of great importance nowadays among Intelligent Transport Systems. These autonomous vehicles must meet high standards of robustness, safety and protection to make autonomous driving safer than human driving, thus improving current accident rates and energy efficiency, so its complete implementation in the market is expected in the long term. However, there are other areas, such as the agricultural and military sectors, where the demand for technologies related to autonomous driving is also high but the applicability could be faster. In this sense, the unmanned ground military vehicles (UGV) are a technology available in international armies for some years, but that is still subject to a series of technical, logistical and material limitations that reduce its applicability. This paper presents an automation kit for transforming military service vehicles into military UGVs, developed through the transfer of autonomous vehicle technology from the civil sector to the military sector. This kit has been installed in a quite common vehicle of the Spanish army, in order to provide two functional capabilities: autonomous navigation and teleoperation, in off-road environments with minimal or no traffic marks or infrastructure aids. The vehicle equipped with this kit is oriented to different missions such as circulation in convoys, operation as tele-operated probe vehicle and perimeter surveillance. The automation kit involves no permanent modifications in the vehicle that can be driven manually when desired, allows vehicles to be used both in missions in conflict zones and to support operations in emergencies and natural catastrophes, and can be easily installed in any kind of vehicle.

Combination of Recurrent Neural Network and Deep Learning for Robot Navigation Task in Off-Road Environment

SUMMARYThis paper tackles the challenge of the necessity of using the sequence of past environment states as the controller’s inputs in a vision-based robot navigation task. In this task, a robot has to follow a given trajectory without falling in pits and missing its balance in uneven terrain, when the only sensory input is the raw image captured by a camera. The robot should distinguish big pits from small holes to decide between avoiding and passing over. In non-Markov processes such as the abovementioned task, the decision is done using past sensory data to ensure admissible performance. Applying images as sensory inputs naturally causes the curse of dimensionality difficulty. On the other hand, using sequences of past images intensifies this difficulty. In this paper, a new framework called recurrent deep learning (RDL) with combination of deep learning (DL) and recurrent neural network is proposed to cope with the above challenge. At first, the proper features are extracted from the raw image using DL. Then, these represented features plus some expert-defined features are used as the inputs of a fully connected recurrent network (as target network) to generate command control of the robot. To evaluate the proposed RDL framework, some experiments are established on WEBOTS and MATLAB co-simulation platform. The simulation results demonstrate the proposed framework outperforms the conventional controller based on DL for the navigation task in the uneven terrains.

Learning to Drive: End-to-End Off-Road Path Prediction

Autonomous driving is a field currently gaining a lot of attention, and recently ?end to end? approaches, whereby a machine learning algorithm learns to drive by emulating a human driver, have demonstrated significant potential. However, recent work has focused on the on-road environment, rather than the more challenging off-road environment. In this work we propose a new approach to this problem, whereby instead of learning to predict immediate driver control inputs, we train a deep convolutional neural network (CNN) to predict the future path that a vehicle will take through an off-road environment visually, addressing several limitations inherent in existing methods. We combine a novel approach to automatic training data creation, making use of stereoscopic visual odometry, with a state-of-the-art CNN architecture to map a predicted route directly onto image pixels, and demonstrate the effectiveness of our approach using our own off-road data set.

Chance constrained reachability in environments with spatially varying energy costs

In off-road environments, energy costs are highly uncertain and variable due to unknown terrains. To plan missions for robots with limited energy storage capacity, a robot’s reachable set must be determined. This work presents a novel approach for learning reachable sets based on data collected during a mission. Leveraging the authors’ previous work, a spatial energy map of an unknown environment, built with data collected by a robot, can be used to compute a chance constrained reachable set (CCRS) based on a user-defined confidence level. Simulations demonstrate that as a robot collects more data on an energy map, the true positive rate of a computed CCRS improves significantly while the false positive rate remains low, implying that a robot’s reachability can be robustly determined for use in future missions.

Development of autonomous tractor model considering the off-road environment

Development of autonomous tractor model considering the off-road environment

Road-Feature-Based Multiparameter Road Complexity Calculation Model of Off-Road Environment

To promote the development of unmanned ground vehicle technologies, it is necessary to design a scientific and reasonable test method. Road is an important part of test environmental elements, and different road conditions can examine the adaptability of unmanned ground vehicles to the environment. Therefore, the scientific calculation of road complexity is of great importance. Previous studies on road are mainly based on the concept of road roughness; however due to the unicity of road feature indicators, road complexity can only be reflected to a certain extent. This paper proposes a new road-feature-based multiparameter road complexity calculation model of off-road environment to show the complexity of road more comprehensively. First, a multi-sensor-based data acquisition mobile platform is established to obtain more complete road data. Then, based on the analysis of road feature, road indicators like three-dimensional scale, average slope, and adhesion characteristics of travelable area are obtained. According to the analysis methods of road roughness, the principle of analytic hierarchy process, and the data collected from off-road environment, the calculation model of road complexity is determined. Finally, by calculating complexities of several cross-country roads, the feasibility of this model is verified, which provides a theoretical support for the scientific calculation and quantitative analysis of different road complexities.

Field terrain recognition based on extreme learning theory using wavelet and texture features

Terrain recognition technology plays a key role in enhancing autonomous mobility for Quadruped robot in off-road environments. However, feature extraction and classification algorithm are the key to accuracy and efficiency of the terrain recognition. Regarding the characteristics of different terrain surface properties and structures, it gets low-dimensional and high dimensional characteristics by texture features and wavelet transform and uses them as training features of classifier. Then its efficiency is not high and convergence speed is slow for traditional learning algorithm, which is difficult to meet the requirements. So the extreme learning machine is used to classify the terrain pictures collected by robot in real time. Experimental results show that the accuracy of extreme learning machine terrain classification is higher than the traditional neural network algorithm and the support vector machine, and algorithm efficiency is raised more than several times for the sample size of 6000, which meet...

Off-road Path Planning Based on Improved Ant Colony Algorithm

Optimal vehicle off-road path planning problem must consider surface physical properties of terrain and soil. In this paper, we firstly analyse the comprehensive influence of terrain slope and soil strength to vehicle’s off-road trafficability. Given off-road area, the GO or NO-GO tabu table of terrain gird is determined by slope angle and soil remolding cone index (RCI). By applying tabu table and grid weight table, the influence of terrain slope and soil RCI are coordinated to reduce the search scope of algorithm and improve search efficiency. Simulation results based on tracked vehicle M1A1 in off-road environment show that, improved ant colony path planning algorithm not only considers the influence of actual terrain and soil, but also improves computation efficiency. The time cost of optimal routing computation is much lower which is essential for real time off-road path planning scenarios.

Scene-Adaptive Off-Road Detection Using a Monocular Camera

This paper studies vision-based road detection for a robot’s path following in off-road environments. We define the problem as detecting the region in front of the robot that is mechanically traversable (i.e., mechanical traversability ), that is apt to be chosen by a human to drive through (i.e., human selection ), and that extends for a distance to show the road’s direction, shape, or even network of the intersection ahead (i.e., far-field capability ). An algorithm framework is designed that contains two parts: inference and learning. In inference, the problem is formulated as a consecutive road type classification and road region segmentation to address the diversity of terrain surfaces. In model learning, the robot is first driven by a human being, with image samples on the track of the robot being collected that meet the prerequisites of both mechanical traversability and human selection. Evaluation measures are defined to examine the three requirements of mechanical traversability, human selection, and far-field capability. The performances of the above aspects are demonstrated on a data set using LiDAR, track and manual references, which will be released together with this publication.

地表土壤探测及传感器优化布设

复杂多变的越野环境对车辆行进的机动性、灵活性具有重要影响，为提高越野车辆机动性必须对地形、坡度、土壤类型、水分、植被等地表土壤属性进行探测。本文首先对最新涌现的探测技术进行分析比较，发现在地表原位布设轻量级的传感器节点可以快速准确全面地监测土壤属性。据此我们提出了土壤传感器节点的优化布设方案，仿真结果表明该方案既保证了监测区域的覆盖率又提高了节点自身的定位精度，具有实际的应用价值。 Complex and changeable off-road environment influences the mobility and flexibility of vehicle a lot. To improve mobility of off-road vehicles, the terrain conditions including slops, soil types, moisture content and vegetative cover must be sensed. In this paper, the state-of-the-art sensing techniques are analyzed and compared firstly, the comparison results show that by deploying light weight sensors in situ soil, the soil properties can be monitored quickly and precisely. Based on this, an optimal soil sensor deployment method is proposed. Simulation results show that this method not only guarantees the coverage ratio of sensing area but also improves localization precision of sensors, which possesses practical value of application.

Rancang Bangun Prototype Robot Pendeteksi Gas Metana Berbasis Mikrokontroler Arduino Untuk Eksplorasi Pertambangan

Accident which is occurred due to the harmful gas always found in the mining location. One example of harmful gas is the explosive methane gas. One of attempt to overcome this problem is by using a methane gas detector robot that can detect the gas early before the mining process is conducted. This kind of robot can be control autonomously from a place which is located far away from the mining area. Therefore, the mining workers can work safely. This research is aimed to design and manufacturing a mobile robot for detecting a methane gas in the mining area. The research is conducted step by step by making the mechanical components of robot, manufacturing the electronic circuits and programming the Arduino microcontroller using the Arduino IDE software. Mechanical design of the robot is targeted for operating the robot in the off-road environments. The main control center of the robot arm uses an Arduino UNO microcontroller. Robot control system has a methane gas detector to sense the methane gas. The command signal is applied autonomously by using Android based smart phone, which move the robot to the desired area. The signal sent by gas sensor MQ-04 will ring the buzzer and turn on the LED of the robot whenever the robot detects a methane gas. The experimental results shows that the robot can detect the methane gas and robot maneuvering can be conducted successfully.

A Framework for Predicting Future System Performance in Autonomous Unmanned Ground Vehicles

The development of complex self-adaptive systems has accelerated rapidly over the past decade, led by the Department of Defense, which has sought to develop and field military systems, such as unmanned aerial vehicles and unmanned ground vehicles, with elevated levels of autonomy to accomplish their mission with reduced funding and manpower. As their role increases, such systems must be able to adapt and learn, and make nondeterministic decisions. To field such systems requires extensive testing, evaluation, verification, and validation—a challenging task. To address this, we apply a novel systems perspective to develop a framework to predict future system performance based on the complexity of the operating environment using newly introduced complexity measures and learned costs. In this paper, we consider an autonomous military ground robot navigating in complex off-road environments. Using our model and experimental data from Defense Advanced Research Projects Agency-led experiments, we demonstrate the accuracy with which our model can predict system performance and then validate our model against other experimental results.

A path planning method of anti-jamming ability improvement for autonomous vehicle navigating in off-road environments

PurposeNavigating in off-road environments is a huge challenge for autonomous vehicles, due to the safety requirement, the effects of noises and non-holonomic constraints of vehicle. This paper aims to describe a path planning method based on fuzzy support vector machine (FSVM) and general regression neural network (GRNN) that is able to provide a solution path for the autonomous vehicle navigating in the off-road environments.Design/methodology/approachThe authors decompose the path planning problem into three steps. In the first step, A* algorithm is applied to obtain the positive and negative samples. In the second step, the authors use a learning approach based on radial basis function kernel FSVM to maximize the safety margin for driving, and the fuzzy membership is designed based on GRNN which can help to resolve the problem that the traditional path planning method is easily influenced by noises or outliers. In the third step, the Bezier interpolation algorithm is used to smooth the path. The simulations are designed to verify the parameters of the path planning algorithm.FindingsThe method is implemented on autonomous vehicle and verified against many outdoor scenes. Road test indicates that the proposed method can produce a flexible, smooth and safe path with good anti-jamming performance.Originality/valueThis paper applied a new path planning method based on GRNN-FSVM for autonomous vehicle navigating in off-road environments. GRNN-FSVM can reduce the effects of outliers and maximize the safety margin for driving, the generated path is smooth and safe, while satisfying the constraint of vehicle kinematic.