Learning multiobjective rough terrain traversability

Wallin, Erik; Wiberg, Viktor; Vesterlund, Folke; Holmgren, Johan; Persson, Henrik J.; Servin, Martin

doi:10.1016/j.jterra.2022.04.002

Cited by 6 publications

(6 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Traversability refers to the capacity of a ground vehicle to traverse a particular terrain area while satisfying predefined objectives and criteria [29]. This capacity of vehicles has also been referred to with keywords like drivability, navigability, trafficability, and mobility [10,30].…”

Section: Terrain Traversability Analysismentioning

confidence: 99%

“…Predicting traversability in advance is fundamental for autonomous path planning in unstructured environments. Inaccurate or inefficient information regarding traversability can result in the generation of substandard paths which not only consume excessive energy and time but also pose unnecessary risks of damaging both equipment and the environment [29]. The global path planner is responsible for selecting the path and optimizing it based on the outputs obtained from the TTA results, such as the traversability map [31,32], traversability cost model [33,34], or terrain classifier [35].…”

Section: Terrain Traversability Analysismentioning

confidence: 99%

“…The geometric features of terrain related to TTA encompass physiognomy and static obstacles. Physiognomy is characterized by elevation, slope, roughness, and curvature [29]. Elevation changes and slope characteristics, including vertical and lateral slopes, are fundamental factors in TTA models [36], impacting climbing ability and potential rollover [10].…”

Section: Terrain Geometrical and Physical Factorsmentioning

confidence: 99%

See 2 more Smart Citations

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

Wang,

Li,

Zhang

et al. 2024

Machines

View full text Add to dashboard Cite

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.

show abstract

Section: Terrain Traversability Analysismentioning

confidence: 99%

Section: Terrain Traversability Analysismentioning

confidence: 99%

Section: Terrain Geometrical and Physical Factorsmentioning

confidence: 99%

See 1 more Smart Citation

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

Wang,

Li,

Zhang

et al. 2024

Machines

View full text Add to dashboard Cite

show abstract

“…A number of hyperparameters, such as the batch size, learning rate, and network parameters, were varied to find the agents with the best performance. The best model was trained using a batch size of 1600, a learning rate of 0.00025, and a feature extractor CNN with (8, 8, 8) filters of sizes [8,4,3] and strides [4, 2, 1], and 64 output features. The fully connected networks have two hidden layers of size (64, 64), with tanh activation functions.…”

Section: Rl Algorithm and Networkmentioning

confidence: 99%

“…Driven by the global trend of big data and the progress in machine learning, the forestry industry is experiencing an increase in the collection and availability of large amounts of data. Harvest areas can be scanned from the air and the ground, and both ground and trees can be segmented [2,3], allowing detailed terrain maps to be created for path planning [4], among other things. Harvesters are increasingly being equipped with high-precision positioning systems, and are able to store the geospatial information of the felled logs [5] as well as the travelled paths.…”

Section: Introductionmentioning

confidence: 99%

Multi-Log Grasping Using Reinforcement Learning and Virtual Visual Servoing

Wallin,

Wiberg,

Servin

2023

Robotics

Self Cite

View full text Add to dashboard Cite

We explore multi-log grasping using reinforcement learning and virtual visual servoing for automated forwarding in a simulated environment. Automation of forest processes is a major challenge, and many techniques regarding robot control pose different challenges due to the unstructured and harsh outdoor environment. Grasping multiple logs involves various problems of dynamics and path planning, where understanding the interaction between the grapple, logs, terrain, and obstacles requires visual information. To address these challenges, we separate image segmentation from crane control and utilise a virtual camera to provide an image stream from reconstructed 3D data. We use Cartesian control to simplify domain transfer to real-world applications. Because log piles are static, visual servoing using a 3D reconstruction of the pile and its surroundings is equivalent to using real camera data until the point of grasping. This relaxes the limits on computational resources and time for the challenge of image segmentation, and allows for data collection in situations where the log piles are not occluded. The disadvantage is the lack of information during grasping. We demonstrate that this problem is manageable and present an agent that is 95% successful in picking one or several logs from challenging piles of 2–5 logs.

show abstract