A Vision-Based Two-Stage Framework for Inferring Physical Properties of the Terrain

Dong, Yunlong; Wei, Guo; Zha, Fusheng; Liu, Yizhou; Chen, Chen; Sun, Lining

doi:10.3390/app10186473

Cited by 15 publications

(14 citation statements)

References 37 publications

(45 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our baseline in this part is the two-stage inference framework proposed by Dong et al 39 . It is the most advanced research in the sphere of inferring field terrain’s physical properties.…”

Section: Methodsmentioning

confidence: 99%

“…Second, there is no information related to physical properties in these datasets. At present, the most advanced research on terrain segmentation and physical properties inference is the two-stage inference framework of Dong et al 39 . In Dong’s work, the image dataset for field terrain was established, and the terrain type, friction property, and stiffness property were predicted.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Monocular vision-based multiinformation map construction framework for field terrain

et al. 2022

Self Cite

View full text Add to dashboard Cite

Three-dimensional (3D) spatial data, terrain type, and physical properties of field terrain are very important information for mobile robots in trajectory planning, force adjustment, motion control, dynamic parameter optimization, and so on. Obtaining the information in advance through vision can provide a prior perception to help make reasonable decisions and improve security and adaptability. We present the first monocular-vision-based multiinformation map construction framework for field terrain. The current state-of-the-art physical properties inference algorithm is promoted in illumination robustness, time consumption, and terrain richness. Our improvement makes it possible to apply physical properties inference to the real-time mapping of field terrain. An improved monocular vision and inertia fused localization scheme is used to introduce the gravity direction. Depending on the obtained terrain segmentation and camera pose, an unsupervised monocular scale-consistent terrain dense depth estimation method with comprehensive geometric constraints is designed. Finally, the components are systematically combined to form a complete framework, and the ambiguity of inference results between multiview frames is addressed. Comprehensive experiments are conducted to verify the effectiveness of each component and the performance of the overall framework. Our multiinformation map construction framework can output a gravity-aligned 3D RGB map, terrain map, friction property map, and stiffness property map at once with high accuracy.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Monocular vision-based multiinformation map construction framework for field terrain

et al. 2022

Self Cite

View full text Add to dashboard Cite

show abstract

“…However, in a more challenging environment, the slope angle may be larger. In the authors' recently published work [36], a vision-based framework to infer the physical properties of the terrain is proposed. Furthermore, the framework in [36] will be introduced into the control system of the large-size hexapod robot, and the leg trajectory generation process in the attitude trajectory optimization method proposed in this paper will be updated to cooperate with the function of the framework to further improve the walking balance of the large-size hexapod robot on more challenging terrains.…”

Section: Conclusion and Further Workmentioning

confidence: 99%

“…In the authors' recently published work [36], a vision-based framework to infer the physical properties of the terrain is proposed. Furthermore, the framework in [36] will be introduced into the control system of the large-size hexapod robot, and the leg trajectory generation process in the attitude trajectory optimization method proposed in this paper will be updated to cooperate with the function of the framework to further improve the walking balance of the large-size hexapod robot on more challenging terrains. More experiments of the large-size hexapod robot walking on challenging terrains will be carried out in the future to validate the effectiveness of the control method proposed, especially the slope-climbing experiment with a large slope angle.…”

Section: Conclusion and Further Workmentioning

confidence: 99%

Attitude Trajectory Optimization to Ensure Balance Hexapod Locomotion

Chen

Wei

Wang

et al. 2020

Sensors

Self Cite

View full text Add to dashboard Cite

This paper proposes a simple attitude trajectory optimization method to enhance the walking balance of a large-size hexapod robot. To achieve balance motion control of a large-size hexapod robot on different outdoor terrains, we planned the balance attitude trajectories of the robot during walking and introduced how leg trajectories are generated based on the planned attitude trajectories. While planning the attitude trajectories, high order polynomial interpolation was employed with attitude fluctuation counteraction considered. Constraints that the planned attitude trajectories must satisfy during walking were well-considered. The trajectory of the swing leg was well designed with the terrain attitude considered to improve the environmental adaptability of the robot during the attitude adjustment process, and the trajectory of the support leg was automatically generated to satisfy the demand of the balance attitude trajectories planned. Comparative experiments of the real large-size hexapod robot walking on different terrains were carried out to validate the effectiveness and applicability of the attitude trajectory optimization method proposed, which demonstrated that, compared with the currently developed balance motion controllers, the attitude trajectory optimization method proposed can simplify the control system design and improve the walking balance of a hexapod robot.

show abstract

“…As a non-interactive approach, the visual terrain classification method can recognize not only the traversing terrain, but also the terrains traversed or that will be traversed [ 15 , 16 , 17 ]. However, it suffers from two issues: (i) vision cannot work in extreme illumination (glare or dark); (ii) vision may be confused by the covering materials, thus it cannot recognize the real terrain [ 18 , 19 , 20 ]. Therefore, the interactive terrain classification, which are often implemented by means of acoustics [ 21 , 22 ], haptics [ 23 , 24 ], or vibration [ 25 , 26 ], is becoming more and more promising in robotic environment perception.…”

Section: Introductionmentioning

confidence: 99%

Frequency-Temporal Disagreement Adaptation for Robotic Terrain Classification via Vibration in a Dynamic Environment

Chen

Jin-yi

et al. 2020

Sensors

View full text Add to dashboard Cite

The accurate terrain classification in real time is of great importance to an autonomous robot working in field, because the robot could avoid non-geometric hazards, adjust control scheme, or improve localization accuracy, with the aid of terrain classification. In this paper, we investigate the vibration-based terrain classification (VTC) in a dynamic environment, and propose a novel learning framework, named DyVTC, which tackles online-collected unlabeled data with concept drift. In the DyVTC framework, the exterior disagreement (ex-disagreement) and interior disagreement (in-disagreement) are proposed novely based on the feature diversity and intrinsic temporal correlation, respectively. Such a disagreement mechanism is utilized to design a pseudo-labeling algorithm, which shows its compelling advantages in extracting key samples and labeling; and consequently, the classification accuracy could be retrieved by incremental learning in a changing environment. Since two sets of features are extracted from frequency and time domain to generate disagreements, we also name the proposed method feature-temporal disagreement adaptation (FTDA). The real-world experiment shows that the proposed DyVTC could reach an accuracy of 89.5%, but the traditional time- and frequency-domain terrain classification methods could only reach 48.8% and 71.5%, respectively, in a dynamic environment.

show abstract

A Vision-Based Two-Stage Framework for Inferring Physical Properties of the Terrain

Cited by 15 publications

References 37 publications

Monocular vision-based multiinformation map construction framework for field terrain

Monocular vision-based multiinformation map construction framework for field terrain

Attitude Trajectory Optimization to Ensure Balance Hexapod Locomotion

Frequency-Temporal Disagreement Adaptation for Robotic Terrain Classification via Vibration in a Dynamic Environment

Contact Info

Product

Resources

About