Active Traversability Learning via Risk-Aware Information Gathering for Planetary Exploration Rovers

Endo, Masafumi; Ishigami, Genya

doi:10.1109/lra.2022.3207554

Cited by 12 publications

(10 citation statements)

References 56 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Numerous risk-aware planning methods have been proposed in which mobility risk is expressed as a function of rover wheel slippage, with the generation of paths that constrain risk ( Lee et al, 2016 ; Endo et al, 2023 ; Park and Chung, 2023 ). Inotsume et al, (2020) proposed a path planning algorithm based on Rapidly-exploring Random Trees (RRT), allowing users to define slip-based risk thresholds.…”

Section: Related Researchmentioning

confidence: 99%

See 1 more Smart Citation

Modeling of slip rate-dependent traversability for path planning of wheeled mobile robot in sandy terrain

Sakayori,

Ishigami

2024

Front. Robot. AI

Self Cite

View full text Add to dashboard Cite

A planetary exploration rover has been employed for scientific endeavors or as a precursor for upcoming manned missions. Predicting rover traversability from its wheel slip ensures safe and efficient autonomous operations of rovers on deformable planetary surfaces; path planning algorithms that reduce slips by considering wheel-soil interaction or terrain data can minimize the risk of the rover becoming immobilized. Understanding wheel-soil interaction in transient states is vital for developing a more precise slip ratio prediction model, while path planning in the past assumes that slips generated at the path is a series of slip ratio in steady state. In this paper, we focus on the transient slip, or slip rate the time derivative of slip ratio, to explicitly address it into the cost function of path planning algorithm. We elaborated a regression model that takes slip rate and traction force as inputs and outputs slip ratio, which is employed in the cost function to minimize the rover slip in path planning phase. Experiments using a single wheel testbed revealed that even with the same wheel traction force, the slip ratio varies with different slip rates; we confirmed that the smaller the absolute value of the slip rate, the larger the slip ratio for the same traction force. The statistical analysis of the regression model confirms that the model can estimate the slip ratio within an accuracy of 85% in average. The path planning simulation with the regression model confirmed a reduction of 58% slip experienced by the rover when driving through rough terrain environments. The dynamics simulation results insisted that the proposed method can reduce the slip rate in rough terrain environments.

show abstract

Section: Related Researchmentioning

confidence: 99%

“… Mizuno and Kubota (2020) also use a RRT* based path planner, but they model slip uncertainty with particle filters. Endo and Ishigami (2022) propose a framework that updates the latent traversability model by exploring informative terrain under the constraints of stochastic rover slip.…”

Section: Related Researchmentioning

confidence: 99%

Modeling of slip rate-dependent traversability for path planning of wheeled mobile robot in sandy terrain

Sakayori,

Ishigami

2024

Front. Robot. AI

Self Cite

View full text Add to dashboard Cite

show abstract

“…However, it suffers from significant computational challenges when transitioning between the current and next states. Several approaches use the Gaussian Process (GP) to predict the likelihood of falling cost in rover operations [9], [10]. However, it suffers from expensive computations during operation.…”

Section: B Previous Work and Motivationsmentioning

confidence: 99%

“…This replicates real-world conditions where sensors often encounter challenges such as reflective or deformable surfaces. This paper incorporates artificial noise directly into the robot's exteroceptive observations, as outlined in the equation (9). In this setup, the signal-to-noise ratio parameter σ H is set to a value of −10% ≤ σ H ≤ 10% specifically for height map measurements.…”

Section: Additional Noisementioning

confidence: 99%

Robust Traversability Prediction Using Multiple Costs for Quadruped Robot in Random Terrains

Muhamad,

Kim,

Park

2024

IEEE Access

View full text Add to dashboard Cite

The quadruped robot has to assess the feasibility of upcoming terrains before making contact to safely traverse various environments. This assessment is called traversability in the literature on quadruped robots. Trasversability has recently posed challenges due to a high-dimensional system that leads to long computational times. Furthermore, exteroceptive observations often suffer from noise that potentially causes misinterpretations of terrains and results in an inaccurate assessment. This paper proposes a robust traversability predictor to tackle these issues by utilizing a Convolutional Neural Network (CNN) encoder, CNN decoder, and Multi-layer Perceptron (MLP) to predict multiple costs associated with traversability. The integration of the CNN encoder and decoder helps mitigate the effect of noise in exteroceptive observations, while the MLP network serves as a predictor for multiple costs. The proposed method utilizes the information collected from a physics simulator to avoid hand-crafted multiple-cost labeling. It can predict a comprehensive set of costs that overcomes the limitations of relying on a single cost metric. It also achieves faster computational time by utilizing neural networks, in contrast to the model-based approach in the literature. The robustness of the proposed method is validated by comparing it to a baseline noise-free prediction model and an existing method in the literature. The results indicate that the proposed method exhibits the lowest prediction errors. Therefore, despite the noise in exteroceptive observations, the proposed multiple cost-based traversability predictor has better accuracy and robustness than the baseline and existing methods.

show abstract

“…Other works focus on improving semantic segmentation by reducing the number of training samples, given the challenges associated with collecting training data in a space exploration scenario (Zhang et al, 2022;Goh et al, 2022). (Endo et al, 2023) assess risk-aware traversability costs by fusing semantic terrain classification and a slip model in a probabilistic manner.…”

Section: Traversability From Semanticsmentioning

confidence: 99%

Locomotion Policy Guided Traversability Learning using Volumetric Representations of Complex Environments

Frey¹,

Hoeller²,

Khattak³

et al. 2022

2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

View full text Add to dashboard Cite

Autonomous navigation at high speeds in off-road environments necessitates robots to comprehensively understand their surroundings using onboard sensing only. The extreme conditions posed by the off-road setting can cause degraded camera image quality due to poor lighting and motion blur, as well as limited sparse geometric information available from LiDAR sensing when driving at high speeds. In this work, we present RoadRunner, a novel framework capable of predicting terrain traversability and an elevation map directly from camera and LiDAR sensor inputs. RoadRunner enables reliable autonomous navigation, by fusing sensory information, handling of uncertainty, and generation of contextually informed predictions about the geometry and traversability of the terrain while operating at low latency. In contrast to existing methods relying on classifying handcrafted semantic classes and using heuristics to predict traversability costs, our method is trained end-to-end in a self-supervised fashion. The RoadRunner network architecture builds upon popular sensor fusion network architectures from the autonomous driving domain, which embed LiDAR and camera information into a common Bird's Eye View perspective. Training is enabled by utilizing an existing traversability estimation stack to generate training data in hindsight in a scalable manner from realworld off-road driving datasets. Furthermore, RoadRunner improves the system latency by a factor of ∼ 4, from 500 ms to 140 ms, while improving the accuracy for traversability costs and elevation map predictions. We demonstrate the effectiveness of RoadRunner in enabling safe and reliable offroad navigation at high speeds in multiple real-world driving scenarios through unstructured desert environments.

show abstract

Active Traversability Learning via Risk-Aware Information Gathering for Planetary Exploration Rovers

Cited by 12 publications

References 56 publications

Modeling of slip rate-dependent traversability for path planning of wheeled mobile robot in sandy terrain

Modeling of slip rate-dependent traversability for path planning of wheeled mobile robot in sandy terrain

Robust Traversability Prediction Using Multiple Costs for Quadruped Robot in Random Terrains

Locomotion Policy Guided Traversability Learning using Volumetric Representations of Complex Environments

Contact Info

Product

Resources

About