Probabilistic Terrain Mapping for Mobile Robots With Uncertain Localization

By Péter Fankhauser, Michael Bloesch, Marco Hutter. IEEE Robotics and Automation Letters, 2018

326 citations recorded by the source corpus.

Mobile robots build on accurate, real-time mapping with onboard range sensors to achieve autonomous navigation over rough terrain. Existing approaches often rely on absolute localization based on tracking of external geometric or visual features. To circumvent the reliability issues of these approaches, we propose a novel terrain mapping method, which bases on proprioceptive localization from kinematic and inertial measurements only. The proposed method incorporates the drift and uncertainties of the state estimation and a noise model of the distance sensor. It yields a probabilistic terrain estimate as a grid-based elevation map including upper and lower confidence bounds. We demonstrate the effectiveness of our approach with simulated datasets and real-world experiments for real-time terrain mapping with legged robots and compare the terrain reconstruction to ground truth reference maps.

Subfields: Aerospace Engineering, Biomedical Engineering, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/8182c78b881844090d4d8423aec68104b40b5dd1

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

In Biomedical Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

Robot Parkour Learning (2023, methodology): “Classical modular methods rely on decoupled visual perception and control pipelines, where the elevation maps [75, 76, 77, 78, 79, 80, 81], traversability maps [82, 83, 84, 85], or state estimators [86, 87, 88, 89, 90, 91, 92, 93, 94, 95] are constructed as intermediate representations for downstrea”
ANYmal parkour: Learning agile navigation for quadrupedal robots (2023, methodology): “The approach in [37] improves standard elevation mapping [38] by adding various filtering operations and post-processing steps, which can explicitly realign the map to compensate for drift in the z-direction, and performs additional visibility checks to clean up outliers.”
Extreme Parkour with Legged Robots (2023, unknown): “…curriculum for overcoming exploration burden in RL. • Our results set a new landmark in learning-driven park-our with high jumps that are 2x the height of the robot, long jumps that are 2x the length of the robot, walking on front two legs (handstand), and jumping over titled ramps directly from a…”
Legged Locomotion in Challenging Terrains using Egocentric Vision (2022, methodology): “241 Locomotion from Elevation Maps To achieve visual control of walking, classical methods de242 couples the perception and control aspects, assuming a perfect output from perception, such as an 243 elevation map, and then using it for planning and control [50, 51, 52, 53, 54].”
Learning robust perceptive locomotion for quadrupedal robots in the wild (2022, background): “5D elevation map representation cannot accurately repre sent overhanging objects such as tree branches or low ceilings (17).”