In this paper, we present a mapping method of stairs for quadruped robots based on point-cloud measurements and stair-modeling. Because of the quadruped robot's physical property, the distance between the robot...
详细信息
ISBN:
(纸本)9781728132327
In this paper, we present a mapping method of stairs for quadruped robots based on point-cloud measurements and stair-modeling. Because of the quadruped robot's physical property, the distance between the robot's vision sensor and the stair is short and the detecting range of the point-cloud sensor is narrow when the robot navigates a stair environment. This causes many problems, for example, difficulties in finding features on the image or tracking them. As a result, vision-only based odometry becomes unreliable. What we propose here is to use the estimation model of stairs fused with point-cloud measurements from a depth sensor. By combining sensor measurement and estimation data from the regular shape of stairs, we overcome the disadvantage of mapping that comes from the limited measurement distance between the object and the sensor. We use a clustering algorithm for stairs based on the surface normal directions of the stair surfaces and their global coordinates and this method provides us robust and reliable clustering results. Finally, we show the performance of the implemented ideas in experiments with hand-held sensors as well as with a quadruped robot.
We address semantic segmentation of road-objects from 3D LiDAR point clouds. In particular, we wish to detect and categorize instances of interest, such as cars, pedestrians and cyclists. We formulate this problem as ...
详细信息
ISBN:
(纸本)9781538630815
We address semantic segmentation of road-objects from 3D LiDAR point clouds. In particular, we wish to detect and categorize instances of interest, such as cars, pedestrians and cyclists. We formulate this problem as a point-wise classification problem, and propose an end-to-end pipeline called SqueezeSeg based on convolutional neural networks (CNN): the CNN takes a transformed LiDAR point cloud as input and directly outputs a point-wise label map, which is then refined by a conditional random field (CRF) implemented as a recurrent layer. Instance-level labels are then obtained by conventional clusteringalgorithms. Our CNN model is trained on LiDAR point clouds from the KITTI [ 1] dataset, and our point-wise segmentation labels are derived from 3D bounding boxes from KITTI. To obtain extra training data, we built a LiDAR simulator into Grand Theft Auto V (GTA-V), a popular video game, to synthesize large amounts of realistic training data. Our experiments show that SqueezeSeg achieves high accuracy with astonishingly fast and stable runtime (8.7 +/- 0.5 ms per frame), highly desirable for autonomous driving. Furthermore, additionally training on synthesized data boosts validation accuracy on real-world data. Our source code is open-source released1. The paper is accompanied by a video(2) containing a high level introduction and demonstrations of this work.
The future of main memory appears to lie in the direction of new technologies that provide strong capacity-toperformance ratios, but have write operations that are much more expensive than reads in terms of latency, b...
详细信息
ISBN:
(纸本)9781538643686
The future of main memory appears to lie in the direction of new technologies that provide strong capacity-toperformance ratios, but have write operations that are much more expensive than reads in terms of latency, bandwidth, and energy. Motivated by this trend, we propose sequential and parallel algorithms to solve graph connectivity problems using significantly fewer writes than conventional algorithms. Our primary algorithmic tool is the construction of an o(n)-sized implicit decomposition of a bounded-degree graph G on n nodes, which combined with read-only access to G enables fast answers to connectivity and biconnectivity queries on G. The construction breaks the linear-write "barrier", resulting in costs that are asymptotically lower than conventional algorithms while adding only a modest cost to querying time. For general non-sparse graphs on m edges, we also provide the first parallel algorithms for connectivity and biconnectivity that require o(m) writes and O(m) operations. These algorithms provide insight into how applications can efficiently process computations on large graphs in systems with read-write asymmetry.
This paper introduces novel insights to improve the state of the art line-based unsupervised observation and abstraction models of urban environments. The scene observation is performed by an UAV, using self-detected ...
详细信息
ISBN:
(纸本)9781538612354
This paper introduces novel insights to improve the state of the art line-based unsupervised observation and abstraction models of urban environments. The scene observation is performed by an UAV, using self-detected and matched straight segments from streamed video frames. The increasing use of autonomous UAVs inside buildings and human built structures demands new accurate and comprehensive representations for their environment. Most of the 3D scene abstraction methods published are using invariant feature point matching, nevertheless some sparse 3D point clouds do not concisely represent the structure of the environment. Likewise, line clouds constructed by short and redundant segments with unaccurate directions will limit the understanding of the objective scenes, that include environments with no texture, or whose texture resembles a repetitive pattern. The presented approach is based on observation and representation models using the straight line segments, whose resemble the limits of an urban indoor or outdoor environment. The goal of the work is to get a better 3D representation for future autonomous UAV.
暂无评论