Towards generalization of semi-supervised place classification over generalized Voronoi graph

Shi, Lei; Kodagoda, Sarath

doi:10.1016/j.robot.2013.04.013

Cited by 10 publications

(5 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We also make comparisons with the results we achieved in our previous work SPCoGVG [23]. SPCoGVG is also a semi-supervised approach, which is composed of support vector machine (SVM) and conditional random field (CRF) to ensure the generalization ability.…”

Section: Fusion Resultsmentioning

confidence: 93%

“…In our past work, we implemented a logistic regression based classifier, as a binary and multiclass problem contributing to higher accuracies [24], [25]. The work was further extended to address the generalizability of the solution through a semi-supervised place classification over a generalized Voronoi graph (SPCoGVG) [23]. In all of these methods, the features were extracted from the laser range data based on statistical and geometrical information, or so-called hand-engineered features.…”

Section: Related Workmentioning

confidence: 99%

“…The general representation of GVG is composed of meet-points (locations of three-way or more equidistance to obstacles) and edges (feasible paths between meet-points which are two-way equidistance to obstacles) [28]. We adopt the same resolution as in our previous work [23] to construct the first layer GVG, and then higher layers of GVGs are constructed to describe the environment at different levels of granularity.…”

Section: Multi-layer Construction and Decision Makingmentioning

confidence: 99%

See 2 more Smart Citations

Place Classification With a Graph Regularized Deep Neural Network

Liao

Kodagoda

Wang

et al. 2017

IEEE Trans. Cogn. Dev. Syst.

Self Cite

View full text Add to dashboard Cite

Place classification is a fundamental ability that a robot should possess to carry out effective human-robot interactions. It is a nontrivial classification problem which has attracted many research. In recent years, there is a high exploitation of Artificial Intelligent algorithms in robotics applications. Inspired by the recent successes of deep learning methods, we propose an end-to-end learning approach for the place classification problem. With the deep architectures, this methodology automatically discovers features and contributes in general to higher classification accuracies. The pipeline of our approach is composed of three parts. Firstly, we construct multiple layers of laser range data to represent the environment information in different levels of granularity. Secondly, each layer of data is fed into a deep neural network model for classification, where a graph regularization is imposed to the deep architecture for keeping local consistency between adjacent samples. Finally, the predicted labels obtained from all the layers are fused based on confidence trees to maximize the overall confidence. Experimental results validate the effectiveness of our end-to-end place classification framework in which both the multi-layer structure and the graph regularization promote the classification performance. Furthermore, results show that the features automatically learned from the raw input range data can achieve competitive results to the features constructed based on statistical and geometrical information.

show abstract

Section: Fusion Resultsmentioning

confidence: 93%

Section: Related Workmentioning

confidence: 99%

Section: Multi-layer Construction and Decision Makingmentioning

confidence: 99%

See 1 more Smart Citation

Place Classification With a Graph Regularized Deep Neural Network

Liao

Kodagoda

Wang

et al. 2017

IEEE Trans. Cogn. Dev. Syst.

Self Cite

View full text Add to dashboard Cite

show abstract

“…In this technique, what needs to be defined are nodes and paths. These nodes and paths are typically defined by using the Generalized Voronoi Graph [138][139][140][141][142][143][144]. In this case, the solution path is already deterministic at the end of the map creation.…”

Section: Topological Path Planningmentioning

confidence: 99%

Path planning for autonomous bulldozers

et al. 2019

View full text Add to dashboard Cite

With increasing demand and awareness from industries for enhanced safety, higher productivity, and better vehicle utilization, coupled with labour shortage, automating heavy equipment is becoming more and more important. Among the different types of heavy equipment, bulldozers fall into the category of machines that have a very high potential of achieving fully autonomous operation. Focusing on this fact, the aim of this research is to develop an automated path planning system for bulldozers. In general, an autonomous system is composed of terrain mapping, path planning and path-tracking. Existing commercial grade methodologies already solve terrain mapping and path-tracking, however path planning will be a key challenge for bulldozer automation.Conventional path planning algorithms are designed typically using a grid-based map, and are very versatile. However, in reality, a substantial effort is still required when applying to specific industry products. In contrast to this trend, the aim of this work is to develop a path planning algorithm specifically suitable for bulldozers, starting from the phase of theoretical development. The novel path planning methodology was developed by incorporating industry feedback, and has successfully resolved the issues which occurred when attempting to apply the pre-existing path planning methodologies. Results obtained from this newly developed methodology will demonstrate that efficient paths can be created. The methodology presented in this thesis has led to the development of automated path planning for bulldozers. This allows bulldozers to function without the need of an operator, contributing to enhanced safety, increased productivity, and solving the shortage of skilled operators.Declaration relating to disposition of project thesis/dissertation I hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or in part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all property rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation.I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstracts International (this is applicable to doctoral theses only). …………………………………………………………… Signature ……………………………………..……………… Witness Signature ……….……………………...…….… DateThe University recognises that there may be exceptional circumstances requiring restrictions on copying or conditions on use. Requests for restriction for a period of up to 2 years must be made in writing. Requests for a longer period of restriction may be considered in exceptional circumstances and require the approval of the Dean of Graduate Research.

show abstract

“…Spectral clustering and cluster growing were applied to extract node regions from gridmaps by representing navigable areas of the environment as a graphical model . Machine learning‐based algorithms can be applied to identify the door, room, and corridor regions . Although the above‐mentioned methods can extract subregions in various environments, most of them are focused on topological modeling or localization and are not appropriate for solving path‐planning problems.…”

Section: Related Workmentioning

confidence: 99%

Incremental hierarchical roadmap construction for efficient path planning

Park

Choi²,

Chung

2018

ETRI Journal

View full text Add to dashboard Cite

This paper proposes a hierarchical roadmap (HRM) and its construction process to efficiently represent navigable areas in an indoor environment. HRM is adopted to solve the path‐planning problems of mobile robots in indoor environments. HRM has a multi‐layered graphical structure that enables it to abstract and cover navigable areas using a smaller number of nodes and edges than a probabilistic roadmap. During the incremental process of constructing HRM, information on navigable areas is abstracted using a sonar gridmap when the mobile robot navigates an unexplored area. The HRM‐based planner efficiently searches for paths to answer queries by reducing the search space size using the multi‐layered graphical structure. The benefits of the proposed HRM are experimentally verified in real indoor environments.

show abstract

Towards generalization of semi-supervised place classification over generalized Voronoi graph

Cited by 10 publications

References 46 publications

Place Classification With a Graph Regularized Deep Neural Network

Place Classification With a Graph Regularized Deep Neural Network

Path planning for autonomous bulldozers

Incremental hierarchical roadmap construction for efficient path planning

Contact Info

Product

Resources

About