Online adaptive rough-terrain navigation vegetation

Wellington, Carl; Stentz, Anthony

doi:10.1109/robot.2004.1307135

Cited by 55 publications

(48 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recent progress has been made by applying methods based on learning from examples, imitation or experience [15], [20], [26]. A commonly used concept in learning for autonomous navigation, known as learning from proprioception [20], [26], is to associate the terrain appearance observed from a distance with the mechanical observations made by the robot (e.g. if the terrain is traversable or not) when the corresponding location is traversed; this association is learned, thus allowing prediction of mechanical traversability properties from vision information only.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Learning slip behavior using automatic mechanical supervision

Angelova

Matthies

Helmick

et al. 2007

Proceedings 2007 IEEE International Conference on Robotics and Automation

View full text Add to dashboard Cite

Abstract-We address the problem of learning terrain traversability properties from visual input, using automatic mechanical supervision collected from sensors onboard an autonomous vehicle. We present a novel probabilistic framework in which the visual information and the mechanical supervision interact to learn particular terrain types and their properties.The proposed method is applied to learning of rover slippage from visual information in a completely automatic fashion. Our experiments show that using mechanical measurements as automatic supervision significantly improves the visual-based classification alone and approaches the results of learning with manual supervision. This work will enable the rover to drive safely on slopes, learning autonomously about different terrains and their slip characteristics.

show abstract

Section: Introductionmentioning

confidence: 99%

“…Although mechanical sensor measurements have been used to characterize terrain [5], [8], [20], [26], they have not been used to close the loop in a fully automatic vision-based learning framework and no principled approach for learning using automatic mechanical supervision has been considered.…”

Section: Introductionmentioning

confidence: 99%

Learning slip behavior using automatic mechanical supervision

Angelova

Matthies

Helmick

et al. 2007

Proceedings 2007 IEEE International Conference on Robotics and Automation

View full text Add to dashboard Cite

show abstract

“…The particular application addresses recognizing terrain types and inherent mobility related to robot slip using visual input, similar to [2], with the difference that learning is done with automatic supervision, provided by the robot, and does not need manual labeling of terrain types, as in [2]. Being able to predict certain mechanical terrain properties remotely from only visual information and other sensors onboard the vehicle has significant importance in autonomous navigation applications, because more intelligent planning could be done [16], [24].…”

Section: Previous Workmentioning

confidence: 99%

“…The collected training examples are used for learning of the mapping between the input visual and geometric features and the output slip. This strategy is commonly applied to learning traversability or other terrain properties from vision [2], [11], [24]. VO [15] is used for robot localization.…”

Section: Experimental Evaluationmentioning

confidence: 99%

Dimensionality Reduction Using Automatic Supervision for Vision-Based Terrain Learning

Angelova¹,

Matthies²,

Helmick³

et al. 2007

Robotics: Science and Systems III

View full text Add to dashboard Cite

Abstract-This paper considers the problem of learning to recognize different terrains from color imagery in a fully automatic fashion, using the robot's mechanical sensors as supervision. We present a probabilistic framework in which the visual information and the mechanical supervision interact to learn the available terrain types. Within this framework, a novel supervised dimensionality reduction method is proposed, in which the automatic supervision provided by the robot helps select better lower dimensional representations, more suitable for the discrimination task at hand. Incorporating supervision into the dimensionality reduction process is important, as some terrains might be visually similar but induce very different robot mobility. Therefore, choosing a lower dimensional visual representation adequately is expected to improve the vision-based terrain learning and the final classification performance. This is the first work that proposes automatically supervised dimensionality reduction in a probabilistic framework using the supervision coming from the robot's sensors. The proposed method stands in between methods for reasoning under uncertainty using probabilistic models and methods for learning the underlying structure of the data.The proposed approach has been tested on field test data collected by an autonomous robot while driving on soil, gravel and asphalt. Although the supervision might be ambiguous or noisy, our experiments show that it helps build a more appropriate lower dimensional visual representation and achieves improved terrain recognition performance compared to unsupervised learning methods.

show abstract

“…As an example of the first approach, the authors of [6] described a learning system for ground height estimation, in which large amounts of labeled data were generated by simply driving a vehicle over interesting terrain. Their solution is highly practical because driving a vehicle represents an easy method of labeling data, and because the system can benefit from adapting its parameters online.…”

Section: Introductionmentioning

confidence: 99%

Active Learning For Outdoor Obstacle Detection

Dima¹,

Hebert²

2005

Robotics: Science and Systems I

View full text Add to dashboard Cite

Abstract-Real-world applications of mobile robotics call for increased autonomy, requiring reliable perception systems. Since manually tuned perception algorithms are difficult to adapt to new operating environments, systems based on supervised learning are necessary for future progress in autonomous navigation.Data labeling is a major concern when supervised learning is applied to the large-scale problems occuring in realistic robotics applications. We believe that algorithms for automatically selecting important data for labeling are necessary, and propose to employ active learning techniques to reduce the amount of labeling required to learn from a data set.In this paper we show that several standard active learning algorithms can be adapted to meet specific constraints characteristic to our domain, such as the need to learn from data with severely unbalanced class priors. We validate the solutions we propose by extensive experimentation on multiple realistic data sets captured with a robotic vehicle. Based on our results for the task of obstacle detection, we conclude that active learning techniques are applicable to our domain, and they can lead to significant reductions in the labeling effort required to use supervised learning in outdoor perception.

show abstract

Online adaptive rough-terrain navigation vegetation

Cited by 55 publications

References 9 publications

Learning slip behavior using automatic mechanical supervision

Learning slip behavior using automatic mechanical supervision

Dimensionality Reduction Using Automatic Supervision for Vision-Based Terrain Learning

Active Learning For Outdoor Obstacle Detection

Contact Info

Product

Resources

About