Crowdsourcing the construction of a 3D object recognition database for robotic grasping

Kent, David; Behrooz, Morteza; Chernova, Sonia

doi:10.1109/icra.2014.6907520

Cited by 15 publications

(13 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The study used nine objects, shown in Figure 1. Further details of the user study can be found in our previous work [1]. Additionally, we compare results against grasps demonstrated by an expert.…”

Section: A Data Collectionmentioning

confidence: 94%

“…Our object model construction algorithm conducts point cloud registration to combine each view of an object into a single model [1]. This algorithm uses a set of metrics to evaluate point cloud registration, which are used to intelligently determine the order in which to perform successive pairwise point cloud registrations based on a graph structure.…”

Section: B Model Constructionmentioning

confidence: 99%

“…This paper builds on our previous work in crowdsourced object recognition database construction [1], in which we conducted a user study to collect object recognition and grasping data, and presented a system to construct an object recognition database using that data. Here, we extend our previous work to include grasping data, while still meeting the original goals of the recognition database, i.e.…”

Section: Introductionmentioning

confidence: 98%

See 2 more Smart Citations

Construction of an object manipulation database from grasp demonstrations

Kent

Chernova

2014

2014 IEEE/RSJ International Conference on Intelligent Robots and Systems

Self Cite

View full text Add to dashboard Cite

Intelligent object manipulation is critical for a robot to effectively operate in a household environment. There are many grasp planners that can estimate grasps based on object shape, but these approaches often perform poorly because they miss key information about non-visual object characteristics. Object model databases can account for this information, but existing methods for database construction are time and resource intensive. We present an easy-to-use system for constructing a grasp database from crowdsourced demonstrations. The method requires no additional equipment other than the robot itself, and non-expert users can demonstrate grasps through an intuitive web interface, with virtually no training required. We show that the crowdsourced grasps can prove sufficient for object manipulation, and furthermore the demonstration approach outperforms purely vision-based grasp planning approaches for a wide variety of object classes.

show abstract

Section: A Data Collectionmentioning

confidence: 94%

Section: B Model Constructionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 98%

See 1 more Smart Citation

Construction of an object manipulation database from grasp demonstrations

Kent

Chernova

2014

2014 IEEE/RSJ International Conference on Intelligent Robots and Systems

Self Cite

View full text Add to dashboard Cite

show abstract

“…To learn different user preferences, we collect data from many non-expert users using a crowdsourcing platform. Prior work has also leveraged crowdsourcing for data labeling or as an efficient platform for transferring human knowledge to robots (Deng et al, 2013; Kent et al, 2014). For example, Sorokin et al (2010) utilized crowdsourcing to teach robots how to grasp new objects.…”

Section: Related Workmentioning

confidence: 99%

Organizing objects by predicting user preferences through collaborative filtering

Abdo

Stachniss

Spinello

et al. 2016

The International Journal of Robotics Research

View full text Add to dashboard Cite

As service robots become more and more capable of performing useful tasks for us, there is a growing need to teach robots how we expect them to carry out these tasks. However, different users typically have their own preferences, for example with respect to arranging objects on different shelves. As many of these preferences depend on a variety of factors including personal taste, cultural background, or common sense, it is challenging for an expert to pre-program a robot in order to accommodate all potential users. At the same time, it is impractical for robots to constantly query users about how they should perform individual tasks. In this work, we present an approach to learn patterns in user preferences for the task of tidying up objects in containers, e.g. shelves or boxes. Our method builds upon the paradigm of collaborative filtering for making personalized recommendations and relies on data from different users which we gather using crowdsourcing. To deal with novel objects for which we have no data, we propose a method that compliments standard collaborative filtering by leveraging information mined from the Web. When solving a tidy-up task, we first predict pairwise object preferences of the user. Then, we subdivide the objects in containers by modeling a spectral clustering problem. Our solution is easy to update, does not require complex modeling, and improves with the amount of user data. We evaluate our approach using crowdsourcing data from over 1200 users and demonstrate its effectiveness for two tidy-up scenarios. Additionally, we show that a real robot can reliably predict user preferences using our approach.

show abstract

“…As an example, this infrastructure has been exploited to construct a 3D object recognition database for grasping purposes by means of crowdsourcing, i.e. , by obtaining information from a large number of people through the Internet [ 16 ].…”

Section: Introductionmentioning

confidence: 99%

A Remote Lab for Experiments with a Team of Mobile Robots

Casini

Garulli

Giannitrapani

et al. 2014

Sensors

View full text Add to dashboard Cite

In this paper, a remote lab for experimenting with a team of mobile robots is presented. Robots are built with the LEGO Mindstorms technology and user-defined control laws can be directly coded in the Matlab programming language and validated on the real system. The lab is versatile enough to be used for both teaching and research purposes. Students can easily go through a number of predefined mobile robotics experiences without having to worry about robot hardware or low-level programming languages. More advanced experiments can also be carried out by uploading custom controllers. The capability to have full control of the vehicles, together with the possibility to define arbitrarily complex environments through the definition of virtual obstacles, makes the proposed facility well suited to quickly test and compare different control laws in a real-world scenario. Moreover, the user can simulate the presence of different types of exteroceptive sensors on board of the robots or a specific communication architecture among the agents, so that decentralized control strategies and motion coordination algorithms can be easily implemented and tested. A number of possible applications and real experiments are presented in order to illustrate the main features of the proposed mobile robotics remote lab.

show abstract

Crowdsourcing the construction of a 3D object recognition database for robotic grasping

Cited by 15 publications

References 9 publications

Construction of an object manipulation database from grasp demonstrations

Construction of an object manipulation database from grasp demonstrations

Organizing objects by predicting user preferences through collaborative filtering

A Remote Lab for Experiments with a Team of Mobile Robots

Contact Info

Product

Resources

About