A Deep Incremental Boltzmann Machine for Modeling Context in Robots

Download
2018-05-25
Doğan, Fethiye Irmak
Çelikkanat, Hande
Kalkan, Sinan
Context is an essential capability for robots that are to be as adaptive as possible in challenging environments. Although there are many context modeling efforts, they assume a fixed structure and number of contexts. In this paper, we propose an incremental deep model that extends Restricted Boltzmann Machines. Our model gets one scene at a time, and gradually extends the contextual model when necessary, either by adding a new context or a new context layer to form a hierarchy. We show on a scene classification benchmark that our method converges to a good estimate of the contexts of the scenes, and performs better or on-par on several tasks compared to other incremental models or non-incremental models.
IEEE International Conference on Robotics and Automation (ICRA)

Suggestions

A behavior based robot control system using neuro-fuzzy approach
Öğüt, Demet; Alpaslan, Ferda Nur; Department of Computer Engineering (2003)
In autonomous navigation of mobile robots the dynamic environment is a source of problems. Because it is not possible to model all the possible conditions, the key point in the robot control is to design a system that is adaptable to different conditions and robust in dynamic environments. This study presents a reactive control system for a Khepera robot with the ability to navigate in a dynamic environment for reaching goal objects. The main motivation of this research is to design a robot control, which i...
A multitasking knowledge-based system for control applications
Tolun, Mehmet; Baykal, Nazife; Abu-Shaar, S. (1999)
Knowledge-based systems can provide several intelligent features for control applications, which decrease their dependency on human operators. As industrial systems become more complex, the response time and the amount of thinking required to control a large number of instruments far surpass the capability of humans. This paper describes a knowledge-based tool architecture that is supported by a multitasking inference engine and an interfacing hardware for data acquisition. The tool Features a high-level...
A control system using behavior hierarchies and neuro-fuzyy approach
Arslan, Didem; Alpaslan, Ferda Nur (null; 2005-09-14)
In agent-based systems, especially in autonomous mobile robots, modelling the environment and its changes is a source of problems. It is not always possible to effectively model the uncertainty and the dynamic changes in complex, real-world domains. Control systems must be robust to changes and must be able to handle the uncertainties to overcome this problem. In this study, a reactive behaviour based agent control system is modelled and implemented. The control system is tested in a navigation task using a...
A behavior based robot control system architecture for navigation environments with randomly allocated walls
Altuntaş, Berrin; Alpaslan, Ferda Nur; Department of Computer Engineering (2003)
Integration of knowledge to the control system of a robot is the best way to emerge intelligence to robot. The most useful knowledge for a robot control system that aims to visit the landmarks in an environment is the enviromental knowledge. The most natural representation of the robot2s environment is a map. This study presents a behavior based robot control system architecture that is based on subsumption and motor schema architectures and enables the robot to construct the map of the environment by using...
A developmental framework for learning affordances
Uğur, Emre; Şahin, Erol; Öztop, Erhan; Department of Computer Engineering (2010)
We propose a developmental framework that enables the robot to learn affordances through interaction with the environment in an unsupervised way and to use these affordances at different levels of robot control, ranging from reactive response to planning. Inspired from Developmental Psychology, the robot’s discovery of action possibilities is realized in two sequential phases. In the first phase, the robot that initially possesses a limited number of basic actions and reflexes discovers new behavior primiti...
Citation Formats
F. I. Doğan, H. Çelikkanat, and S. Kalkan, “A Deep Incremental Boltzmann Machine for Modeling Context in Robots,” presented at the IEEE International Conference on Robotics and Automation (ICRA), Brisbane, AUSTRALIA, 2018, Accessed: 00, 2020. [Online]. Available: https://hdl.handle.net/11511/37700.