In this paper the problem whether or not a human being can plan a tour for a mobile robot with a camera so that all changes in the room are detected within a time limit is investigated. A tour simulator and a game based on it are developed. The plan of the


Annales UMCS Informatica AI 3 (2005) 263-271
Annales UMCS Informatica
Lublin-Polonia Sectio AI
doc.guandang.net
Experimental data driven robot for pattern classification
Krzysztof Sapiecha, Barbara ukawska*, Pawe Paduch
Department of Computer Science, University of Technology, Al. 1000-lecia Państwa Polskiego, 25-314 Kielce, Poland
Abstract
In this paper the problem whether or not a human being can plan a tour for a mobile robot with a camera so that all changes in the room are detected within a time limit is investigated. A tour simulator and a game based on it are developed. The plan of the winner of the game is used for a driving real robot. It is shown that this plan is two times shorter (as Java bytecode for Lego RCX) than the average one.
1. Introduction
Most today guarding systems use stationary cameras and mobile guards. However, while cameras are getting better and cheaper guards are still unreliable and very expensive. It is almost certain that in the near future mobile robots equipped with cameras will replace today guarding systems. In this new approach the robot will acquire information and also process it [1-5]. To do this effectively the robot should properly tour the area it guards and correctly classify patterns it sees.
Visibility-based problems of surveying, guarding, or searching have a long history in the area of computational optimization. Determining of minimum stationary cameras for guarding a region is the well-known art gallery problem [6]. The watchman problem [7] is to find the shortest tour along which one the mobile guard can see the entire region. If the region is unknown in advance, we are faced with the online watchman problem. All these problems are NP-hard [5].
In the paper an efficient method of planning a tour for a mobile robot in a room (including positions where pictures should be taken), so that all changes in the room are detected within a time limit is presented. The method is based on an experiment in which human intuition and perceptivity are involved. In section 2 the problem is formulated. Section 3 describes the method we use to solve the
*
Corresponding author: e-mail address: b.lukawska@tu.kielce.pl
Experimental data driven... 暂无评价 9页 免费 Affective pattern classi...... Higher order autocorrelations for pattern classification The use of higher-...
Data-Driven Language Und... 7页 免费 Experimental data driven... 暂无... Data-driven Classification of Linguistic Styles in Spoken Dialogues Language ...
Pattern Classification Using Polynomial Neural ...The assumption process is driven by fuzzy c-...As shown in results on the experimental studies,...
Section 5 will be dedicated for experimental ...pattern used in (Pan et al., 2002) and (...driven algorithms for highlight classification ...
This paper presents a swarming agent architecture for distributed pattern-detection and classification. Driven by the need for more efficiency and agility in ...
Classification, NN-nearest neighbor, NS-nearest ...(2010-) Experimental Results (1) Objects Airplane...机器学习模型 Data Driven Models 大脑如何处理与表示...
Keywords: Learning; Model Comparison; Evaluation Criteria JEL classification: ...Erev (1995), “Learning in Extensive Form Games: Experimental Data and ...
controller for hydraulically driven six degree of freedom parallel robot...Experimental validation of the above method is being investigated. ...
Model-Driven and Pattern... 暂无评价 32页 免费 Spatial process and data...Journal of Experimental Psychology: General, 127, 107–140. G¨ rdenfors, ...
1990. Data and model driven foveation. In ...Pattern Classification and Scene Analysis. John ...Experimental Brain Research 70 216-220. Goddard, ...
热门文档