A context-aware approach in realization of socially intelligent industrial robots. (February 2016)

• Record Type: Journal Article
• Authors: Stipancic, Tomislav; Jerbic, Bojan; Curkovic, Petar
• Is Part Of: Robotics and computer-integrated manufacturing. Volume 37(2016)
• Page Start: 79
• Page End: 89
• Publication Date: 2016-02
• Subjects: Cognitive robotics -- Context-awareness -- Probabilistic graphical models -- Ontology -- Ubiquitous computing; Robots, Industrial -- Periodicals; Computer integrated manufacturing systems -- Periodicals; Robotics -- Periodicals; Robots industriels -- Périodiques; Productique -- Périodiques; Robotique -- Périodiques; 670.285
• DOI: 10.1016/j.rcim.2015.07.002
• Languages: English
• ISSNs: 0736-5845

Abstract

Abstract: Contemporary industrial environments are usually constrained or limited in order to fit a fast, cheap and non-error prone production. Human-like system capabilities are not generally desirable there. But, recent trends in industrial robotics demand robust, flexible and efficient robots with a certain level of autonomy. Therefore, new and different approaches and perspectives in designing of industrial facilities are required. This paper reveals how a context-based reasoning can be used to achieve an intelligent robot group behavior. In order to achieve adaptivity, self-recovery or scalability of the system, a COgnitive MOdel for the Robot group control (COMOR) is developed. COMOR can be understood as an interpreter used to transform high-level context to low-level data, allowing machines to make context-based decisions. COMOR has three main parts and relies on a simulated Social Capital phenomenon as a feature of people. The first part is used to collect significant information from the environment. The second part is used to provide a set of possible solutions respecting the semantic domain description. The last part of COMOR is used to provide a behavioral component ensuring an optimal solution to given environmental conditions. Highlights: A cognitive model for a robot work (COMOR) on industrial assignments is proposed. A partial contextual awareness is achieved through three model components. Sensors are placed seamlessly within the environment to ensure ubiquitousness. Formal knowledge is stored in so-called the Core Ontology as a domain description. Social Capital is achieved in Bayesian Network to provide unambiguity in solutions.