Robotics is receiving a great deal of attention from industry, academia, and the general press. The literature in Robotics, however, is not easy to follow; the few articles published in journals or books are spread over a wide spectrum of disciplines and the motivational material linking the articles often is missing. The purpose of this column is threefold: to keep the general AI reader informed of progress in Robotics, to provide a source of references for in-depth exploration of topics in Robotics, and to supply some of the missing connections between the individual results.
Much of the early work on Robotics, especially in the use of sensors and models, was done at Artificial Intelligence projects such as those at Stanford, MIT, SRI, and JPL. Because of this, AI researchers could keep abreast of progress in Robotics as they did in other areas of AI: by glancing through the pages of the IJCAI proceedings or by getting technical reports from the few active laboratories. The rapid growth of all areas of AI and the diversification of research in Robotics has rendered this strategy ineffective. The difficulty stems from the variety of the technical fields relevant to robotics: mechanical design, electronics design, control engineering, computer systems, and artificial intelligence. Many key problems in Robotics are now viewed as belonging to these different disciplines; as a result, publications in the area have been distributed over a disconcerting variety of journals, conferences, and technical reports. Until very recently, there were no journals or books in which the varied areas of Robotics were treated as part of a single discipline. This situation is changing and one of the most exciting developments in Robotics today is the emergence of integrated approaches to several central problems in Robotics.
The principal goal of this column will be to identify relevant publications in Robotics and to assess their impact on other areas of Robotics and AI. In Artificial Intelligence 19 (1982) 137-143