This article describes an expert system called ENGRAVE that produces musical notation. This task is a resource allocation problem involving the layout of musical objects in space. to do this, we use a spatial planning technique called constraint guidance to plan the layout of the music. an example is given to illustrate the technique. an additional example is given in another domain to show that constraint guidance has general applicability for spatial planning problems.

The results of a model-driven touch sensor recognition experiment are reported. The touch sensor used is a large-field tactile array. Object features appropriate for touch sensor recognition are extracted from a geometric model of an object, and a dual spherical image is formed. Both geometric and dynamic features are used to identify objects and their position and orientation on the touch sensor. Experiments show that geometric features extracted from the model are effective but that dynamic features must be determined empirically. Correct object identification rates, even for very similar objects, exceed 90%, a success rate much higher than would have been expected from only two-dimensional contact patterns. The position and orientation of objects, once identified, are very reliable. The authors conclude that large-field tactile sensors could prove useful in the automatic palletizing problem when object models (from a computer-aided design system, for example) can be utilized.>

The authors hypothesize the human-computer interaction can be significantly improved by using human-human communication protocols at the interface. These human communication principles are presented for a system built as a natural-language interface to an airline information and reservation system called simulated airline reservation agent (SARA). In particular, they present an extension to human communication theory that makes it possible to create a natural-language interface using human communication principles. Particular attention is given to the use of metacommunication to help manage the shared communication context. The current implementation can parse and generate >5000 surface versions of a single underlying request content, which can help the interface infer some of the user's goals and intentions. It is concluded that this technology can be used as a front end for any task domain, yielding much of the full expressiveness of English.>

A time/space planning system to coordinate the actions of two robot manipulators for transfer movements in a "sparse" environment is reported here. The collision avoidance reasoning guarantees that arms will arrive safely at their destination by temporally delaying or by altering the path of one arm. End effectors are constrained to follow elliptical motions. The performance of the system is sufficient in normal circumstances to drive an execution module in real time with tool tip speeds of about 3 in/s.

Three-dimensional object representation significantly impacts computations for manipulation, spatial reasoning and vision. This paper introduces a technique called the spherical dual image, derived from Gaussian sphere and dual space concepts. This representation technique will be defined and a number of its properties will be proven, leading to a reconstruction algorithm for convex and concave polyhedra. Finally, the application of this representation for selecting grasping points to manipulate a polyhedron is discussed.

A time/space planning system to coordinate the actions of two robot manipulators for transfer movements in a "sparse" environment is reported here. The collision avoidance reasoning guarantees that arms will arrive safely at their destination by temporally delaying or by altering the path of one arm. End effectors are constrained to follow elliptical motions. The performance of the system is sufficient in normal circumstances to drive an execution module in real time with tool tip speeds about three inches per second.

High quality human/computer interfaces have become a major topic of research. This paper describes a new method for modeling, designing, and developing dialogues, a method that has a strong formal basis and allows a uniform syntactic and semantic specification. This formal descriptive technique has the added advantage of being executable, that is, it has widely available translators. The technique chosen here allows a very high level specification of human/computer interaction enabling rapid development and easy modification. This paper describes the nature of the formal specifications written in first order logic using Prolog, and the successful specification and development of a carrier air traffic controller (CATC) dialogue. These experiments demonstrate the utility of Prolog as a high level specification language and point the way to a full dialogue development system that can incorporate a multi-layered concept of human/computer interaction.

As the complexity of human-computer interfaces increases, those who use these interfaces as well as those reponsible for their design have recognized an urgent need for substantive research in the human factors of software development [2], [5]. Because of the magnitude of the task of producing software for individual human-computer interfaces, appropriate tools are needed for defining and improving such interfaces, both in research and production environments. This paper describes the research being carried out to construct DMS (Dialogue Management System), which is a complete system for defining, modifying, executing, and metering human-computer dialogues.

A method is developed to represent movement of convex blocks in three-dimensional space from a sequence of two-dimensional camera images. The goals are to determine the objects' movement toward or away from the camera as well as left/right and up/down movement in the image plane and to build models of the blocks. The movement information is used as part of a hierarchical matching process that determines the correspondence of blocks between scenes.