The advent of low cost microcomputer hardware has provided the capability for sophisticated control of robot machinery, allowing such features as straight line trajectory motions, sensory interaction, workstation integration, external database access and off-line programming. The foundation of these capabilities lies in the architecture of the computer hardware and software system. This real-time control architecture must provide the framework for a logical partitioning and structuring of the control tasks into functionally separate modules that are bounded by well-defined data interfaces. It is through this architecture definition that robot controllers can be designed and build that allow enhancement of control strategies through upgrades in isolated modules, addition of sensors for real-time modification of behavior, and integration into workstations communicating with external knowledge based through the use of defined data interfaces. This paper will report on such an architecture for real-time control. The fundamental building block is the input-process-output structure. For this is derived a generic control level that can be stacked into multiple levels to provide an environment for hierarchical decomposition of the task. The more complex the task, the more levels that are required. Efficiency and reliability are obtained through a highly interactive user interface with diagnostic probing and display capability that allows quick evaluation and modifications as required to accomplish the intended tasks.
, Barbera, T.
and Albus, J.
Real-Time Control Systems for Robots, 1985 SPI National Plastics Exposition Conference, Chicago, IL, USA, [online], https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=820178
(Accessed June 7, 2023)