Physical modeling of tools necessary for robot manipulation

Abstract

Previous research on modeling general processes has focused on physical and empirical modeling of dedicated process machines which have sufficient stiffness and accuracy. For example, machining centers have relatively high stiffness and exhibit negligible deflections. Robot manipulators have low stiffness which easily allows undesirable deflections under large reaction forces. Also, models for intuitive decision surfaces for users, decision making systems, or system controllers have not been embedded or otherwise deployed effectively. The objective of this research is to suggest graphical and parametric models of robotic processes suited for intuitive user friendly graphs and modeling nonlinear systems and to initiate a program which proves usefulness of performance maps. Performance maps are primitive surface representations which can be used to create decision surfaces. General robotic processes considered are robotic drilling, grinding, deburring, chiseling, sawing, peg insertion, force-fit insertion, forming for assembly, screw fastening, and riveting. To achieve the objective, a framework for in-depth parametric and analytic modeling of robotic processes is presented. First, relevant process parameters such as process operating variables, system condition parameters, and process performance criteria are defined. Process operating variables are the robot controller inputs. System condition parameters are process parameters which define system constraints and characteristics. Process performance criteria are critical parameters to define, anticipate, and evaluate product quality, system stability, economic performance, and system performance. Second, performance maps which describe the graphical relationship of the relevant process parameters are developed. A performance map is the surface representation of a process performance criterion as a function of process operating variable(s), system condition parameter(s), or other process performance criteria. Third, the application of performance maps of robotic drilling is simulated, to illustrate their advantages. Whether robot deflections generated during the process satisfy tolerance requirements or not is evaluated and suitable robot postures are recommended. Pertinent literature is extensively reviewed to recognize current research trends in modeling and parameterizing relevant process parameters. Approximately, 100 performance maps were created as graphical and parametric models based on process performance. Two performance envelopes were developed. Proper robot postures were suggested to drill a hole with constant feed-rate or bounded ranges of that feed-rate.