We use a B/W CCD camera mounted on a robot arm, and a coordinate measuring machine (CMM) to sense the mechanical part. This allows us to combine vision, which can be relatively quick, and touch, which is typically slow but produces high accuracy. Vision is used not only to sense the part, but also to observe the probe, which must be manoeuvred to contact the part, running the risk of breakage. This combination requires control of disparate systems.
DEDS are suitable for modeling robotic observers as they provide a means for tracking the continuous, discrete and symbolic aspects of the scene under consideration [17][16][4]. Thus the DEDS controller will be able to model and report the state evolution of the inspection process.
In inspection, the DEDS guides the sensing machines to the parts of the objects where discrepancies occur between the real object (or a CAD model of it) and the recovered structure data points and/or parameters. The DEDS formulation also compensates for noise in the sensor readings (both ambiguities and uncertainties) using a probabilistic approach for computing the 3-D world parameters [19]. The recovered data from the sensing module is then used to drive the CAD module. The DEDS sensing agent is thus used to collect data of a passive element for designing structures; an exciting extension is to use a similar DEDS observer for moving agents and subsequently design behaviors through a learning stage.
A dynamic recursive finite state machines (DRFSM) implementation of a discrete event dynamic system (DEDS) algorithm is used to facilitate the state recovery of the inspection process. DRFSM are a particularly apt choice for the exploration of a set of machined parts which exhibit a recursive nature. As described below, we consider a part's visual features to be related as in tree structure. A feature which is contained within another feature will generally be of a smaller scale than its parent, even though it might be of the same type and require the same sensing technique. With the DRFSM implementation, the smaller scale feature can be explored using the same technique, but with more appropriate parameters. For more detailed information about experiments with DEDS and DRFSM, please see our related technical reports ([22][21][20]).
The CAD modeller we use is _1, from the University of Utah.
This modeller offers us a comprehensive representation for complex
geometries, as well as a diverse set of machined features
and interfaces to
automated milling and rendering tools. Among the machined features
supported by _1which are useful to our research are: