Discrete Event Control Previous: A Graphical DRFSM

Experiment

In conducting our experiments, we use a B/W CCD camera mounted on a Puma 560 robot arm (see Figure 10), and simulate the operation of a CMM probe. Control signals that were generated by the DRFSM were converted to simple English commands and displayed to a human operator so that the simulated probe could be moved.

In order for the state machine to provide control, it must be aware of state changes in the system. As exploration takes place, the camera supplies images that are interpreted by a set of 2D and 3D vision processing algorithms and used to drive the DRFSM. These algorithms are described in greater detail in a technical report [4], but include thresholding, edge detection, region growing, stereo vision, etc. The robot arm is used to position the camera in the workplace and move in the case of occlusion problems. Our latest experiments used the robot and GIJoe-generated automata. One of them is described below.

The DRFSM generated by GIJoe is shown in figure 11. This machine has the following states:

• A: The initial state, waiting for the probe to appear.
• B: The probe appears, and waiting for it to be close.
• C: Probe is close, but not on feature.
• D: The probe visually appears to be on feature but physical touch with the CMM machine has not occurred.
• E: Physical touch has happened. If the current feature represents a closed region, the machine goes one level deeper to get the inner features by a recursive call to the initial state after changing the variable transition parameters. Otherwise, the machine looks for another feature on the same level.
• F: This state is used to solve any vision problem happens during the experiment. For example, if the probe is occluding one of the features, then the camera position can be changed to solve this problem.
• ERROR1: There is time limit for each part of this experiment to be done. If one of the modules does not finish within the limit, the machine will go to this state, which will report the error and terminate the experiment.

A part similar to the fuel pump cover from a Chevrolet engine was used in the experiment to test the exploration automaton. This piece offers interesting features and has a complex recursive structure. The piece was placed within view of the camera. Lighting in the room was adjusted so as to eliminate reflection and shadows on the part to be explored.

Some of the images from the experiment are shown in sequence in Figure 12.

Discrete Event Control Previous: A Graphical DRFSM