Next: References Up:

Industrial Inspection and Previous: Summary


Conclusions and Future Work

The current implementation has allowed us to explore the constraints needed to make progress in the reverse engineering problem. We have come to the conclusion that a machined feature representation is a worthwhile target and may be useful in determining sensing strategy. Interacting features have been identified as a source of complexity. This information will be used to construct a new reverse engineering system.

In our current implementation, each contour is treated as at most one machined feature (some contours may be islands and therefore part of another contour's feature). Future work may allow contours to be made from multiple features if appropriate. For example, combinations of drilled holes, slots, and pockets may be produced (see Figure 7), based on a machining/inspection time/cost analysis which will include such factors as time needed to select and load tools (operator), change tools, etc. This problem has some characteristics that may be best solved through artificial intelligence or optimization techniques.

However, for the reverse engineering problem, autonomy may not be an important requirement. By allowing interaction with a machinist, we could overcome the problems of features interacting with each other and more complex geometries. A conceptual graph for such a system is shown in Figure 8. Some support for this type of system is found in [4] and [1].

Such system could provide valuable information about how to build reverse engineering systems in the future. For example, it could be used to perform protocol analysis to determine how an expert machinist solves the reverse engineering problem. Lessons learned would enable us to build a robust autonomous system.

Acknowledgements

This work was supported in part by ARPA under ARO grant number DAAH04-93-G-0420, DARPA grant N00014-91-J-4123, NSF grant CDA 9024721, and a University of Utah Research Committee grant. All opinions, findings, conclusions or recommendations expressed in this document are those of the authors and do not necessarily reflect the views of the sponsoring agencies.



Next: References Up:

Industrial Inspection and Previous: Summary



sobh@bridgeport.edu
Mon Sep 19 19:38:55 MDT 1994