An on-line system for drill wear detection has been developed by using a sensor fusion strategy. Both acceleration and thrust signals were analyzed. Flank wear area was used to evaluate drill wear states. The drill wear area was measured by a vision system and classified into two groups: usable and worn-out. The wear prediction model was obtained by a two-category linear classifier. On-line detection tests indicate that the prediction model has over a 90 percent success rate.
On-line routing and wavelength assignment for dynamic traffic in WDM ring and torus networks