Abstract
As the need for qualified Machinists increases so has the need for quality CNC machinist training One of the hurdles towards creating the pool of qualified candidates needed to fill these positions is the costs associated with running quality programs at the high school level. These cost concerns are not primarily associated with equipment purchase, but rather the insurance and ongoing costs that are often due to mishaps that occur while students use the machines. The most notable and frequent occurrence is spindle crash. Students are often able to successfully use CAM packages to generate tool paths that can cut out the parts they need, but frequently do not account for workholding equipment. This will often lead to collisions between the spindle and the workholding equipment when toolpath trajectories used for changing tools or even plunging different parts of the workpiece do not account for the workholding equipment. These occurrences can cause damages to the machine, workholding equipment and tooling and can also generate safety concerns which can be enough to shut down an entire program due to the strict safety guidelines employed by school districts. CAM packages offer the ability to set up workholding equipment in the virtual environment so tool paths can be tested against the possibility of collision. Students often choose not to take this approach because of the time commitment. This justifies the need for a cheap, effective online solution for collision avoidance. The solution would need to be easily adaptable to 3-axis CNC routers which are commonly used in high schools or other training environments. The essential question is how can we intelligently decide the difference between workholding equipment and the workpiece in a simple, safe and cost effective way? The proposed method is to use a marker-based vision system to detect potential tool crash and interrupt the executed g code inline. This would be done by attaching a vision camera on (or near) the spindle, and attaching one or more markers to the work-holding equipment, such that the angle and distance between camera and marker could be accurately measured and used to abort a running program if collision is imminent. The purpose of this work is to validate weather distance (XYZ coordinate system) and angle data can be accurately measured via vision camera in a variety of environments. Should this hurdle be effectively overcome, then further work could be justified towards implementing a fully functional inline system.