Phase II Amount
$4,800,000
The authors establish a novel metrology system capable of determining the pose of a robotically-mounted End of Arm Tool (EOAT) in Six Degrees of Freedom (6-DOF) to a capability better than True Position (TP) 0.010 in. The metrology system consists of low-cost, modular components in the form of motorized gimbals that steer laser beams onto robot-affixed, photosensitive chips typical of those found in digital cameras. The system was conceived to provide real-time pose information of an EOAT to better than True Position (TP) 0.0035 in. at the Tool Center Point (TCP). The system is scalable, deployable, and affordable, making it a viable candidate for guiding machines, particularly articulated arm robots, to aerospace tolerances over large volumes. Monte Carlo simulations, which propagate error through the system out to the EOAT, have shown the ability to determine the TCP to no worse than TP 0.002 in. over the volume of an F-35 Lightning II Wing Box robotic drilling system. Manufacturing Readiness Level 5 (MRL-5) testing of the system in a less than optimal environment showed it capable of determining the TCP to TP 0.0119 in. when compared to a laser tracker device. The current system updates the EOAT pose at 10 Hz. Significant improvement in both speed and accuracy will be achieved with minimal engineering changes to the system components, along with better mounting hardware and a more manufacturing-caliber test environment. It is intended that such a system will enable further penetration of Commercial, Off-the-Shelf (COTS) articulated arm robots into applications that previously required Numerically Controlled (NC) gantry machines, resulting in either less expensive systems or systems that provide significant span time reduction for approximately the same cost. In addition, the system could also be used to retrofit NC machines so that they achieve higher accuracies; to align large structures such as fuselages and wings during aerospace assembly; and to serve as a tool that enables ''birth certificate'' modeling of enhanced accuracy robots.
Benefit: As mentioned previously, VRSI will be using a laser tracker as the benchmark for the novel guidance technology. The main points of comparison include cost, accuracy, speed, reliability, and scalability. With regard to cost, a typical laser tracker solution for robot guidance is ~ $210k not including integration services. This includes the laser tracker ($120k), optical corner cube clusters that need to be attached to the EOAT ($15k per EOAT), twenty additional Spherically Mounted Retroreflectors (SMRs) required as fixed work cell ''fiducials'' ($1k ea, for $20k + $10k for support tooling), laser tracker mounting hardware ($10k if stationary tracker at a single position), and controls software ($30k). Of course a spare tracker would add another $120k. VRSI''s AARG solution comes in at ~ $197k for hardware and software. This includes four Beacons (one arc second accuracy at $25k ea, for $100k), eight Active Targets attached to the EOAT ($1.5k ea, for $12k per EOAT), twenty additional Active Targets required as fixed work cell ''fiducials'' ($1.25k ea, for $25k + $10k for support tooling), Beacon mounting hardware ($20k), and controls software ($30k). A spare Beacon would add another $25k. So although the upfront cost is not necessarily compelling, it becomes more so when spare parts are considered ($330k versus $222k). Keeping with the one arc second accurate Beacons, computer simulations have shown that VRSI''s AARG system should perform as accurately as a laser tracker in determining the position of the EOAT TCP, at around TP 0.0035 in. This comparison takes into consideration published laser tracker accuracies in conjunction with the typical geometry between the EOAT corner cube cluster targets and the TCP. If greater accuracies are desired, a half arc second Beacon is currently available for $45k ea, pushing the system cost from $197k to $277k, but allowing the TCP to be determined to an accuracy of TP 0.002 in. The AARG system currently generates the 6-DOF pose of the EOAT at 10 Hz. Our experience with laser trackers shows a typical pose acquisition time of eight seconds. Assuming that each of the F-35 Wing Box holes will require two EOAT pose measurements (initial measurement, then validation measurement following a corrective offset), this comes to a guidance reduction of just under four hours per Wing Box surface (2 measurements x 950 fasteners x 7.5 sec/fastener = 237.5 min) if a single robot is being used to drill. This advantage is lessened as more robots are added to the drilling cell, as the other robots may be busy drilling while one is being guided. Concerning reliability, laser trackers are reasonably robust. However, as VRSI currently owns five laser trackers and has used them for years, it is not uncommon for a laser tracker to need emergency service once per year. The Beacons that VRSI has employed in Phase I come directly from the field of land surveying, and are designed for outdoor conditions involving high humidity and temperature variations (arctic versions are available), commonly being used in mining and construction applications. Since thousands of them are in use across the world, it is safe to consider them as being MRL-10 devices. The modularity of the AARG solution combined with the lower component cost per line of sight affords it an advantage over laser trackers when it comes to scalability. VRSI has also outlined procedures for ''plug and play'' swapping of defective Beacons and Active Targets, and would like to validate these methods as part of a Phase II effort.
Keywords: Affordable, Accurate, Metrology, Robotic, Drilling, Milling
