Collaborative Robotic Painting with Real-Time Coverage Modeling and Planning

Labori Systems LLC, a small, innovative developer of autonomous robotic systems, proposes to develop Medusa, an autonomous, stand-alone aircraft painting system that can accurately and rapidly paint an aircraft without the need for specialized facilities or detailed a-priori information about the aircraft.  Medusa simply rolls up to an aircraft, takes its own high accuracy measurements of the aircraft, and autonomously plans the painting job.  The operator is in a supervisory role and requires far less training than in traditional painting operations.        Accurate painting of aircraft is critical in managing lifecycle costs of the Air Force’s large fleet of aircraft.  Paint in not only expensive, but it adds weight to an aircraft that adversely affects fuels consumption.  Paint weight on an aircraft can reach several hundred pounds, and even small over-applications can result in significant process and operating costs.  Further, personnel on the paint line are exposed to hazardous chemicals, requiring specialized training and expensive safety measures.  Since properly trained human resources are typically scarce, throughput is often reduced and the training burden is high.  To address the technical and logistical challenges, Medusa uses a novel real-time paint coverage model to estimate the amount of paint applied along a surface over time with a given robotic motion path.  The real-time model lets the system estimate its current progress and effectiveness and is used with an advanced multi-agent planner to optimize motion of multiple robot arms to cover the aircraft surfaces effectively and rapidly.  Medusa’s physical base is a commercially available boom lift, that mounts a set of robotic arms, each with a paint sprayer.  The system is outfitted with sensors to ensure safety around aircraft and humans.  Each arm is outfitted with a depth camera that measures the surface geometry in the vicinity of the robot arm to millimeter accuracy.  That information is used with a short-horizon planner to optimize robot arm and nozzle motion.  Medusa can be outfitted with different nozzle types to address different geometries, and the planning algorithms account for these geometry differences.  We have generated coverage and control algorithms results that give confidence that the proposed solution is viable, feasible, and provides substantially greater value to the Air Force than typical static or gantry-mounted painting robots, which have significant facilities and planning requirements.  Labori Systems will bring to bear its expertise in developing mission-focused robotic systems with high safety standards. We have partnered with Scientific Systems Co., Inc, an industry leader in multi-agent planners.  SSCI’s planners have been developed to operate in difficult tactical environments and to provide graceful degradation of system performance under uncertainty and component failures and are well suited for this time-critical application.