Cognitive Robotics for Aircraft Disassembly

Over 30,000 commercial aircraft are estimated to be decommissioned over the next 20-years. It is identified recycling these aircraft is key to decarbonising the global aviation industry. The current aircraft dismantling is a 100% manual practice, posing many health and safety hazards, including but not limited to a) working at heights, b) the presence of hazardous materials, c) high-stress structural components, d) tool and equipment hazards, e) fire hazards, etc.

Further, the increased presence of carbon and glass fibre in modern aircraft significantly reduces the valorisation ratios, leading the aeronautical industry to deposit more than 40,000 tons of composite material waste in landfills annually.

The recent developments in the automation industry led to the emergence of collaborative mobile manipulator robots, particularly combining the capabilities of robotic arms with mobile platforms. Their human-robot-collaborative capabilities, combined with state-of-the-art Artificial Intelligence and Machine Learning (AI&ML) techniques, these platforms offer a unique exploitation opportunity in aircraft End-of-Life (EOL) management.

The disassembly cannot be considered the reverse of the aircraft assembly process. The multifold challenges of automated aircraft disassembly arise due to a) high uncertainty in aircraft EOL conditions, b) model-related variations, c) complexities in process planning and operations and d) external factors such as unavailability of complete knowledge of aircraft to the 3rd party dismantlers, market-driven influences, etc.

At a planning level, Automated Disassembly Sequence Generation remains a fundamental challenge for the robotics system. Here, the robot must identify components to be disassembled, exploit known priories, account for uncertainties and generate optimal disassembly procedures with minimal or no human intervention. The system must be robust to uncertainties, E.g., missing components, changes to components against the original Bill of Materials, etc.

At an operational level, the robot should generate safe robotics trajectories, adapt to dynamic environments in real-time, use optimal tool configurations, etc.

This research aims to research and develop enabling technologies to automatically dismantle EOL aircraft using mobile robotic manipulators to remove health and safety hazards, preserve the maximum value of the disassembled components and reduce environmental impact. The principle of cognitive robotics will be used to resolve the uncertainties in the EOL aircraft and disassembly processes.

This will generate a flexible robotic system capable of disassembling given aircraft components regardless of the exact product structure and geometry information. The fundamental technology will then be translated to a laboratory-scale demonstrator using a KUKA KMR mobile robotic platform and real aircraft components (provided by Spirit AeroSystems).

Key objectives:

a. Examine, research and develop a Cognitive Disassembly Sequence Planner robust to uncertainties using Mathematical, Heuristic and AI&ML models for EOL aircraft dismantling. b. Research and develop a novel robotic Disassembly Process Planner for EOL aircraft dismantling.

c. Examine and analyse viable sensor technologies (Ultrasound, X-ray, 3D vision, etc.) for the robotic system to sense and reason its environment to generate disassembly sequences.

d. For robotic disassembly path planning, exploit sensor technologies such as Laser, LiDar, Force-Torque, 3D vision, Vibrometers, etc. e. Research and develop novel reasoning and robotic path-planning algorithms for automated disassembly. f. Validate the outcome of a-e using Spirit Aerosystem part and mobile robotic manipulator (TRL 3) at SEARCH.

g. Translate the technology validated in f) to a case study at the A*STAR facility

h. Building on a-g, generate an industry-focused Proof-of-Concept at A*STAR and/or NMIS facilities to demonstrate robotic disassembly.