A team of robots guided by machine learning will take over the painstaking job of assembling lasers and other precision optical systems—work currently done by hand by highly skilled engineers, often with PhDs. This matters because lasers are critical components in aircraft gyroscopes, telecommunications networks, surgical tools, and manufacturing equipment. Each system contains dozens of optical components that must be placed with enormous accuracy; even a tiny misalignment can ruin performance or destroy the entire system. Today, assembly teams rely on diagnostic equipment and make minute adjustments by hand, making the process slow, expensive, and difficult to scale. Changing production volume or design specifications is particularly hard. If successful, this project will create an automated robotic system that combines observations of human operators, feedback from diagnostic tools, and machine-learning search algorithms to control the alignment process. The system would adapt to variations in parts, supply chain changes, or shifts in demand. This could fundamentally change how optical systems are designed and manufactured, potentially lowering costs, speeding production, and enabling performance levels that are currently out of reach.
View original technical description
Aircraft gyroscope, telecommunications, manufacturing, and surgical tools to name a few; optical systems, and especially lasers, are critical components in a host of modern devices. The manufacture of these systems supports a massive, global industry. Many of these are extraordinarily complex with dozens of optical components each of which needs to be placed in the system with enormous accuracy; any misalignment will result in poor performance, or the failure of the entire system. Currently this is accomplished by using highly qualified (even up to PhD level) and highly experienced system assembly teams who rely on a whole host of diagnostic and test equipment to make minute adjustments to the placement of each component. This is both time consuming and very expensive. It is also very difficult to modify production either terms of scale or specification. As a result these systems are very expensive and slow to respond to changing demand or potential for technical improvement. This project will develop an automated robotic and mechatronic system for assembling lasers and other optical systems. We will combine; observations of highly skilled human operators; feedback from automated diagnostic and test equipment; robotic alignment tool wielding robots; and a combination of machine learning and search algorithms which will be used to control the alignment process. The resulting system will be adaptive, able to cope with variations in part production, changes to the supply chain, modifications to the design specification, as well as being able to rapidly adapt to changes in demand. It will also result in a fundamental change to the way these systems are designed and developed and the levels of performance which can be achieved.
Matthew Esser (Co-Investigator)Mike Chantler (Co-Investigator)Mustafa Suphi Erden (Co-Investigator)Richard Carter (Principal Investigator)Xianwen Kong (Co-Investigator)
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know