Novel optical instrumentation for robotic manufacturing

The aim of this proposal is to undertake research into novel optical instrumentation to support future manufacturing aims for more agile and flexible manufacturing systems as well as to improve precision of traditional robot-based manufacturing operations. New optical instrumentation will be developed combining two complimentary optical measurements techniques, range-resolved interferometry and laser speckle pattern correlation. These will be capable of remote, high quality and high-bandwidth measurements of the relative motion and orientation between a robotic end-effector mounted sensor and the workpiece, cumulating in a combined multi-parameter sensor.

The applications of these sensors are common to many areas of manufacturing and can be grouped into three main areas:
i) The stabilisation and control of the end-effector relative to the workpiece.
ii) Motion tracking to lock the robot to a moving assembly line.
iii) Relative positioning operations.
In these application areas there is a need for new effective, low cost instrumentation to provide real-time feedback about their relative motion and orientations, which this proposal will attempt to address.

In the first application area, the stabilisation and control of a tool's motion, positioning and orientation is required in the presence of end-effector and/or workpiece vibrations. For example contact tasks such as polishing, drilling, and riveting are often challenging tasks due to the comparatively low mechanical stiffness of typical industrial robots. This can result in excessive tool-tip deflection, vibration, and resulting poor quality of the finished part. Other methods of monitoring the motion suffer from limitations; vision systems can have limited update rates, whilst laser scanning/tracking systems are expensive and inflexible in that the scanning system needs to be mounted externally to the robot and maintain a line-of-sight.

The motion tracking of robotic manipulators to follow moving assembly lines is an important application in many manufacturing operations. Currently vision-systems can provide a flexible sensor capable of determining the relative position and orientation of objects to good precision, but their slow update rate and latency make them unsuitable for many feedback loops for real-time motion control. In addition, these visual systems sometimes require visual markers or beacons for operation and unless the cameras are mounted directly on the end-effector, the field-of-view can sometimes be obscured towards the end of the manoeuvring operation, at the point where control is most critical.

Finally, in the area of relative positioning many applications require high accuracy relative to a reference and this makes the adoption of standard robots difficult. For example, in airframe manufacturing the high accuracies of between 0.2 mm and 0.02 mm relative to a local reference a few meters away are required for drilling, fettling and component location operations and in remote laser cutting there is a requirement for high precision in the positioning of the cutting beam between scans, with any misalignment leading to multiple grooves and failure to cut the workpiece. Here laser speckle correlation may offer a solution with the proposed sensors capable of precise relative positioning between the end-effector and workpiece (potentially down to <1 micron in mm's and <50 micron in m's) while relying on other means, such as the robot encoders or vision systems, to reference to an absolute position.

In many areas of robotics there is requirement for robots capable of performing complex tasks autonomously, and this is of great importance in manufacturing to fulfil the desire to move to more agile and flexible manufacturing systems, permitting smaller production runs and faster reconfiguration to new product lines. One proposed means of achieving this is via a new generation of flexible robot systems, with a higher degree of autonomy. As such the positioning and control of robotic end-effectors with respect to the workpiece is of great importance and there is a need for new effective, low cost instrumentation to provide real-time feedback about their relative motion and orientation.

This is an applied research proposal, requesting funds to research and construct novel optical instrumentation to support future manufacturing aims for more agile and flexible manufacturing systems but also to increase the operating precision of traditional robotic manufacturing operations; by providing new instruments capable of remote, high quality and high-bandwidth measurements of the relative motion and orientation between a robotic end-effector mounted sensor and the workpiece. The ultimate goal for such a project is the development of the resulting prototypes into commercial devices for the technical instrument market, benefiting both scientific research and the UK economy. At this stage, the aim is to demonstrate that the technology can be successfully implemented.

At first the potential impact will be mainly academic, providing manufacturing and automation researchers both in academia and industry with a new sensing technique. The intended applications of these sensors are common to many areas of manufacturing and can be grouped into three main areas; i) the stabilisation and control of the end-effector relative to the workpiece, ii) motion tracking to lock the robot to a moving assembly line and iii) moderate-to-high accuracy relative positioning operations. As the applications of the sensors are wide there is potential impact across a broad range of manufacturing areas. For example in automobile assembly lines it is desired to have robotic manipulators follow moving assembly lines whilst performing operations. In this area vision systems are limited by high processing requirements and their need for visual markers, which can become obscured during operation, while laser scanner systems are accurate, but are expensive, less flexible and have limited update rates. The proposed sensors will offer complementary data providing fast tracking information at higher data rates. In other areas the stabilisation and control of a tool's motion, positioning and orientation in the presence of end-effector and/or workpiece vibrations is required to ensure quality of the manufactured parts, in particular contact tasks such as polishing and drilling are often challenging due to the comparatively low mechanical stiffness of typical industrial robots. The provision of new, low cost, sensors capable of high date rates will enable researchers to investigate new control implementations. Later, economic and societal impact is envisaged in two ways, through the economic exploitation of the sensor concept by UK instrumentation companies, and by using the sensor as enabling technology to improve manufacturing processes within the UK manufacturing community.

To facilitate uptake and implementation of the sensors by researchers investigating solutions to specific manufacturing problems, we will publish open-access papers in relevant journals and conferences. In addition, reference designs and algorithm implementations will be made openly available and technology demonstrator systems will be constructed in the project. For dissemination to the wider manufacturing community, it is intended to demonstrate the sensor capabilities at manufacturing and automation conferences, trade fairs and exhibits, as well as to potential collaborators and end-users.