Carnegie Mellon University Robotics Institute Homepage

Carnegie Mellon University Robotics Institute Research Guide

Carnegie Mellon University, Robotics Institute, Research Guide

Manufacturing

RI has a long history of research in manufacturing, with work ranging from the development and application of robotics technologies for automation, innovation and improvement of manufacturing processes to the development of software technologies for support of “white collar” activities such as managing factory operations, coordinating interactions with supply chain partners, and product and facility design. The current trends in manufacturing continue to drive the small batch production of products. To make this cost effective, the time to setup machines must be minimized by: scheduling products with similar setup requirements one after the other, designing products to minimize the use of specialty tools and setup changes, and to partially automate aspects of the setup process itself. To accomplish partially automated setups, we have focused on computational tools to allow robots to follow human instructions (RHI) and to allow humans to follow robot instructions (HRI) based on maximizing the utility of the natural abilities of either the human or robot.

Manufacturing research at RI aims generally at the creation of technologies that enable more flexible and more efficient manufacturing operations. One broad challenge is “custom manufacturing” – the ability to produce small quantities of customized (unique) parts rapidly and effectively. Some of our faculty members are focusing on vertically integrated unit processes that compress and automate processes of transforming designs into physical artifacts for particular classes of products (Bourne). Others are emphasizing aspects of manufacturing agility in multi-stage manufacturing processes, including rapid configuration of product-specific “minifactories” (Hollis), and infra-structure and collaboration mechanisms for creation and coordination of “virtual organizations” (Smith, Sycara).

Another area looks at decreasing setup time, so as to allow more flexible manufacturing as the market demands change. We see this in our autobody painting work (Choset) and flexible manufacturing for aircraft assemblies (Choset). For the former, it is well established that robots are widely used for spray-painting in the automotive industry. Owing to a lack of fully automated trajectory generation tools, paint specialists often require 3 to 5 months to produce the coverage trajectories for a new automobile model. This programming time is a critical bottleneck in the “concept-to-consumer” timeline for bringing a new automobile to the market. Automating the process of path planning (Choset) helps the paint specialists significantly reduce this programming time by offering them reasonable guidelines for effective paths. In the latter case of manufacturing in confined spaces, such as wing-boxes on airplanes, is a tedious and time-consuming process which is prone to error and injury, both incurring a high cost. Highly articulated robots can reach into difficult to access areas to do work which require people to contort themselves to achieve or perhaps disassemble just to reassemble again. We have developed strong, yet precise, articulated robots to carry out manufacturing tasks in the aerospace industry.

Current projects in manufacturing at RI continue to reflect this broad perspective and include the development of new manufacturing processes for micro-electromechanical systems (MEMS), novel high-precision modular robots, and the design of execution-driven planning and scheduling tools for small manufacturing enterprises. Some research focuses generally on execution-driven logistics management, where planning and execution processes are tightly coupled through real-time data streams to the current factory state and current performance trends. One recent line of work seeks to couple data mining and machine learning techniques with scenario-based optimization procedures, to develop production and supply network plans that accurately reflect the demand/supply patterns and uncertainties evident in historical manufacturing data. Other work is developing algorithms for coordinating the movements of multiple, material-handling robots for just-in-time delivery of tooling, fixtures and parts in a mobile manufacturing assembly setting (Smith).

Research in MEMS-based manufacturing processes and technologies (Fedder) will provide new opportunities for custom manufacturing of highly integrated microsystems. The applicable systems span inertial sensor arrays for personal navigation to miniature biomedical implants. A relatively new branch of related research is exploiting MEMS manipulators to create highly parallel tip-based nano-manufacturing systems (Fedder, Sitti). The potential exists in the future to embed these emerging micro- and nano-scale manufacturing capabilities into the "mini-factory" system (Hollis). Future commercial growth of 3D- heterogeneous microsystems will increase motivation for agile robotics manufacturing at these scales.

For example, we are automating the assembly of cell phones, which is focused on the "last inch of motion" while mating parts (Bourne). The fine motion plan for assembly requires multiple sensory modes to control the assembly process, which can later be analyzed to determine assembly success or failure. Furthermore, dexterous and modular serial-link snake-like manipulators are being considered for use in aircraft assembly and inspection (Choset).

Manufacturing is an increasingly critical research area due to its position as a U.S. strategic security issue. Trends toward globalization in recent years have forced companies to focus on core competencies and rely on partners to supply other components and capabilities. This has enabled other emerging countries to capture increasing share in mass-produced, non-unique product manufacturing markets. Advancement of technologies that enable custom manufacturing capabilities is key to re-energizing US manufacturing competitiveness.

The current RI manufacturing faculty is comprised mainly of senior people, and to sustain leadership in this technology area into the future, it is important that RI recruit new junior faculty in manufacturing. In conjunction, it is an opportune time to consider a broader initiative in the custom manufacturing area. For example, there are now a number of mature component technologies at RI that could be brought together to achieve a magnifying effect on productivity and provide a visionary integrated framework for on-demand manufacturing.

Manufacturing research in RI collaborates with other manufacturing-related activity on the Carnegie Mellon campus. Some of our faculty (Smith, Sycara) participate in a joint Ph.D. program in Robotics and Management of Manufacturing Operations with the Tepper School of Business. Other collaborations include the joint SCS/Tepper School MS program in Electronic Commerce.

Continue Reading: Medical Robotics


Faculty

  1. David
    Bourne

  2. Howie
    Choset

  3. Gary
    Fedder

  4. Ralph
    Hollis

  5. Metin
    Sitti

  6. Stephen
    Smith

  7. Katia
    Sycara


Project Images

  • Aircraft/Automobile Painting

  • Aircraft Assemby & Inspection

  • Boeing Arm

  • Cell-Phone Assembly

  • Minifactory

  • Scenario-Based Manufacturing Logistics Planning