Advanced Robot Manufacturing Driving Growth through Increased

Production Flexibility

Larry M. SweetGeorgia Tech Institute for Robotics and Intelligent Machines

Explosive Robot Growth through 2025

• 11% CAGR unit growth through 2025

Source: BCG report “The Robot Revolution”, September 2015

• Cost reductions: • Robot HW / SW, perception• System engineering• Safety systems

0

10

20

30

40

50

60

70

80

2015 2020 2025

Global Robot Market, $B

Military

Industrial

Commercial

Personal

0

20

40

60

80

100

120

140

2015 2020 2025

Typical Total System Costs, $K

Robot

Peripherals

System engineering

Project management

Potential for Automation

Source: McKinsey & Company, “The Technical Potential for Automation in the U.S.”, 2016

D

D

Growth is Top Priority for Industry

• OEM’s, Tier 1 suppliers, SME’s all see growth as priority

• Increased flexibility: increase capacity for new customers and products

• Major positive economic impact through the entire supply chain

Will Robots Provide Sufficient Flexibility?

How to produce higher product mix demanded by customers?

How to delivery when they want it (and return it too)?

Advanced Robot Manufacturing Priorities

• Increase production rates 30% to 50%

• Mass manufacturing -> mass customization

• Multiple manufacturing processes

• Reduce cost of integration

• Humans and co-robots working in limited floor space

Potential for 30% to 50% Gains

-

0.5

1.0

1.5

2.0

Prior New

-

0.5

1.0

1.5

2.0

Prior New

-

0.5

1.0

1.5

2.0

Current New

Consumer goods(in production)

Automated DC(in production)

Aerospace(opportunity)

High Impact Emerging Technologies

Topics for today:• Autonomous mobility

• Fixtureless manufacturing

• Collaborative robots

• Virtual product / process development

Flexible Consumer Goods Production

• New material flow, food processes, controls, robotics

• All products, flavors, sizes on each line

• 50% improvement in asset utilization

• Sales growth through increased product mix

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

1

2

3

4

5

6

7

8

1

2

3

4

5

6

7

8

Production hours

Lin

esLi

nes

Optimized flexible system

Legacy system

50% boost

Why Are Mobile Robots Disruptive?

• New levels of performance• Transactions per hour

• Sequencing

• Storage density

• Easy to re-configure• Change in SKU mix or velocity

• Change in business model

• New algorithm challenges• Replace fixed hardware with flexible

software

• Leverage research on autonomous mobile robots

Sequencing with Autonomous Robots

1

3

5

2

4

12345Output sequence 1

3

5

2

4

1 12345Output sequence 2

5

3 1

2

4

54321Input sequence 1

1 2 3 4 5 6 7 8

Production order: 395 SKU, 476 cases

Aisle number

Fully Automated Distribution Centers

• DC’s for retail, grocery, food service, beverage

• 40% gain in storage density, sequenced output for mixed palletizing

• Reduce need for capital expansion, largely pays for automation

• Locate closer to urban centers for same day delivery

• Five to ten leading integrators moving to mobile robot technology

Aerospace Fixed Assets

Increase asset utilization• More units / month

• Lower fixed costs / unit

• Lower changeover costs

Painting

InspectionAssembly

Drilling

Composites

Flexibility Options

Collaborative robots Mobile robotsIndustrial robots

Flexible fixtures No fixtures

Assembly with Shims and Positioners

Aircraft assembly shims

Composite structure

Adjustable assembly positioners

Sources: Aviation Week, April 17, 2015; Duqueine Group, Airbus A350; Assembly Magazine, October 3, 2014..

3D Surface Scanning of Workpiece

• Measurements of mating surfaces

• Laser tracking

• Laser scanning

• Photogrammetry

• Machine excess material to net shape

Source: “ Interface Management in Wing Box Assembly”, B. Chuovion, A. Popov, S. Ratchev, C. Mason, M.

Summers, SAE 2011-01-2640.

Cutting path to eliminate clash between rib feet and cover (U. Nottingham, Airbus)

3D workpiece surface scan (Georgia Tech)

Fixtureless Aerospace Manufacturing

Technical challenges• Robot absolute position accuracy

• Robot compliance subject to machining forces

• Resonance excitation during machining

• Variable depth of cut due to workpiece tolerances

Future vision

Flexible Robotic Machining Technology Roadmap

• Robot, workpiece, fixture compliance models

• Redundancy resolution to maximize stiffness

• Real-time robot control

Shreyes add sensor photo here

Robot complianceMetrology Process knowledge

• Workpiece and feature

location in world, robot

coordinates

• 3D surface scanning

• Secondary encoders

• Wireless sensing for in-situ process monitoring

• Process modeling

• Sensor and process model integration into robot control

Metrology and Enhanced Accuracy

ElectroimpactTM contributions:• Role of laser tracking of EOAT for absolute accuracy

• Detailed kinematic and stiffness models

• Secondary encoders

• Demonstration of accurate drilling under +/- 0.125 mm

Source:. “Applied Accurate Robotic Drilling for Aircraft Fuselage”, R. DeVlieg and T. Szallay,

SAE Paper 2010-01-1836.

Laser tracking EOAT Joint secondary encoder Laser tracking EOAT

Compliance Optimization

• Detailed measurements of joint and link static stiffness

• Redundancy resolution exploits extra degrees-of-freedom to find optimal stiffness

• Optimization should be task specific, along vector of machining forces

Source: “Joint stiffness identification of six-revolute industrial serial robots”, C. Dumas, S. Caro, S. Garnier, B.

Furet, Robotics and Computer-Integrated Manufacturing 27 (2011) 881–888.

Source: T. Cvitanic, V. Nguyen, L. Sweet, Collaborative and Advanced Robotic Manufacturing Lab, Georgia Tech

Stiffness Optimization for Milling

Fraunhofer Institute (IPA), TU Stuttgart contributions:• Non-linear joint compliance experiments

• Redundancy resolution for optimal stiffness

• Open loop machining yields significant tolerance variation

Source: Julian Ricardo Diaz Posada, Ulrich Schneider, Arjun Sridhar and Alexander Verl, “Automatic Motion

Generation for Robotic Milling Optimizing Stiffness with Sample-Based Planning”, MPDI Machines, 2017, 5, 3.

Machining Process Sensor Feedback

Low cost, thin film PVDF wireless sensing for in-situ process monitoring of cutting forces up to 10 MHz

Machining process modeling

Sensor and process model integration into robot control yields > 70% error reduction

Wireless tool force sensor Undercut coordinates CMM measurements

Source: “A Wireless Force-Sensing and Model-Based Approach for Enhancement of Machining Accuracy in Robotic

Milling”, L.Cen, S.Melkote, J. Castle, H. Appelman, IEEE/ASME Trans. on Mechatronics, vol. 21, no. 5, Oct 2016.

Dynamics and Chatter Control

Georgia Tech contributions:• Congruent Convergent Transformation (CCT) based mode coupling chatter

avoidance

• Expands range of permissible tool feed orientation or workpiece orientation

Source: “ CCT-based Mode Coupling Chatter Avoidance in Robotic Milling ”, L.Cen, S.Melkote, Journal of

Manufacturing Process (in review)

Robot mounted milling tool Chatter avoidance demonstration

Roadmap to Accurate Machining

Future of Collaborative Manufacturing

Aerospace industry example

Standards for Safe Human Interaction

• ISO/TS 15066 Technical Specification• Safety-rated monitored stop

• Hand guiding

• Speed and separation monitoring

• Power and force limiting

• Still requires full risk assessment • Includes all equipment and processes in

cell

• following industry best practices (FMECA)

Series elastic actuators (ReThink)

Current / encoder based torque estimates (Universal)

Pressure sensitive skin (Kuka)

Load cell in base (Fanuc)

Power and force limiting approaches

Sequential Process Example

Collaborative Operations

• Collaborative operations• Robot and human perform tasks

simultaneously during task execution

• Sequential process flow• Most common implementation today

• Human and CoBot have separate tasks

• Buffers in between to prevent starving / blocking

• Task sharing• Human and CoBot work at team

• Requires higher level of coordination, communication, flexibility

Extending CoBot Perception to 3D

Part flat on table Part tilted

2D vision pick limited to parts on flat surface 3D vision tracking on flexible conveyor

3D Vision Tracking for Moving Targets

Robot Operating System (ROS) Enabled Integration

Universal Robot Kuka KR 500 ReThink Robot Kuka KR 210 (3)

Laser tracker / scanner Tool force sensor Milling head 3D vision tracker

Liquid cooled GPU-based controller

Collaborative and Advanced Robotic Manufacturing Lab (Georgia Tech)

Hardware Robot Operating System (H-ROS)

Source: “The shift in the robotics paradigm — the Hardware Robot Operating System (H-ROS), an infrastructure to

create interoperable robot components”, V. Mayoral, A. Hernandez, R. Kojcev, I. Muguruza, I. Zamalloa, L. Usategi,

A. Bilbao.

Multi-Level Human Collaboration

• Planning, scheduling, flow

• Re-planningTask level

• Robot path / process planning

• Mobile robot route planning

• Adaptive controlControl level

• Machines, processes, fixtures

• Machine, process, perceptionMachine level

Planning and control levels Human interaction

Contact

Larry M. Sweet

Associate Director

Georgia Institute of Technology

Institute for Robotics and Intelligent Machines

Collaborative and Advanced Robotic

Manufacturing Laboratory

Atlanta, Georgia

USA

770.862.1953

larry.sweet@cc.gatech.edu

www.gtcarmlab.net