Safety of Human-Robot Systems in Flexible Factory Environments

Industry is interested in leveraging the dexterity and versatility of people and the precision and repeatability of robots by enabling collaboration in dynamic and reconfigurable manufacturing environments.  Such collaborations, however, are not possible today.  Robots are still not capable of safely interacting cooperatively with their human coworkers in highly variable task scenarios.  This project will focus on developing the measurement science for characterizing and modeling interactions between humans and robots in flexible factory environments.  Project results will improve quality control, productivity, and safety of both robots and humans.

Objective

To develop the measurement science for characterizing and modeling the safety of collaborations between humans and next-generation robots in flexible industrial production environments by 2014.

What is the new technical idea?

Current robot standards focus on making robots safe enough to operate in close proximity to humans.  Collaborative robots use advanced sensors and mechanical designs to ensure they can observe and touch their environment safely, but guarantee safety only by limiting their interactions with humans.  Manufacturers want to implement robotics in flexible factory environments in which the robots and people can move about freely to accommodate shifting production requirements.  Successfully integrating next-generation collaborative robots with human coworkers rests on the capacity to measure and understand their interactions and react accordingly.

The new technical idea is to develop methods for modeling and evaluating the performance of safety systems for robots in flexible environments.  This project will specifically develop

1.   modular physical and virtual test beds for measuring and modeling human-robot interactions for collaborative tasks including small-parts assembly and bin picking, and
2. performance measures for testing and characterizing robot safety systems involving humans working with mobile manipulators.

The test methods and metrics developed in this project will provide robot vendors and users with the confidence that collaborative robots are capable of providing the required levels of safety while maintaining the ability to work productively.

What is the research plan?

The project will develop and integrate sensor-, world-, and task-based models and procedures for the characterization of human-robot interactions for safety in flexible manufacturing environments.  The research plan consists of the following component tasks:

• Develop state-of-the-art physical and virtual robot test beds.  A joint test bed consisting of physical and simulated components is needed for perception and safety research. The test bed will integrate force- and position-measuring sensors, simulated human avatars, and state-of-the-art robot controls to help create vendor-neutral manufacturing measurement techniques for the evaluation of robot safety.
• Design, implement, and evaluate task-driven manufacturing scenarios to characterize safety performance. The nature and risks of extended interactions between robots and humans are not well understood.  A framework consisting of parameterized task interactions and measurement strategies for evaluating a robot’s passive and active safety features is needed to remedy this deficiency.  Such a framework will be developed based on safety performance during human-robot collaborations toward joint task completion, and correlations between task motions and human-robot safety will be established.
• Design and integrate a mobile manipulator.  Current safety standards do not account for the risks posed by a robot arm mounted to a mobile base.  This project will develop measurement science to address the unique safety issues mobile manipulators present to flexible and collaborative workspaces.  Scenarios involving a variety of arm and mobile base motions and sub-tasks will be explored.

Recent Results: 

Outcomes: 

• Designed and constructed a reconfigurable work cell and safety system for robot safety performance testing and evaluation.
  
• Developed requirements for performance measures and standards for safety systems for mobile manipulators.
  

• Designed metrics and procedural documentation to support conformance tests for robot safety standards.
  
• Papers published:
• • W. Shackleford, S. Szabo, R. Norcross, and J. Marvel.  “Integrating Occlusion Monitoring into Human Tracking for Robot Speed and Separation Monitoring.” Proceedings of the Performance Metrics for Intelligent Systems (PerMIS) 2012 Workshop.  20-22 March, 2012.
    
• Papers awaiting publication:
• • J.A. Marvel.  “Performance Metrics of Speed and Separation Monitoring in Shared Workspaces.”  IEEE Transactions on Automation Science and Engineering.  Conditionally accepted.

Standards and Codes:

The work supports ISO TS 184/SC 2/WG 3 Industrial Safety for ISO TS 15066 (Robots and Robotic Devices – Collaborative Robots) and ANSI/Robotic Industries Association (RIA) 15.06, and the development of standards for industrial robot safety, Automated Ground Vehicle (AGV) safety, and knowledge representations for manufacturing environments.  The measurement science developed in this project will form the basis for future standards that are anticipated to be needed for implementation of robotics in flexible factory floor environments.