Case Study: Industrial Robot-Human Interaction
https://www.youtube.com/watch?v=TQ0DeWo5Jsg
Overview
General Motors is collaborating with Dr. Croft and her team in the CARIS Lab to improve interaction between industrial robots and employees on GM’s door assembly line. The project is supported by NSERC CRD & NSERC Discovery grants.
Human-Robot interaction is often a limiting factor for performance on the assembly line. It reduces the potential for multitasking and high level interactions between the robot and the employee. To speed up interactions and make the assembly line more flexible, more advanced systems are needed.
Funding
This project is funded by the following:
- NSERC CRD Grant
- NSERC Discovery Grant
- GM Canada ltd.
Solution
The project is supported by a consortium of institutions and companies, including UBC, McGill University, Université Laval and GM Canada Ltd.
The research focuses on the potential for smart machines to improve the ability of workers to handle an increased range of products and operations while improving:
- Interaction Speed
- Production Line Flexibility
- Product Quality
- Worker Safety
Three research streams and thrusts are followed to solve this problem. The three research streams are:
- Define the Human Interaction with the robotic assistant and the task
- Develop the robotic assistant’s relevant Situational Awareness and Data Representation needs
- Coordinate developments through an Integration Framework
The three research thrusts that develop the robot assistant technology comprise:
- Advanced human-robot communication and cooperation schemes
- Safe interaction design and control to support HRI
- Vision and Non-Contact Sensing
At a time of fierce global competition, this research could give Canada a significant edge in advanced manufacturing and further develop a highly skilled workforce as the marketplace undergoes rapid change.
Results
The project is in the early stages. Based on the industrial success of the CARIS Handover Interaction Package and the expertise of the leading collaborating labs and partners, the project is expected to bring solutions soon.
The Labs
The CHARM project brings together five separate research groups from three Canadian universities, in partnership with a group centered at GM Research.
Click on any of the following titles to expand its content and learn more about the collaborating labs.
CARIS
The CARIS research group is experienced in human-robot interaction communication and control design. This group has particular expertise in HRI scenarios with industrial robot manipulators. This group is responsible for proposing, testing, and implementing novel HRI interaction methods specifically suitable for vehicle manufacturing environments.
Sensory Perception and Interaction Laboratory
The SPIN research group studies innovative user interfaces, focusing on physical and multimodal interface hardware, interaction design, and the human perception that underlies the use of these interfaces. The SPIN group’s key contributions to the project will be a human-robot task definition, developed in collaboration with GM, and the design of human-robot communication methods.
Laboratoire de Vision et Systems Numeriques
This laboratory research group is experienced in 3D vision and environment modeling, and cognitive vision. This team collaborates closely with the McGill research group on the development of the vision system. This research group also leads the development of an Integration Framework for CHARM system components. This framework supports efficient technical developments and smooth technology transfer to GM.
Laboratoire de Robotique de l'Universite Laval
Laval's robotics group has expertise in safe dynamic control design for robotic systems. This group has previously developed a robotic intelligent assist device in partnership with GM. Their work on this project is to expand on the interaction control strategies for the robotic platform and to support integration of work from the other research groups to achieve a robotic assistant able to meet the challenge case study.
Artificial Perception Laboratory
The Artificial Perception Laboratory (APL) at the McGill Center for Intelligent Machines contributes expertise in sensors, 3D reconstruction, and the interpretation of sensor data. This research group is developing a system to estimate the pose of one or more human workers in the workspace for safe human-robot interaction. By pose, we refer to the worker’s position and orientation in the workspace as well as the configuration of the various parts of the body (e.g. head, torso, limbs). This research team is also developing a fast and reliable “Situational Awareness Database” (SADB) for the RA system - a common database for modeling the workspace shared by one or more humans and a robotic assistant. The motivation for a common database is to facilitate integration of the various modules by avoiding time-consuming and error-prone conversions between representations.
Aiming for a quick technology transfer of the research results, a diverse group from GM has been engaged. From GM’s research and development department: researchers, managers and directors provide valuable input by guiding the CHARM project. In addition, GM’s manufacturing engineering department has been involved as the custumer of all innovations created by CHARM. The overall group consists of GM personnel from various industrial backgrounds including dynamics and control, computer vision, quality, safety, ergonomics, kinematics, management and system integration. Members of this group work with academic research groups to provide industrial support, vision, and information on GM’s existing assembly processes. GM’s guidance ensures a smooth technology transfer of the results.