IEEE Robotics & Automation Magazine - June 2018 - 118

Conclusions
An SCR was designed, developed, and validated to collaborate with humans in automotive assembly. It can hold a heavy
payload for humans and enable people to use their motions
and finger forces to transport and manipulate heavy parts in
all 6 DoF. By taking advantage of the robot, human workers
can be assisted in effectively, flexibly, and conveniently handling heavy parts in automotive assembly, which has great
potential benefit for increasing automotive assembly production efficiency and quality as well as improving ergonomics.
The SCR is a representative example of how we can leverage
both human and robot capabilities in manufacturing, where
the human deals with dexterous assembly tasks while the
robot handles heavy payloads of automotive parts. The
application of such a robot system is clearly not limited to
automotive assembly. Any manufacturing task that involves
heavy payload transportation and manipulation tasks could
be benefited by this type of robot. In addition, the robot
could have a wide range of potential applications in other
areas, such as helping the elderly transport and manipulate
heavy objects at home [23], [24] and even assisting soldiers
in carrying heavy burdens on battlefields, like a robotic
mule, by leveraging its advantages of compact size, heavy
payload, high flexibility, and intuitive user interfaces.
Significant scope exists for improving the performance of
the SCR, which we are pursuing in future work.
● First, the kinematic control and actuator limits sometimes
generate bumpy motions. We are examining the use of
dynamic modeling and controls of the SCR to achieve
smoother, human-like motions.
● Second, added sensing (in the form of stereo cameras, Lidars,
radars, inertial measurement unit, and indoor global positioning systems) and postprocessing (simultaneous localization
and mapping) can enhance the robot's situational awareness.
● Third, situational awareness unlocks the potential for
greater autonomy (in the form of autonomous navigation,
obstacle avoidance, and part loading) that can further
improve the efficiency of human-robot collaboration.
● Last, but not least, advanced learning and adaptation algorithms can be developed to make the SCR more intelligent
and able to recognize human needs and then actively assist
humans to accomplish assembly tasks instead of being passively person-guided.

(a)

(b)

(c)

Acknowledgments
We wish to thank the BMW technicians for sharing their user
cases and requirements in realistic automotive assembly and
supporting the experiments in this research work. This work
was supported in part by a subaward (ARM-17-QS-F-02)
from the ARM Institute (W911NF-17-3-0004) and the
National Science Foundation Grant IIS-1710898.

(d)
Figure 14. The part assembly process. Part A's edges are denoted
by red lines and part B's by yellow lines. (a)-(b) The SCR
prepares for assembling the two parts by lifting part A up to
part B. (c) The robot implements some fine maneuvers, so part
A's bolts come to align with part B's holes. (d) Then the SCR
connects the two parts.

118

IEEE ROBOTICS & AUTOMATION MAGAZINE

june 2018

References
[1] M. Akamatsu, P. Green, K. Bengler, "Automotive technology and
human factors research: Past, present, and future," Int. J. Veh. Technol.,
vol. 2013, 2013. doi: 10.1155/2013/526180.

IEEE Robotics & Automation Magazine - June 2018

Table of Contents for the Digital Edition of IEEE Robotics & Automation Magazine - June 2018