SoPrA:Fabrication & Dynamical Modeling of a Scalable Soft Continuum Robotic Arm with Proprioception

"SoPrA: Fabrication & Dynamical Modeling of a Scalable Soft Continuum Robotic Arm with Integrated Proprioceptive Sensing"
accepted for presentation at 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2021)
Yasunori Toshimitsu, Ki Wan Wong, Thomas Buchner, Robert Katzschmann
arxiv.org/abs/2103.10726
This work presents a design and modeling process for a Soft continuum Proprioceptive Arm (SoPrA) actuated by pneumatics. The integrated design is suitable for an analytical model due to its internal capacitive flex sensor for proprioceptive measurements and its fiber-reinforced fluidic elastomer actuators. The proposed analytical dynamical model accounts for the inertial effects of the actuator's mass and the material properties, and predicts in real-time the soft robot's behavior. Our estimation method integrates the analytical model with proprioceptive sensors to calculate external forces, all without relying on an external motion capture system. SoPrA is validated in a series of experiments demonstrating the model's and sensor's accuracy in estimation. SoPrA will enable soft arm manipulation including force sensing while operating in obstructed environments that disallows exteroceptive measurements.

[@kiwanwong8600]

Nice work;

[@keldantv2540]

This reminds me of an elephants trunk, is that what it is based on?

[@robertkatzschmann1390]

yes, this arm is inspired by an elephant trunk

[@user-lp7fm6sk5e]

Nice，but I noticed that flex can only return a curve of two axis.How to detect the 3-dimension with only one flex sensor?
Does it have some math work about the 3-d curve calculation?

[@softrobotics]

Thanks for the comment! Our model uses a piecewise constant curvature assumption, where each segment is assumed to have constant curvature throughout, and therefore it can be described with two angles. It assumes that there is no lengthwise stretching nor twisting, so the two values coming from the sensor are all we need to figure out the pose of the robot.
- Yasunori

[@BHARGAV_GAJJAR]

Adding gripper looking thing at the end is vital otherwise....

[@user-lt5no1xt1z]

What is the proprioceptive sensor? Is it FBG fibers?

[@user-lt5no1xt1z]

great work, by the way!!

[@softrobotics]

Thank you for your comment! We’ve used the 2-axes flex sensors from Bend Labs. www.bendlabs.com/
We did discuss using FBG sensors, but for this particular project we decided to go with a much simpler and lower cost (about $100 each) solution.
- Yasunori

[@user-lt5no1xt1z]

@@softrobotics Very neat! Looking forward to your future work

[@user-lt5no1xt1z]

@@softrobotics Also, a follow up, How did y'all map the simulated result coordinates to the 2-d image of the ground truth ?

[@softrobotics]

@@user-lt5no1xt1z Thanks for the comment, the overlay effect from 0:45 - 0:58 was achieved with the Qualisys Track Manager (QTM) software that we use for motion capture. The camera view is from Qualisys' RGB camera, whose pose is known in the motion capture space since it is calibrated. We import the simulated trajectory into QTM using the motion capture coordinates, and we export the video with the overlay enabled.
- Yasunori