The Development of Bow.

Project Overview:

Bow. is an interactive silicone robotic display. The display utilizes a camera to detect a user’s hand (wave) and, when a hand is present, reacts to the user. The display’s reaction is a mix between a wave and a bow.

Project Objectives:

The objective of our project was to create a finger-like pneumatically actuated silicone finger. Our immediate goal was to create a part which, when inflated via an air pump, bent in a way which mirrored a human finger. The loftier goal, which we were unable to complete during the duration of the course, was to develop the part so that it could attach itself to a user’s hand similarly to a brace or exoskeleton. In developing this part, we aimed to further our understanding of cavity geometry, specifically how it affected the inflation and actuation of a soft robotic part.

Creative Design Reflection:

The use of silicone in casting our part allowed us to mirror the fluidity and softness of a real finger’s movement. Additionally, with regard to the concept of attaching it to the user’s hand, we liked the fact that, although the part’s actuation would guide the user’s movement, the softness of the material would allow for the user to pushback/reject movements with far less strain than would be required to pushback against a hard exoskeleton.

In the end, we really enjoyed the fact that, although we did not attach the part to users’ hands, the softness of the material enabled far more user interaction than a harder material would have. At multiple times during the show, we witnessed users poking, prodding, and physically interacting with the parts. Being able to poke them — both while inflated and not — and watch their gelatinous recoil ended up being a fan favorite of those who came by our piece; while this was not intended, it added to the display greatly by allowing physical interaction, rather than only being able to engage through a computer screen.

Lastly, the use of soft technology added a shock-factor and intrigue to our display. When the piece inflates, two air pockets very obviously grow. Because of the softness of the material we used, these pockets seemingly came out of nowhere (as opposed to say, a balloon which is visually limp and deflated). Before engaging with our display, users weren’t quite able to tell what the movement of the part was going to be due to the parts’ opaqueness/translucency, so when the inflation and air pockets came seemingly out of nowhere, they were further enthralled.

Outcomes:

Successes

1. We were able to trial many different cavity designs and settle on both a shape and size (relative to the length and width of the part) which best achieved the inflation angle we desired
2. We designed, fabricated, and programmed an Arduino circuit which used computer vision input to control air pumps
3. We created an interactive display showcasing our research and creative display design.

Failures

1. We were unable to iterate on our design so that the user would be able to wear our part to undergo guided movement. We prototyped multiple different designs for this concept but did not get them to a showcase-able state in time for Rubbery Things.
2. We aimed to have a more complicated script which took into account specific user gestures to control the piece in a more complex way. Due to fabrication timing as well as the limitations of OpenCV via a serial bus, we were unable to update the motor state as rapidly as this goal would require to prevent popping the piece.
3. Although we were able to work around the issue by using an Arduino, we were not able to get a Raspberry Pi Pico to successfully trigger the motors. We’re still not sure why this failed, but were luckily able to switch gears and re-adapt our code to utilize an Arduino shortly before the show.

Technical Documentation

CAD Files:

All of the files linked below were created using SolidWorks, however, many of the designs were originated/conceived using Rhino software.

We designed our part by first 3D modeling the part we wished to obtain by pouring silicone into a mold. From there, we split the part vertically, and used each half to create cavities in a solid block. Lastly we needed to perform a complex split of the remaining part(s) of the block to create our mold top and bottom. The CAD files for our first iteration can be found here.

We continued iterating on our part by adhering a solid piece of silicone to the bottom in order to increase the thickness of the bottom of the part to achieve a more ideal bend. We do not have CAD parts for this piece, as it was poured from a stock mold found in the fabrication lab. The piece was trimmed to fit the surface area of the original, bonded part; the piece thickness was ~6mm.

Due to 3D printing delays, we were unable to iterate on the cavity shape as frequently as we had liked. The makeup for lost time, we developed a part which would allow us to cast and test six of our most promising cavity designs with 1 3D print. The files for this design can be found here.

Lastly, we combined the observations and results of testing multiple designs with minimal 3D prints into our final design. Luckily, this worked nearly perfectly, was easy to both pour and pull, and exceeded our expectations for Rubbery Things. The files for the final iteration can be found here.

Code Files:

Our original code can be found here. This code contains both Python files intended to be run on a computer, as well as MicroPython files intended to be run on a Raspberry Pi Pico. Some test files work as intended, however, the final display shown at Rubbery Things is not able to be achieved with this code.

Our final code can be found here. This is the code that was used during the final show. The folder contains Arduino code intended for an Arduino Uno and Python code for a computer. In order to successfully run this code, some elements (including port names, pin numbers, etc.) will need to be updated by the user.

Citations:

“Creating a Hand Tracking Module Using Python, Opencv, and MediaPipe.” Section, https://www.section.io/engineering-education/creating-a-hand-tracking-module/.

M. N. Golchin, A. Hadi and B. Tarvirdizadeh. Development of A New Soft Robotic Module Using Compressed Air and Shape Memory Alloys. In 2021 9th RSI International Conference on Robotics and Mechatronics (ICRoM), Tehran, Iran, Islamic Republic of. 517-522. https://doi.org/10.1109/ICRoM54204.2021.9663519

Partner Contributions

Maddie:

1. Created final CAD files for finger parts (SolidWorks)
2. Designed and fabricated circuit
3. Wrote Raspberry Pi Pico, Arduino, and Python scripts
4. Maintained and authored documentation
5. Created conceptual sketches

Xiaofan:

1. Spearheaded initial design research
2. Designed and built finalized display box
3. Designed hinge components
4. Created final project video
5. Wrote weekly updates for the group

Both:

1. Conceived final project display concept
2. Pour and bonded silicone parts
3. Continuously researched and updated cavity design

It seems that the design of the hinge solves the problem of power distribution to some extent. The next plan is to optimize the silicone part so that it has less bulge at the top. A better solution, for now, would be to thicken the top and bottom and then reduce the thickness of the ends. When air is injected, the entire silicone model will lengthen rather than bulge.

We rebounded the old silicone as a test model, and were finally able to get a seal without leaks. The test was a success:

The new ABS model has been pointed out by Skylab. Next, we will test the new model.

This week we attempt to use two different designs to solve the problems that arose last week. The first design is a hard ABS with a hinge structure. As you can see in the screenshot, the gray part is the ABS rigid tissue and the blue part is the silicone and the air hose. There are two ring-like objects linked to the ABS parts, which are meant to be fixed to the finger.

Use: When we inject air into the silicone part, the silicone part expands and squeezes the surrounding walls to force the ABS part to rotate. This is a good solution to the problem of force distribution that occurred in the previous model. We used the hinge to concentrate the force in one direction.

The second design is an improvement of the old design.
You can see that we have used two triangles instead of one to provide a more consistent force output. We think the problem that can occur with one triangle is that it causes “ballooning”. Using two triangles allows them to squeeze each other to extend in the opposite direction. This would improve the second problem that appeared last week.

Now that we have the models 3D printed, the next step is to test them.

Unfortunately, this first version of the model did not work well. You can see that the silicone bends at an angle of only 10-15 degrees. This is far below our expectations. Another problem is the “balloon”.

Also this week we tried to cast the wire in silicone and tried to connect to a power source to watch the energy transfer. Unfortunately, we could feel a lot of temperature while the wire inside of the silicone barely straightened.

(Above is a two-part model. The blue numbers “1” and “2” in the picture can form the upper part of our model. The red number “3” is a simple 5-6mm flat surface to help us improve the bottom strength.)

We first tried to fix the bending angle problem. We think there are two reasons for not being able to reach the angle. First, the structure and location of the air region. Second, the force of the air is not concentrated in one area.

We tried to adjust the structure and position of the air area to make the wall more inclined. We also added some small channels to try to see if the silicone would help increase the angle after it was inflated.

For the second problem, we agreed that the main reason was that the bottom silicone was not thick enough. In the first version it only had a bottom of about 2.5mm. In the second version we increased the bottom to 5mm to see if it made a big difference.

After casting the new three-piece design in silicone, we realized that there were some measurement errors that we believe are the result of transferring the STL between CAD programs. Nevertheless, we bonded the pieces together as best as possible, and hope to gain something from this iteration.

We will be re-modeling this design, as well as fleshing out a more triangular design we discussed with Garth last week (rough sketches below):

Another interesting idea is to add some fibers to the bottom layer to further increase the strength. We are looking for materials that still do. We have not implemented this concept in any way yet, so stay tuned!

This piece is called “Umbrella”. It has the hinge structure from the real umbrella to perform the folding action. The artist uses feathers to replace the nylon of the real umbrellas. Feathers give a soft and light feeling. Just as this piece can be an earring or hairpin, it has lightweight.

https://www.behance.net/gallery/90849839/Umbrella?tracking_source=project_owner_other_projects

Another Generosity, designed 2018 by Eero Lundén, Ron Aasholm, and Carmen Lee of Lundén Architecture Company, in collaboration with Bergent, BuroHappold Engineering, and Aalto University (Courtesy of the designers)

This is one of the pieces from an exhibition I found while visiting the Philadelphia Museum in 2018. The name of the exhibition is “design for future”. The work is called “Another generosity”. It uses a lot of flexible plastic and rubber so the piece is like having multiple smaller balloons in one big balloon. It provides gas and electricity through several hoses to ensure the volume and lighting of the “balloon”. Almost the only metal parts used in the whole work are the valves. It has sensors so that the work can detect the carbon dioxide content of the air through movement and light color.

Another version of this work is that it provides a small space for people to walk into the “balloon”. The authors designed a structure in the balloon to ensure that after inflation there is a free space for a person to walk. When people walk inside the “balloon”, a soft ball with lights and a frosted surface will surround them. It allows the viewer to feel like they are in another space.

Our environment is changing dramatically. Wether our architecture is adapting to these changes? Through this work, the authors wanted to find the relationship between architecture and nature and how to make the two coexist.

Trung Thien Hoang, Luke Sy, Mattia Bussu, Mai Thanh Thai, Harrison Low, Phuoc Thien Phan, James Davies, Chi Cong Nguyen, Nigel H. Lovell, and Thanh Nho Do. A Wearable Soft Fabric Sleeve for Upper Limb Augmentation. Sensors, 21(22):7638, January 2021. https://www.mdpi.com/1424-8220/21/22/7638

1. Do you have any conflict of interest in reviewing this paper? A “conflict of interest” is defined as follows:

NO

2. Expertise. Provide your expertise in the topic area of this paper.

1 – No Knowledge

3. Summary. Please summarize what you believe are the paper’s main contributions to the field of soft robotics.

Because soft robots can interact with humans, they are used in many wearable devices. The main problem at the moment is the poor experience caused by the ductility of the material and the slow response time of the system. To solve the material problem, they used rubber tubes instead of non-extendable fibers. At the same time, the authors have developed a liquid metal-based piezoresistive sensor to track the non-linear system response problem. This technology could be used in the future for therapeutic, defense and industrial applications.

4. Strengths and Weaknesses. What are the main strengths and weaknesses of this work? Does the paper have strengths in originality and novelty?

The strength of the article is the detailed description of the human tissue corresponding to each mechanical part, for example by varying the fluid pressure to increase the torque around the elbow joint to achieve arm flexion in a specific direction. The authors also take full account of the mechanics, using fabric around the chest to enhance structural stability. They also take into account the use of Velcro straps to avoid undesirable muscle deformations caused by excessive mechanical movements. I think the downside of the article is that all the experiments and examples are too narrow in terms of the scenarios in which they can be applied. One can only use the material in the article to achieve certain results. I think the article is originality and novelty. the authors have solved the problem of the weight of the wearable exoskeleton device to a certain extent, while the device has enough strength to assist human movement.

5. Soundness. Are the ideas, algorithms, results or studies technologically/methodologically sound?

I think the ideas and algorithms in this article are sound. In order to ensure that the fluid can flow smoothly in the pipe and provide sufficient pressure, the authors have provided extensive algorithmic simulations and material-related knowledge to ensure that the mechanism works smoothly even in a non-linear situation.

6. Does the paper adequately describe related and prior work?
Please write a sentence or short paragraph.

I think this article has been well prepared. The design team has built a successful exoskeleton and experimented with it to obtain the extensive data and algorithms in the article.

7.  Is the paper well organized, well written and clearly presented?
Please write a sentence or short paragraph.

This article is well organised. It begins with a brief introduction to what the whole article will contain and presents the problems faced by the current wearable device. The middle section explains how the design team solved the problem through several experiments and used the device to lift heavy objects or to help humans perform actions. In the conclusion, the authors describe the possible applications of the technology and its contribution to humanity.

8. Do you have suggestions for improving this paper?
Please write several paragraphs detailing specific points of the paper which merit reconsideration. Be sure to address the text, figures and tables, mathematics, and grammar and spelling.

Firstly, I think the article could have been enhanced above the figure. For example, figure1 only briefly describes the components of this wearable device. If the reader wants more clarity and precise detail, the author should add pictures to explain the textual content. Although the author uses a lot of text to justify his research, multiple photos about the details could greatly help those who want to understand this field.

Second, regarding the composition of the article I think it could be improved. In the introductory section, the authors point out the importance of materials in wearable soft robots. However, in the second part, “Materials and Methods”, the authors do not go into detail about the data and properties of materials. Almost all of the content is about how to properly assemble and make the device “move”. I think an introduction to the use of materials would have helped the reader to understand the text more easily.

9. Comments to Committee (Hidden from authors). Does the paper have enough originality and importance to merit publication? Is the paper relevant to the field? These comments will NOT be sent to the authors:

I believe the paper have enough originality and importance to merit publication. This article is about the study of wearable robots for the upper limbs. The article analyses the areas of application of wearable robots and the current problems faced. In the content section, the authors use ample experiments and data to demonstrate the feasibility of new materials and liquid pressure devices. The research component of the article addresses to some extent the power and weight balance of wearable soft robots. I think this brings new ideas to the field of wearable robotics. Therefore this article is worth publishing.

10. Overall Rating. Provide your overall rating of the paper (5 is best)

5 – Definite accept: I would argue strongly for accepting this paper.

Source:

Automatic Design and Manufacture of Soft Robots, DOI: 10.1109/TRO.2011.2172702

Root:

Automatic Design of Soft Dielectric Elastomer Actuators With Optimal Spatial Electric Fields. DOI: 10.1109/TRO.2019.2920108

Related works:

Bioinspired Dynamic Affect-Based Motion Control of a Humanoid Robot to Collaborate with Human in Manufacturing. DOI: 10.1109/HSI47298.2019.8942609

Topology Optimization of Skeleton-Reinforced Soft Pneumatic Actuators for Desired Motions. DOI: 10.1109/TMECH.2021.3071394

Customizable Three-Dimensional-Printed Origami Soft Robotic Joint With Effective Behavior Shaping for Safe Interactions. DOI: 10.1109/TRO.2018.2871440

Pedro P.; Ananda C.; Rafael P.B.; Carlos A.R.; Alexandre B.C. Closed structure soft robotic gripper. IEEE International Conference on Soft Robotics, July 2018. doi:10.1177/0278364914543793.

2. Bubble casting soft robotics

Trevor J. Jones, Etienne Jambon-Puillet, Joel Marthelot, and P.-T. Brun. Bubble casting soft robotics. Nature, 599(7884):229–233, November 2021. doi:10.1038/s41586-021-04029-6.

3. Bioinspired locomotion and grasping in water: the soft eight-arm OCTOPUS robot

M. Cianchetti, M. Calisti, L. Margheri, M. Kuba, and C. Laschi. Bioinspired locomotion and grasping in water: the soft eight-arm OCTOPUS robot. Bioinspiration & Biomimetics, 10(3):035003, May 2015. doi:10.1088/1748-3190/10/3/035003.