8. Soft robotics¶
I couldn’t be more excited about this week’s topic! The idea of making our environments — materials, objects, and spaces — more interactive, responsive, and alive feels deeply inspiring to me. It’s like giving the world around us a gentle breath of life, allowing it to communicate and move with us.
Soft robotics challenges the traditional idea of machines as rigid and mechanical. Instead, it explores how soft, flexible, and inflatable materials can mimic natural motion — the way an octopus arm curls or how a flower slowly opens and closes. This approach opens up new possibilities not only in engineering and design, but also in art, wellness, and human experience.
Since this is my first encounter with soft robotics, I approached the week with a lot of curiosity and playfulness. My goal was to explore how different shapes, structures, and wall thicknesses influence the way silicone moves when air becomes the driving force. Through these hands-on experiments, I’m learning how form, material, and motion can work together — almost like a small choreography between matter and breath.
To approach this week thoughtfully, I also looked into Fabricademy alumni’s projects for guidance and inspiration. Their work helped me understand different ways to design molds, control motion, and combine air with soft materials.
Marisa Satsia’s Documentation of Her Previous Work in Soft Robotics
A few of my favorite references are listed below:
Research¶
This week, I wanted to understand how to manipulate silicone so that it can move in the ways I imagine. To do that, I began studying bending behaviors and different soft actuation mechanisms — exploring how changes in structure, layering, and pressure can bring motion to life.
Elve Flow's Bending Mechanism of a Soft Elastomer Actuator, referenced in Grecia Bello
Elve Flow's Types of Soft Actuation Mechanisms, referenced in Grecia Bello
References & Inspiration¶
I was inspired by Bruna Goveia’s work on embroidered inflatables, which beautifully combines softness and structure. This made me curious to experiment with how different stitch patterns or geometries could influence the way inflatables expand and move.
Bruna Goveia's Embroidered Inflatables Research
Experimentation #1 — Exploring How Different 2D Lattices (Grid Patterns) Affect Expansion¶
What is a 2D Lattice¶
A 2D lattice is a repeating grid-like structure made up of intersecting lines or cells.
In soft robotics, the geometry of these grids influences how silicone expands when inflated.
Dense or symmetrical grids limit movement, while open or asymmetric ones allow more flexible deformation. For this experiment, I explored three types of lattices: grid, honeycomb, and isogrid.
Different 2D Lattice Configurations by Seokpum Kim, Aslan Nasirov, Deepak Kumar Pokkalla, and Vidya Kishore
Process and Workflow¶
1. Modeling — I designed the lattice pattern and mold in Rhino. To reduce printing time, I only printed the top part of the mold and planned to cast a flat silicone sheet in a bakery mold for the bottom layer.
3×3 Collection of Mold Inserts Comparing Concentric Rings with Varying Thickness, Positive Grids, and Inverted Grids for Ecoflex 00-30 Expansion Studies
2. 3D Printing — Printed with a Prusa i3 MK3. The mold took around five hours to complete. Optimizing slicer settings such as infill or layer height could help reduce printing time in future trials.
3D Printing on Prusa i3 MK3
💬 Notes: - Refine the Mold for Smooth Surfaces and Minimal Gaps
This mold took about 3.5 hours to print. I switched from a 0.2 mm Speed profile to a 0.3 mm Draft profile to reduce the print time from 4 hours.
The downside is that the draft quality created tiny gaps in the walls, allowing silicone to seep in and making demolding a bit messy. A smoother, tighter print would have prevented this. - Optimize Print Settings Further or Use a Faster, Well-Calibrated Machine I printed this mold on an older Prusa i3 MK3, which increased the print time and introduced some surface roughness. The MK3 is reliable, but it prints much slower than newer high-speed machines, so fine details can take longer and require more careful tuning.
3. Preparing the Mold — Sprayed the mold with Easy Release to prevent the silicone from sticking.
Mold Preperation
💬 Notes: - Please take the time to prepare your workspace properly. Make sure your surface is flat and place a protective layer under your mold before casting. I rushed and poured directly into the mold without thinking about overflow—silicone spilled onto my workspace and took much longer to clean. The uneven mold surface also led to small leaks and gaps in the final cast.
4. Casting Silicone — Mixed and poured Ecoflex 00-30, then left it to cure for four hours. You can watch 1 minute instruction video on how to cast with Ecoflex™ 00-30 here.
💬 Notes: - Make sure to mix the silicone slowly and evenly. Trapped bubbles can create weak spots in thin areas, which may turn into tiny air leaks when you inflate the silicone.
5. Demolding — Removed the cured silicone piece carefully from the mold.
Cured Silicone Sample Casts
6. Adding a Fabric Base — Pouring a large silicone sheet evenly into a bakery mold was difficult, so I opted to use cotton fabric as the bottom layer instead. This gave me more control over the thickness and helped save silicone. To create the base, I applied a thin coat of uncured silicone to both sides of the cotton sheet, forming a flexible and reinforced bottom layer.
Uneven Poured Silicone Sheet (Left) and the Making of a Cotton Fabric Base (Right)
💬 Notes: - Make sure to pour the silicone evenly if you plan to use it as a base sheet, or ensure that your fabric lies completely flat. Any crease in the fabric can create weak points that turn into air leaks when you inflate the piece.
Results¶
These tests clearly show that the printed grid pattern has a strong influence on how the silicone expands. The following trends were observed:
-
Dense or tightly packed grids restrict expansion, resulting in smaller, stiffer inflation shapes.
-
Patterns with fewer cells or larger openings allow greater expansion, producing bigger, rounder forms.
-
Different geometries (squares, hexagons, circles, triangles) generate distinct deformation signatures visible in both the top and side views:
-
In terms of thickness, each radial element was printed at 0.5 mm (shown in the second image of the top row). This thickness is relatively thin and did not produce noticeable differences in expansion. With a larger range of thickness values, I expect the impact of thickness on inflation behavior to become more apparent.
Overall, these results demonstrate that pattern density and geometric topology are the dominant factors shaping how the silicone expands and the final inflation profile it produces.
Top View Comparison of Expansion Behavior Across Different Grid Patterns — Pattaraporn (Porpla) Kittisapkajon
Side View Comparison of Expansion Behavior Across Different Grid Patterns — Pattaraporn (Porpla) Kittisapkajon
Observations and Errors¶
Error #1: — Air Leak Caused by Trapped Bubbles
Testing Expansion of a 0.5 mm Ecoflex Sample (Air Leak Caused by Trapped Bubbles)
You can fix this by mixing the silicone slowly and thoroughly from the start, or by applying an additional thin layer of silicone afterward to seal any gaps.
Error #2: — Silicone Surface in the Expansion Area Bonded to the Fabric Base
Silicone Expansion Surface Bonding to Fabric Base
Make sure you apply the silicone adhesive carefully and intentionally. In my case, the silicone touched areas it wasn’t supposed to, causing the expanding surface to stick to the fabric base. A better approach would have been to apply the adhesive only on the selected areas—for example, adding a controlled layer on the silicone side—so the fabric bonds only where needed.
However, this mistake led to an unexpected, dynamic, and surprisingly beautiful result. It opened up a new direction of exploration for me: designing non-uniform expansion to create more complex motions and functional behaviors.
Error #3: — Air Leak Caused by Fabric Base Crease
Air Leak Caused by Fabric Crease
Experimentation #2 — 3D Printing an Existing Gripper to Understand Soft Actuation¶
For this experiment, I printed an existing pneumatic gripper file shared by Anastasia to understand how air channels and wall thickness influence bending.
During printing, I manually adjusted the process and accidentally lost the cap piece—later realizing it was crucial for sealing the chamber and maintaining air pressure.
Mold for Soft actuator/gripper - Parametric - OpenSCAD - JB86
Results¶
Even without the cap, the structure partially inflated and demonstrated how air channels guide motion.
This experiment highlighted the importance of airtight sealing and consistent wall thickness for smooth, repeatable actuation.
Partial Inflation Test Demonstrating Airflow and Motion
Experimentation #3 — Creating Self-Supportive Shell Structures¶
What is a Shell Structure¶
A shell structure is a curved form—such as an arch, dome, or saddle—that distributes stress and pressure through its geometry.
Design¶
My goal was to create arch, saddle, and dome structures that mimic the soft actuation of a gripper. Instead of individual finger-like parts, the curvature itself becomes the responsive structure.
Concept — Self-Supporting Geometry Informed by Air-Channel Orientation
Results¶
Self-Supporting Surface — Pseudopneumatic Actuation Test
The results did not perform as expected. I believe the air channels were too small, and some of them may have filled with silicone during attachment, preventing proper airflow. In several samples, the expansion became uneven or overly dramatic because the top and bottom layers did not bond or cure consistently. I also overlooked how much Ecoflex expands, which means the channels likely needed more spacing and design adjustments to achieve the intended effect.
Experimentation #4 — Smocking Pattern with Soft Robotics¶
Continued to be inspired by traditional smocking techniques, I wanted to explore how similar folding or pleating logic could be recreated through air motion.
By embedding air channels in silicone sheets arranged in smocking-like grids, I tested how inflation could generate surface textures that move dynamically, almost like fabric.
Design¶
Using the same principles as the soft robotic gripper, I designed a series of air channels arranged in a smocking-inspired pattern. The goal was to guide inflation in specific directions and recreate the folding behavior seen in textile smocking through controlled pneumatic expansion.
Smocking-Inspired Inflation Mold — Pattaraporn (Porpla) Kittisapkajon
Results¶
Smocking Pattern with Soft Robotics Test — Pattaraporn (Porpla) Kittisapkajon
The results were not as expected—there was no visible inflation. I believe the air channels may have been too narrow for airflow, and in some areas the silicone may have filled or blocked the channels entirely, preventing proper expansion.
Key Takeaways¶
This week taught me how essential it is to co-design with the material—its properties directly shape the outcome. I also realized the importance of working mindfully and attentively; many of my mistakes could have been avoided with more care during design, makรng, and preparation. Finally, I gained a deeper appreciation for the value of craft and practice. Even simple tasks, like gluing silicone, require skill and experience to achieve consistent, high-quality results.