Data from: 3D printed digital pneumatic logic for the control of soft robotic actuators
Data files
Jan 19, 2024 version files 12.57 MB
Abstract
Soft robots are paving their way to catch up with the application range of metal-based machines and to occupy fields which are challenging for traditional machines. Pneumatic actuators play an important role in this development, allowing the construction of bioinspired motion systems. Pneumatic logic gates provide a powerful alternative for controlling pressure-activated soft robots, which are often controlled by metallic valves and electric circuits. Many existing approaches for fully compliant pneumatic control logic suffer from high manual effort and low pressure tolerance. In our work, we invented 3D printable, pneumatic logic gates that perform Boolean operations and imitate electric circuits. Within 7 hours, an FDM printer is able to produce a module that serves as either an OR, AND or NOT gate; the logic function is defined by the assigned input signals. The gate contains two alternately acting pneumatic valves, whose work principle is based on the interaction of pressurized chambers and a 3D printed 1 mm tube inside. The gate design does not require any kind of support material for its hollow parts, which makes the modules ready to use directly after printing. Depending on the chosen material, the modules can operate on a pressure supply between 80 and over 750 kPa. The capabilities of the invented gates were verified by implementing an electronics-free drink dispenser based on a pneumatic ring oscillator and a 1-bit memory. Their high compliance is demonstrated by driving a car over a fully flexible, 3D printed robotic walker controlled by an integrated circuit.
README: 3D printed digital pneumatic logic for the control of soft robotic actuators
https://doi.org/10.5061/dryad.jq2bvq8gv
This dataset contains the data and scripts that were used to generate the figures of the main text and supplements. All data was generated with pneumatic elements that were fabricated on a 3D printing platform as detailed in Conrad et al. (2020) (https://doi.org/10.1007/978-3-030-64313-3_6). The flexible parts, like valves and logic gates, described in this work have been printed in one continuous process with 'Recreus FilaFlex TPU A60' (shore hardness A 63), 'Recreus FilaFlex TPU A70' and 'Recreus FilaFlex TPU A82' with fitting print parameters (detailed in the supplements). This includes pneumatic connectors, channels and expandable membranes. Most parameters were set in the software 'PrusaSlic3r 2.4', which was used to prepare our models for printing. In a post processing step the files generated by the software were additionally modified to adapt the flow rate to the special needs of the flexible filament.
Description of the data and file structure
In the dataset the files are sorted by the figure or table they belong to. Inside the respective folders there are the raw data as well as the RStudio scripts that were used to calculate statistical tests as well as generate the figures.
Explanation for data files
Figure2File1.csv:
Column 1 consists of the material of the used PLG in Shore hardness. Column 2 is the ID of the PLG. Column 3 is the valve of the PLG, where V1 is the normally-open valve and V2 is the normally-closed valve. Column 4 is the pressure at the socket S_t that was required to close the respective valve in bar.
Figure2File2.txt:
Column 1 shows the set pressure in bar at the input socket S_c1 of the PLG. Column 2 shows the measured mean pressure in bar at the input socket S_c1 for that pressure step. Column 3 is the mean pressure in bar at the socket S_t. Column 4 is the mean volumetric flow in ln/min through the PLG. Columns 5,6 and 7 are the respective standard deviations of the measured values for that pressure step. Column 8 is used to differentiate between the two experiments of measuring flow ('Durchfluss') and closing the valve ('Schliessen'). Column 9 specifies the tested valve and column 10 is the concatenated value of column 8 and 9.
Figure2File3.txt:
The columns are the same than in Figure2File2, but instead of specifying the valve, column 9 shows the direction of the signal (ASC = ascending/rising = 0 -> 1, DESC = descending/falling = 1 -> 0).
Figure3File1-125.txt:
Column 1 shows the set pressure in bar at the input socket S_t of the PLG. Column 2 shows the measured mean pressure in bar at the socket S_t. Column 3 shows the measured mean pressure in bar at the socket S_c. Column 4 is the mean volumetric flow in ln/min through the PLG. Columns 5,6 and 7 are the respective standard deviations of the measured values for that pressure step. Column 8 shows the direction of the signal (ASC = ascending/rising = 0 -> 1, DESC = descending/falling = 1 -> 0). Column 9 shows the set system pressure in bar at the socket S_p+. Column 10 specifies the valve of the PLG, where V1 is the normally-open valve and V2 is the normally-closed valve. Column 11 is the concatenated value of columns 8, 9 and 10.
Figure3File1-250.txt:
The columns are the same than in Figure3File1-125, but the system pressure is at 250 kPa instead of the 125 kPa.
Figure3File2-125.txt:
The columns are the same than in Figure3File1-125, but the values are from the NC valve instead of the NO valve.
Figure3File2-250.txt:
The columns are the same than in Figure3File1-250, but the values are from the NC valve instead of the NO valve.
Figure3File3.txt:
This file shows the maximum values for each combination of PLG, system pressure and direction of the signal. Column 1 shows the set pressure in bar at the input socket S_t of the PLG. Column 2 shows the measured mean pressure in bar at the socket S_t. Column 3 shows the measured mean pressure in bar at the socket S_p+. Column 4 is the mean volumetric flow in ln/min through the PLG. Column 5, 6 and 7 are the respective standard deviations of the measured values for that pressure step. Column 8 shows the direction of the signal (ASC = ascending/rising = 0 -> 1, DESC = descending/falling = 1 -> 0). Column 9 shows the set system pressure in bar at the socket S_p+. Column 10 is the concatenated values of column 8 and 9. Column 11 is the ID of the PLG.
Figure4File1.txt:
This file shows the measurement for the NOT-PLG. Column one shows the time in seconds. Column 2 shows the measured pressure at socket S_p+ in bar. Column 3 shows the measured pressure at socket S_c1 in bar. Column 4 shows the measured pressure at socket S_t in bar. Column 5 shows the measured pressure at socket S_c2 in bar. Column 6 shows the measure pressure at socket S_out in bar.
Figure5File1.txt:
This file shows the measurement for the AND-PLG. Column one shows the time in seconds. Column 2 shows the measured pressure at socket S_p+ in bar. Column 3 shows the measured pressure at socket S_t in bar. Column 4 shows the measured pressure at socket S_c2 in bar. Column 5 shows the measured pressure at socket S_out in bar.
Figure5File2.txt:
This file shows the measurement for the OR-PLG. Column one shows the time in seconds. Column 2 shows the measured pressure at socket S_p+ in bar. Column 3 shows the measured pressure at socket S_c1 in bar. Column 4 shows the measured pressure at socket S_t in bar. Column 5 shows the measured pressure at socket S_c2 in bar. Column 6 shows the measure pressure at socket S_out in bar.
Figure6File1.txt:
Column 1 shows the set system pressure in bar. Column 2 shows the measured mean time in seconds needed for an ascending input signal at socket S_t to create a status change in the output socket S_out. Column 3 shows the measured mean time in seconds needed for a descending input signal at socket S_t to create a status change in the output socket S_out. Columns 4 and 5 show the respective standard deviations of the measure values. Column 6 specifies the number of PLG connected in a row.
Figure6File2.xlsx:
Columns 1 to 6 show the same data than in Figure6File1.txt. They were used to calculate the normal equations from the observation equations and to fit a trend to the normal equations. The trend for the ascending slope is calculated in columns 8 to 16, the trend for the descending slope is calculated in columns 18 to 26. This was then used as the trend seen in Figure 6 B and C. Additionally, we subtracted the delay for a theoretical number of gates of 0, in order to remove the delay introduced by the measuring instruments. We included a reference in the script "Figure6.R" to show this.
FigureS1File1.csv:
Column 1 shows the set system pressure in bar. Column specifies the tested valve . Column 3 shows the direction of the signal (ASC = ascending/rising = 0 -> 1, DESC = descending/falling = 1 -> 0). Column 4 is the measured mean transition interval for the switching of the valve in bar.
FigureS1File2.csv:
This file has the same data than FigureS1File1.csv, but the two directions of the signal are pooled into one dataset. It is also the dataset used for figure S1. Please see the script "Figure S1.R" for the statistics.
FigureS4File1.txt:
Column 1 shows the maximum compression force in kN. Column 2 shows the ID of the PLG. Column 3 specifies which channel is beeing measured (1 = S_c1, 2 = S_c2). Column 4 shows the measured mean volumetric flow through that channel in ln/min. Column 5 shows the standard deviation of the measured mean volumetric flow.
FigureS5File1.csv:
Column 1 shows the ID of the PLG. Column 2 shows the maximum compression force in kN. Column 3 shows the direction of the signal (ASC = ascending/rising = 0 -> 1, DESC = descending/falling = 1 -> 0). Column 4 shows the repetition. Column 5 is the beginning of the edge of the signal in seconds. Column 6 is the end of the signal in seconds. Column 7 is the difference between the end and the start of the signal in seconds. Column shows the difference between the end and the start of the signal in seconds after subtracting the delay the measuring system introduces (calculated in Figure6File2.xlsx).
TableS1File1.xlsx:
Column shows the Shore hardness on the A scale of the PLG. Column 2 shows the ID of the PLG. Column 3 shows the maximum air pressure it resisted in bar.
Code/Software
All scripts were made and run in RStudio. The version used is RStudio 2023.03.1+446 "Cherry Blossom" Release (6e31ffc3ef2a1f81d377eeccab71ddc11cfbd29e, 2023-05-09) for windows.
Methods
The described pneumatic elements were fabricated on a 3D printing platform as detailed in Conrad et al. (https://doi.org/10.1007/978-3-030-64313-3_6). The flexible parts, like valves and logic gates, described in this work have been printed in one continuous process with “Recreus FilaFlex TPU A60” (shore hardness A 63), “Recreus FilaFlex TPU A70” and “Recreus FilaFlex TPU A82” with fitting print parameters (please see supplements).
In all pneumatic experiments, the pressurization of chambers and channels was controlled by a custom test stand. This setup contained six pneumatic proportional valves driven by a LabView program. The software supports the creation of linear patterns and repeating sequences, which simplified the systematic characterization by applying rising and falling pressure ramps to the logic elements. For the characterization of a single valve, unused inlets were sealed and the gate output socket was connected to atmospheric pressure via a flow sensor to measure the volumetric stream. After changing the state on the input connectors, the software waited four seconds for the module to reach a new static internal flow and then recorded this value. To test the behavior of the complete logic gate, a pressure sensor was connected to the output. Here, the applied pressure was measured continuously to characterize the dynamic behavior. For deformation experiments we used a mechanical testing machine equipped for compression tests up to 10 kN.
The data was processed using RStudio 2023.03.1+446 "Cherry Blossom" Release (6e31ffc3ef2a1f81d377eeccab71ddc11cfbd29e, 2023-05-09) for windows. The same software was used to create figures.