Tug-of-War-Style High-Force Fluidic Actuation for Small Diameter Steerable Instruments

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This paper presents a novel approach to amplifying the force output of fluidic actuators for use in miniature, steerable medical devices, which introduces a "Tug-of-War" style actuation method that amplifies force by connecting pneumatic artificial muscles (PAMs) in parallel and distributing them serially along the instrument’s axis. This paper is well-structured and well-written. Before publication, there are some questions to be solved.

1. The author mentioned “it is important to develop new steerable MIS devices that can be navigated within the patient to access and treat anatomically remote surgical sites.”, more state-of-the-art human-robot interaction can be cited: DOI: 10.34133/cbsystems.0062; DOI: 10.34133/cbsystems.0007; DOI: 10.34133/cbsystems.0110.

2. How were the individual lengths and tension of each collaborative PAM calibrated during assembly, and how do you ensure uniform force distribution among the actuators?

3. The results show a significant hysteresis effect in both single and collaborative PAMs at higher frequencies. Could the authors elaborate on the causes of this hysteresis and whether material selection or other modifications might reduce this effect?

4. The single PAM used as a control was 250 mm long, while the collaborative PAMs were 50 mm each. How was the decision to use these lengths made, and would altering the lengths of individual PAMs in the collaborative design affect performance?

5. Did the authors explore alternatives to the steel transmission cables? Could more flexible or lightweight materials affect the overall actuation performance and interaction with the PAMs?

Author Response

Thank you for your feedback and questions, please find our responses below marked in red.

This paper presents a novel approach to amplifying the force output of fluidic actuators for use in miniature, steerable medical devices, which introduces a "Tug-of-War" style actuation method that amplifies force by connecting pneumatic artificial muscles (PAMs) in parallel and distributing them serially along the instrument’s axis. This paper is well-structured and well-written. Before publication, there are some questions to be solved.

1. The author mentioned “it is important to develop new steerable MIS devices that can be navigated within the patient to access and treat anatomically remote surgical sites.”, more state-of-the-art human-robot interaction can be cited: DOI: 10.34133/cbsystems.0062; DOI: 10.34133/cbsystems.0007; DOI: 10.34133/cbsystems.0110.
   1. Thank you for your feedback. After reviewing the suggested literature, we agree that they are relevant, and have added these references.
2. How were the individual lengths and tension of each collaborative PAM calibrated during assembly, and how do you ensure uniform force distribution among the actuators?
   1. Thank you for your question. We have addressed this in the revised text by clarifying and expanding the description of the assembly process in the methodology section, quoted below:
   2. “The PAMs are joined progressively, with the cable at the completed end of the chain held in a vice and the in-progress end brought to a neutral tension while the cables of successive PAMs are fixed to the connectors of their neighbors, maintaining a straight alignment of the flexible components while minimizing any variable pre-tensioning of the individual PAM segments. To ensure balanced actuation between the PAMs, dimensions of the braid, tubing, end connectors, and lengths of cable between PAMs were carefully controlled throughout assembly. The final actuatable lengths of the assembled collaborative PAMs were measured to be between 49.3 − 50.2 mm, within 2% of the targeted length. The single PAM was measured to be 249.5 mm in length”
   1. The potential effects of any variations in pre-tension or PAM lengths are also addressed in the discussion section, quoted below:
   2. “Any mismatch among the PAMs in length or initial tension would cause a difference in deadband effect between a PAM and its collaborators, effectively making it slack and non-productive until it has sufficiently contracted to tension its connection to the actuation cable.” 
3. The results show a significant hysteresis effect in both single and collaborative PAMs at higher frequencies. Could the authors elaborate on the causes of this hysteresis and whether material selection or other modifications might reduce this effect?
   
   1. Thank you for your question. Hysteresis effects in the output of pneumatic artificial muscles are common even for PAMs that are not collaborative. It is commonly caused by friction between the bladder and braid strands and the elastic deformation of the components. Note that while complex, different methods have been developed to deal with this hysteresis behavior, such as in the study from our group:  "Hysteresis Modeling of Robotic Catheters Based on Long Short-Term Memory Network for Improved Environment Reconstruction", doi: 10.1109/LRA.2021.3061069.  Additionally, the level of hysteresis could be reduced by optimization of the braid and bladder material properties, improvements to the manufacturing process to simplify uniform assembly and component pre-tension for each actuator segment, the use of a more rigid material for the transmission cables to transmit forces, or the use of a less compressible fluid to drive actuation. These are important areas for further investigation, and significant research has been done focused on reducing hysteresis for large-scale production of PAMs. We believe the current materials and assembly method, which were consistent between the introduced actuator and control actuator, sufficiently demonstrate the capabilities of the proposed design relative to the state-of-the-art. This hysteresis effect is addressed in the section below from the initial submission text: 
   2. “Both prototypes displayed a significant increase in hysteresis and reduction in force output as frequency increased, with the collaborative prototype reaching 18.3 N or only 70.1% of the maximal output force and the single PAM reaching 6 N or only 88.2% of the maximal output force at 1.5 Hz. This difference in maximum achieved force may be a function of limitations of the pneumatic throughput of the selected valve. The larger drop in performance of the collaborative PAMs merits further investigation, but may be influenced by the rate of pressure propagation through the muscles or deformation of the additional lengths of cable needed to connect their ends.”

1. The single PAM used as a control was 250 mm long, while the collaborative PAMs were 50 mm each. How was the decision to use these lengths made, and would altering the lengths of individual PAMs in the collaborative design affect performance?
   1. Thank you for your question. We have reworded the following section in the text to clarify the selected actuator dimensions and effects of altering lengths:
   2. “To evaluate performance of the proposed design, two prototype actuators were constructed: one using a group of collaboratively connected PAMs, and a second with a single PAM of equivalent length to the sum of the collaborative PAMs. A total actuator length of 250 mm was selected to fit within the 300 mm shaft length of currently available fetoscopic devices used for laser ablation procedures and a previously developed flexible fetoscope [18]. For the collaborative group, the 250 mm actuator length was split evenly between 5 PAMs of equal length. This number was selected to ensure the actuator would be capable of providing useful forces at stroke lengths up to 5 mm, 10% of the resting length of each PAM, to actuate the device bending segment through its maximum range of motion. Subdividing the actuator length further (building 6 or more collaborative PAMs and keeping overall actuatable length constant) would produce higher forces at short stroke lengths, but the force output as a function of contraction length will decrease more quickly as the resting length of the PAMs is reduced.”
2. Did the authors explore alternatives to the steel transmission cables? Could more flexible or lightweight materials affect the overall actuation performance and interaction with the PAMs?
1. The steel cables were selected to provide a relatively stiff coupling between components, while still being able to bend around the curvature of the flexed device bending segment. Flexible transmission cables would stretch too much under tension, introducing additional hysteresis effects to the system and allowing the PAM segments to contract further, reducing the amount of force they are able to apply to the load. We believe that for the targeted application, flexible or lighter transmission cables would not provide an advantage.

Reviewer 2 Report

Comments and Suggestions for Authors

Please find my comments in the attached file.

Comments for author File: Comments.pdf

Author Response

Thank you for your feedback and questions, please find our responses below marked in red.

 

Review of the ID actuators-3227903 manuscript entitled ‘Tug-Of-War Style High Force Fluidic Actuation for Small-Diameter Steerable Instruments” by Robert Lathrop, Mouloud Ourak, Jan Deprest, and Emmanuel Vander Poorten to be published in MDPI ‘Actuators. 

The reviewed manuscript presents a miniature yet powerful fluidic actuation system for bending of a steerable microinstrument device for minimally invasive surgical procedures. The required end effector force and bending degree boosting was achieved using parallel interconnection of 5 collaborative miniature pneumatic artificial muscles (PAM) which were located serially along and interlocked by short and stiff load and anchor cables. 

The presented solution is very interesting, the developed prototype intrinsically actuated drive system consisting of 5 PAM segments with a maximum diameter of approximately 2.6mm was made and tested both in the basic configuration as a linear actuator (obtaining approximately 10mm stroke length) as well as in the target configuration as a prototype laser coagulation fetoscope bending actuator, achieving almost 70° bending of its active tip. In addition, this miniature actuator provides a high bending force that exceeds 26 N. The paper is well written, and the ideas and results are clearly presented and discussed. The conclusions in the conclusion of the paper are also correct. 

However, in my opinion, the following remarks should be addressed before the manuscript can be published. 

1. I wonder whether the presented actuator was tested in the target wet and viscid environment. In my opinion it is mandatory to test the dynamic properties (i.e. the frequency response) of the multi-segment actuator in a ‘working’ fluid simulating e.g. amniotic fluid. For sure, the actuator frequency response would differ significantly in a ‘sticky’ environment in relation to the measured results presented obtained in air (which are not of great value in a real operational environment).
2. Thank you for your questions. For this study, the maximum frequency response of the actuators was assessed to explore the possible design space that this actuation scheme enables for use in future devices, up to 1.5 Hz. This showcases the actuator’s potential (which would also be relevant for design of devices for laparoscopic or endoscopic tools functioning in a dry field). For the targeted clinical application of this study, rapid device movements are not desirable and could potentially be dangerous. Performing a laser ablation procedure on the placenta requires a surgeon to carefully maneuver to different vessels on the placental surface, orient their device perpendicular to the tissue surface as best they can, and hold the device tip in a fixed position while activating the laser to coagulating tissue uniformly. We do not believe that high frequency testing in a viscous environment would be clinically relevant to the target application for this study. We also would not expect amniotic fluid to produce significant resistance to the motion of a flexible device driven by over 20 Newtons of force.  We believe that the current evaluation demonstrates the value of the proposed design and its potential for use in the design of devices suited for clinical validation testing in a fluid environment.
3. Has a complete MIS device consisting of the discussed multi-segment PAM actuator, endoscopic camera, and the laser coagulation fiberoptic line been constructed? Has it been tested? Did it work as expected?
4. Thank you for your question. The intention of this work is to evaluate the potential of the proposed actuation system and characterize its performance, thereby enabling optimization of a fully functional MIS device design. The additional force produced by the introduced actuator opens up a wide array of new possibilities for the design, construction, and material selection of MIS bending segments and structural components. The construction of a complete MIS device leveraging the performance of the introduced actuator is planned for future work.
5. Figures presenting the test data: The points visualising real test values should be made visible in the figures, especially in Figures 7 and 8. Moreover, at least basic information on the measurement uncertainty should be provided, especially in the case of those figures (7 and 8). Were the test results repeatable? To what degree? Please provide some basic statistical data or quantities (at least the standard deviation of the maximum extension and maximum deflection). The presented measurement data seem to be recorded just for one actuator run…
6. Thank you for your feedback. To address this point, we have additionally performed a series of five repeated tests at each pressure level for both actuators, including the results in Table 1 and adjusting Figure 7 to show average and standard deviation for the measurements. The diameter measurements are also moved to Table 1, with Figure 8 removed as we believe it would be redundant to the table.  We have also evaluated the actuator’s performance after repeated actuation cycles, which resulted in slightly reduced but comparable bending performance (3.6 % reduction after 100 actuations). This is addressed in the revised text quoted below:
7. “Tests were repeated five times by removing and re-applying the target pressure, with average achieved bend angle +/− standard deviation reported in Table 1 and displayed in Fig.7. The collaborative PAMs produced a significantly higher average bend angle at 65.9 degrees compared to the single PAM’s 10.2 degrees, while occupying a similar footprint. To assess stability over repeated actuation, the bend angle produced at maximum pressurization was re-measured after applying 100 cycles of a sinusoidal pressure input with 6 bar amplitude at 0.5 Hz frequency, with results reported in Table 1. A small decrease in bend angle is observed after repeated cycling, likely due to relaxation of the actuators and transmission cables. This could be addressed in future designs by applying a controlled stretching regimen to the actuators before they are installed, to induce any potential relaxation before the actuators are pre-tensioned."
8. At least some basic test results should be made available regarding the actuation length and bending angle change in time or along the number of actuation cycles. Do these values change with the number of actions?
9. This point is addressed in the response to point 3 above.
10. A closer analysis of the data presented in Figure 5 shows that the actuator system does not produce the initial 0 N force when unactuated. What was the reason for that? Does it change with time or number of actuations?
11. Thank you for your question. The data presented in Figure 5 indeed shows a non-zero initial force due to the pretension of the actuators during testing. Pre-tension is minimized for this study, as it could lead to undesirable deformation at rest for a cable-driven MIS device. During force and stroke testing, each actuator was suspended between two stages (setup shown in Figure 4), with pre-tension controlled by adjusting the stage position until the actuator was under the minimal tension at which no slack remained in the transmission cable. We did not observe changes to this baseline pretension over time or following multiple actuations. We have added the following section to address this directly in the text:
12. “While fixing each actuator in place for testing, its pre-tension was adjusted to the minimum amount required to remove slack from the driving cables. Pre-tension was minimized for this study, as significant pre-tensioning can deflect the neutral tip position of flexible MIS devices. This resulted in a baseline tension of 0.48 N for the collaborative actuator and 0.14 N for the single PAM, with the collaborative actuator requiring a slightly higher tension to counteract the additional weight of its transmission cables.”

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

This paper is ready for publication.