Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.
Communications Engineering volume 3, Article number: 117 (2024 ) Cite this article ce certification butterfly valve
To achieve coordinated functions, fluidic soft robots typically rely on multiple input lines for the independent inflation and deflation of each actuator. Fluidic actuators are controlled by rigid electronic pneumatic valves, restricting the mobility and compliance of the soft robot. Recent developments in soft valve designs have shown the potential to achieve a more integrated robotic system, but are limited by high energy consumption and slow response time. In this work, we present an electropermanent magnet (EPM) valve for electronic control of pneumatic soft actuators that is activated through microsecond electronic pulses. The valve incorporates a thin channel made from thermoplastic films. The proposed valve (3 × 3 × 0.8 cm, 2.9 g) can block pressure up to 146 kPa and negative pressures up to –100 kPa with a response time of less than 1 s. Using the EPM valves, we demonstrate the ability to switch between multiple operation sequences in real time through the control of a six-DoF robot capable of grasping and hopping with a single pressure input. Our proposed onboard control strategy simplifies the operation of multi-pressure systems, enabling the development of dynamically programmable soft fluid-driven robots that are versatile in responding to different tasks.
The emerging field of soft robotics has been demonstrated to be suitable for operations in complex and unstructured environments1,2. Taking inspiration from the ways that biological systems like octopuses coordinate multiple appendages and exploit the compliance of their soft tissues, soft robots take advantage of distributed actuation and the mechanical properties of soft materials to achieve levels of dexterity that are challenging to attain with rigid-linked robots3,4,5,6. Progress in manufacturing techniques has enabled multiple distributed actuators to be embedded in soft robots7. However, efforts to enable the coordination of actuators in a multi-degree of freedom (DoFs) soft robot that mimic the complex coordination seen in nature are stymied by reliance on control hardware that can hinder the dexterity of the robot itself8,9,10. This is particularly relevant to fluid-driven soft robots wherein a pressurized fluid is used to induce non-linear deformations in the form of extension, contraction, bending, or twisting11,12. To achieve specific, pre-programmed sequence of actuation, existing fluidic soft robots typically require multiple pressure input lines in the form of tethered, pneumatic tubings controlled by electromechanical valves and external pressure sources for independent control of each DoF, adding bulk, limiting mobility and adaptivity, and negatively impacting the dynamic performance of the system13,14,15.
Enabling on-board control of soft robots is critical to afford its autonomy for deployment in real-world applications9. Building upon valve designs developed for microfluidics and progress in manufacturing soft materials, valves have been designed for use in soft robotics applications for low-level control functions16,17. In these cases, one pressurized input is used to control flow through one or more other pressurized inputs to achieve a combination of pressurized outputs. While these valves function as electronic-like devices, as exemplified by the development of soft ring oscillators and digital logic gates, they must be dynamically reprogrammed and is primarily used to sense and respond to environmental stimuli following Boolean logic18,19,20,21,22,23,24. Recent work has demonstrated the use of a fluidic demultiplexer to decrease the input-to-output ratio, but is limited to sequential activation of individual output lines25,26. The aforementioned devices are suitable for applications where electricity may be unreliable, such as remote locations or environments with high radioactivity24,27. However, they sacrifice the complexity afforded by electronic control in an effort to make strictly mechanical systems that operate with some degree of autonomy. Furthermore, some fluidic logic components are also limited by their relatively bulky and hard infrastructure, precluding their integration into applications requiring low-profile design for navigation in confined spaces. This led to recent interest in developing a class of soft fluidic devices and actuators manufactured from sheet-based materials such as textiles28,29,30,31,32 and thermoplastic films. In33, this layered manufacturing approach is used to fabricate multi-degrees-of-freedom stacked balloon actuators (SBAs). The minimal thickness in the inflated states of these sheet-based actuators makes them useful for operations in small, delicate environments33,34,35,36,37,38.
In contrast, electronic control offers fast reprogramming capabilities and implementation of more advanced control algorithms. A number of electronically controlled valves for fluidically actuated soft robots have been introduced in recent years. For example, Xu et al. presented a valve that uses a dielectric elastomer to control an indenter that blocks a pneumatic channel39,40 with a maximum blocked pressure of 20.6 kPa. This approach involves the use of high operational voltage, with the dielectric elastomer requiring 1500 V. Wang et al. introduced a valve that uses magnetorheological elastomer (MRE) controlled via an electromagnet to block flow up to 41.3 kPa41. However, the valve’s reliance on continuous electrical current to adjust magnetic fields in MRE valves leads to increased system temperature, potentially challenging control and performance and necessitating extra thermal management strategies. Electropermanent magnets (EPMs) offer an energy-efficient approach to control mechanisms in valve designs for actuation of multi-DoF soft robotic systems42. EPMs operate by combining two permanent magnets of differing material properties with an electromagnet and steel end caps to create an assembly that switches between magnetized (ON) and demagnetized (OFF) states by applying a small, instantaneous—typically in magnitudes of microseconds—current pulses43. As EPMs can hold their state indefinitely (i.e., maintain a constant magnetic field over long durations) between pulses of current, this reduces the overall energy consumption of the system and can enable actuation without continuous control. One valve design uses the magnetization of the EPMs to manipulate the movement of a metal ball inside a check valve to block airflow through a tube at pressures up to 25 kPa44. We have previously introduced a method that uses EPMs for induced stiffening of magnetorheological (MR) fluid—a smart fluid with iron particles in a carrier fluid like water that stiffens in the presence of an applied magnetic field—to actuate soft actuators42,45,46. However, this method is limited in actuation speed and ability to generate sufficiently large magnetic fields, resulting in changes of pressure up to 10 kPa. Additionally, this method is not compatible with working fluids which are unaffected by the changing magnetic field of the EPMs (e.g., compressed air and water). While other works in literature explore electroprogrammable stiffness of smart fluids for flexible valve designs, they suffer from similar challenges in speed, high voltage requirements, and reliance on specialty materials47,48,49,50,51,52. Table 1 shows a comparison of electronic control hardware for pneumatic soft robots including the valve presented in this work.
Size also plays a crucial role in valve design, especially concerning its impact on soft robot performance. Even if a valve is rigid, its significance diminishes if its size is small relative to the characteristic size of a soft robot. As shown in Table 1, current valves are larger than 1 cm and only the commercially available solenoid valves have maximum pressures exceeding 50 kPa. Being able to scale the valve to ≤ 1 cm in all dimensions is an important step towards integrated, on-board flow control with minimal effect on the overall mobility and dexterity afforded by the soft structure. This is particularly relevant for mesoscale pneumatic soft robots with typical actuators such as fiber-reinforced actuators measuring ≈1.3 cm in diameter and 16.5 cm in length53 and PneuNets measuring ≈2 cm in width and 10 cm in length54. There currently exists no electronically controlled valve that has been integrated into fluidically actuated soft robots manufactured from sheet-based materials.
This work introduces an electronically controlled EPM valve to achieve independent control of a multi-DoF soft robot powered by a single pressure source (Fig. 1). Our proposed valve architecture is designed to overcome the limitations of current state of the art soft valves and commercialized solenoid valves for fluid-driven soft robotic systems through the following features: (i) a low-profile architecture integrated with channels fabricated from commercially available sheet materials (i.e., thermoplastic films) following a stacked layer lamination process, (ii) electronic control enabling reprogrammability without altering the robot’s mechanical construction (e.g., rewiring of pneumatic tubings or reconstruction of the valve components), (iii) no direct contact with the robot working fluid (iv) modular and compact design to facilitate the scalability of the system for multi-DoF systems, (v) low energy consumption of 47 mJ with 20 V electrical pulse for 500 μs required to switch the magnetization of the EPMs, and (vi) ability to distribute pressure from a single pressure source into different fluidic chambers of the robot. Such input can be in the form of a constant or a sinusoidal pressure input.
A Picture of the EPM valve. B Qualitative demonstration of the flexibility of the integrated EPM valves on a thermoplastic structure. C The EPM valve uses a pinching effect to obstruct fluid flow: the magnetic force between the magnetized EPM end cap and the keeper plate results in the collapse of the TPU channel. C-i A side view of the valve in its closed state with a graph of the corresponding pressure over the length of the valve, and C-ii A side view of the valve in its open state with a graph of the corresponding pressure over the length of the valve. D Fabrication process of an EPM valve involving (1) thermal bonding of stacked layers of thermoplastic films to create the channel, (2) insertion of tubing to facilitate pneumatic connections, (3) integration of the EPM assembly, and (4) placement of the EPM into the alignment slot.
To understand the blocking capabilities of the EPM valve, we first present a model of its force generation based on the geometry. We next validate the model by performing a parametric analysis and optimization study of the EPM geometry using COMSOL. We then experimentally determine the pressure blocking capabilities of the EPM valve as well as cycle testing to evaluate the behavior of the EPM valves in response to repeated switching. To highlight the reprogrammability and scalability of our proposed valve design, we develop a multi-function soft robot with two three DoF actuators powered by a single pressure input while all of its six DoFs are electronically controlled. Besides enabling solenoid-like flow control (i.e., selectively pressurize the different flow channels), the EPMs can act as switching flow controllers to direct the pressure into the desired actuator chamber. We demonstrate this by using configuring the three EPM valves in parallel to a single pressure source and programming the three DoF actuator to reach all the positional states. We also demonstrate the various modalities the robot can achieve with a single pressure input including (i) pinching grasp, (ii) hopping locomotion, and (iii) dual-curvature bending by combining two of the same three DoF actuators end-to-end.
The valve exploits the force between a magnetized EPM and a steel keeper plate positioned below a flow channel (Fig. 1A, C). This force then deforms the wall of the underlying fluidic channel (Fig. 1C). With the valve closed in this way, the EPM blocks flow so long as the flow pressure does not exceed the pressure exerted on the channel between the keeper plate and the end caps of the EPM (Fig. 1C(i)). This prevents areas downstream of the valve from pressurizing. With the EPM demagnetized, the flow channel is free to expand underneath the EPM, allowing fluid to flow downstream (Fig. 1C(ii)).
The EPM valve design consists of three main components: (i) a soft, fluidic channel, (ii) an EPM, and (iii) supporting structures (Fig. 1D). The fluidic channel is fabricated through heat sealing two sheets of thermoplastic polyurethane (TPU) (38 μm thick), with a Teflon masking layer (25.4 μm thick) between them, to define the internal geometry with the middle rectangular region of the channel measuring 1 mm wide and 8.5 mm long (Fig. 1D–1). Slits matching the dimensions of the EPM end caps are then laser cut into the TPU layer. Flexible tubings are inserted at both ends of the channel (Fig. 1D–2) to facilitate connections with other fluidic components (e.g., actuators or sensors) and to reduce the collapse length of the TPU channel if an applied vacuum is used—such in the case of actuator deflation—where the TPU channel may flatten and restrict the passage of air. The TPU valve is enclosed between supporting structures including an alignment plate fabricated from laser-cut DuraLar with a single slit matching the surface area of the contacting end cap and a keeper plate that are secured by metal fasteners (Fig. 1D–3). The EPM is then aligned to the valve by placing one of the end cap inside the slot of the alignment plate (Fig. 1D–4). The detailed fabrication process of the TPU channel and EPM valve assembly is provided in the Supplementary Materials S0.5. While the EPM is rigid, its small nature still allows to maintain flexibility as shown in Fig. 1B.
The design of the EPM is guided by our analytical models and simulations aimed at achieving a maximum blocking force while limiting the operational voltage of the valve, as elaborated in the subsequent section.
The EPMs used in this study consist of a soft permanent magnet, AlNiCo 5, a hard permanent magnet, Neodymium 42, two endcaps made of Stainless Steel and copper coil wrapped around the two magnets as shown in Fig. 2A. A more thorough overview of the EPM fabrication process is detailed in Supplementary Materials S0.4 and S0.7.6. AlNiCo 5’s magnetic field can be re-oriented in the presence of a strong external magnetic field, as supplied by the current carrying copper coils; while Neodymium retains its original magnetic field due to higher coercivity. This ability to switch AlNiCo’s magnetic field allows us to “magnetize” or “demagnetize” the EPMs based on the amplitude and duration of the current supplied in the copper coil43. The performance of the proposed EPM valve design is characterized by the maximum blocked pressure or the critical pressure threshold at which the valve can withstand without experiencing leakage. For our EPM valve, this corresponds to the force exerted by the EPM on the steel keeper plate that pinches the underlying TPU channel. To predict this force, we used the model derived by Knaian43 which describes the magnetic circuit based on Ampere’s Law and Gauss’s law:
where d is the diameter of each permanent magnet, Nrods is the number of magnets (here equal to two), Balnico is the magnetic flux density of the alnico magnet which itself is a nonlinear function of the magnetic field, Hm(t), and time, t, due to hysteresis. Br is the remanent flux density of the neodymium magnet, μ0 is the permeability of free space, a and b are the width and thickness, respectively, of either end cap, g is the length of the air gap between either end cap and the keeper plate, \({{{{\mathcal{P}}}}}_{leak}\) is the leakage permeance of the magnetic circuit, N is the number of turns in the electromagnet coil, I(t) is the current through the coil, and l is the length of each permanent magnet.
A The dimensions of the EPM viewed isometrically and from a cross-sectional perspective. The coil modeled does not show the full 150 turns of the actual EPMs. B Force vs. air gap for the EPM as determined via the analytical model, via simulation in COMSOL, and experimentally. C Cross-sectional view of the EPM assembly, highlighting the elements contacting to induce the pinching effect of the TPU channel.
The value of Hm(t) was solved numerically in MATLAB and then used to compute the force, F, exerted on the keeper plate using the following equation derived from the Maxwell stress tensor. When the EPM is in its ON state, achieved by supplying an electrical pulse with an appropriate driving voltage and sufficient duration, the magnetic field within the material stabilizes. Thus, we assume static magnetization, simplifying the time-dependent variables to I(t) = 0 and Balnico approximating Br. For our EPM a = 8 mm, b = 1.6 mm, d = 3.175 mm, and l = 6.35 mm (Fig. 2A, B). The force was predicted for values of g in the range of 0.001 mm to 1 mm using MATLAB.
A simulation of the EPM and keeper plate was then conducted in COMSOL to provide an estimate for \({{{{\mathcal{P}}}}}_{leak}\) in Equation (1) and to validate the forces predicted by Equation (2) (see Methods and Materials for details). \({{{{\mathcal{P}}}}}_{leak}\) was estimated to be 65 nH by sweeping over a range of values in MATLAB such that the error between the COMSOL results and MATLAB prediction was minimized.
As indicated in Equation (2), the magnetic force is impacted by the dimensions of the EPM. Hence, we conducted additional COMSOL simulations to investigate the impact of various geometric parameters on the valve performance, particularly those not captured explicitly by the analytical model. These parameters are however collectively captured by the influence of leakage permeance, \({{{{\mathcal{P}}}}}_{leak}\) . Characterizations of the EPMs with different design dimensions are illustrated in Figure S1. A more detailed discussion of the results of the COMSOL simulations to characterize the ideal EPM dimensions is included in Supplementary Materials S0.1. Based on these simulations, it was determined to make the end caps 8 mm high and 1.6 mm thick. A slit pattern matching the dimensions of the end caps was incorporated into the TPU layer to mitigate any potential loss of magnetic force that might arise from the presence of a non-magnetic material between the contacting surfaces of the EPM end caps and the keeper plate. The keeper plate was made to be 1.6 mm thick and with an area of 8 mm by 10 mm to just accommodate the footprint of the EPM.
The number of electromagnet coil turns to magnetize the EPM was determined using an analytical model of its electrical behavior (Equations S1–S4). We selected 150 coils to reduce the overall power consumption of the EPMs by minimizing the resistance of the wire while ensuring a sufficient amount of current required to switch magnetization states is supplied with a short pulse duration. With the selected geometry and wire dimensions, each EPM consumes 47 mJ of instantaneous energy when switching (Equation S5). The activation of the EPM involves changing the orientation of the magnetic domains in the soft magnet. However, achieving uniform and consistent magnetization across the soft magnet and minimizing the potential for incomplete switching due to hysteresis effects of magnetic material—the lag between the applied magnetic field and the resulting magnetization—requires higher activation voltage. Through experimental validation, we have determined that setting the EPM at 20 V for 500 μs consistently yields reliable and reproducible results.
Based on the analytical and simulation results, EPMs were constructed with the determined dimensions and electromagnet coils. An experiment was then conducted to validate the force between the EPM and steel keeper plate as predicted by COMSOL and equations (1) and (2). Figure 2 shows the experimental data plotted with the values predicted analytically and via COMSOL.
We calculate for the maximum theoretical pressure blocked by the EPM to be 682 kPa. A detailed derivation of this result is included in Supplementary Materials S0.3. We note that the theoretical blocking pressure predicted by this model provides initial estimates of the EPM’s blocking capability, disregarding the material properties of the soft valve specifically, the elastic yield strength of the TPU, as well as potential manufacturing inconsistencies. As the transient response of the valve is not captured in our theoretical model, leakage at low pressures may be present due to pressure build up as the contained fluid in the upstream has no pathway to exhaust.
Experiments were conducted to characterize the pressure blocking performance of the EPM valves (Fig. 3). The blocking performance of the valve is characterized by the maximum pressure the valve can withstand without leaking when the EPM is magnetized (Fig. 3A). The inlet of the valve receives compressed air at a constant supply pressure while the outlet is connected to a needle of small diameter (0.31 mm) connected to atmosphere. This needle, characterized by high fluidic resistance, enables the gradual release of accumulated air in the downstream areas at a controlled flow rate. The intentional design of the needle, analogous to a pull-down resistor, ensures that the pressure remains relatively constant when the system is pressurized (Fig. 3C). The needle also allows for a controlled drop to atmospheric pressure when the system is not pressurized. We subjected the valve to continuous operation to evaluate its durability over an extended period of time (Fig. 3B). Specifically, we magnetized the EPM (i.e., switched to the ON state) and conducted tests to detect any potential leakage after a 15 min duration. This is a sufficient time frame to capture the reliability of the valve in blocking flow. A control strategy where the input pressure does not need to be constantly engaged can be used so long as the soft actuators inflate in times on the order of seconds. Such an approach is demonstrated in the following sections. In addition to continuous operation, we subjected the valve to periodic operation involving repeated magnetization and demagnetization of the EPMs for approximately 5 min to assess the performance of the valve over cyclic conditions (Fig. 3B). This process included allowing and blocking a constant air flow supply of ≈40 kPa, maintaining each state for 10 s.
A 3D model of the EPM TPE valve showing a view of the valve in its closed state. B-i Responses of the EPM valve to a constant pressure of ≈150 kPa over 15 minutes. B-ii Switching of the EPM valves when supplied with a constant pressure of ≈ 50 kPa to demonstrate its performance over cyclic conditions for 15 min. C Schematic of the experimental setup used to obtain the flow response of the EPM valve. D Responses of the EPM Valve to incremental pressure step inputs with the EPM valve initially in the closed position D-i and with the EPM valve initially in the open position D-i The shaded error bars of each test show the mean and standard deviation for three trials. E A − 100 kPa step input with (left) the valve open and (right) with the valve closed starting from an initial pressure of 30 kPa. F The transient response of downstream pressure under a ramp-up staircase waveform pressure input. The experimental curves show the average magnetic flux with one standard deviation.
We evaluated the behavior, or the flow response, of the EPM valve when switching between its magnetized and demagnetized states to determine its effectiveness in blocking the upstream and downstream flow, critical in maintaining the pressure state of the connecting actuator (Fig. 3D). When the EPM undergoes a transition from magnetized to demagnetized states (ON to OFF), the valve facilitates the downstream flow of pressurized air, leading to the pressurization of the system (Fig. 3D–i). In the reverse scenario, transitioning from demagnetized to magnetized states (from OFF to ON), the valve inhibits the pressurization of the downstream areas (Fig. 3D–ii), thus maintaining the pressure in a channel even after the input is switched off. This is evident through a discernible drop in pressure, signifying the release of the accumulated air into the atmosphere once the input pressure is set to atmosphere. Experimental results for both cases with input pressures up to ≈146 kPa revealed that the EPM effectively occludes flow in either condition as illustrated in Fig. 3D. Furthermore, the response time to reach the input pressure was ≈0.147 s. Operating above this pressure deforms the thin channel which, in turn, may impact the dynamic behavior of the EPM valve.
We then conducted further experiments to characterize the negative pressure blocking performance of the EPM valve. For each trial, the pressure regulator provided a 30 kPa step input to the inlet of the valve splitter until the pressure equalized across the valve. A -100 kPa of vacuum was then supplied at the inlet. The test ran until the outlet pressure sensor measured a value of 0.5 kPa. Three trials were completed with the EPM in its magnetized and demagnetized states. Figure 3E shows the results of this test. We also demonstrated the ability of the EPM valve to maintain the pressure state of the downstream flow areas by supplying the EPM valve with an incrementally increasing constant pressure input from 0 to 40 kPa (Fig. 3F). The EPM is switched to its off state (i.e., demagnetized) to equalize the upstream and downstream pressure after ≈10 s of the increase in pressure. This capability can be expanded to the selective inflation of multi-DoF pneumatic actuators by timing when the DoF connected to a corresponding channel will be pressurized.
Two failure modes were observed when operating the EPM valve. The first pertains to the potential leakage in the downstream areas when operating beyond the maximum blocking pressure due to insufficient magnetic attraction between the end caps and keeper plate. When the pressure is large enough, it deforms the channel to let air break through which results in the lifting of the end caps, allowing the pressurized air to escape downstream. To address this limitation, scaling the EPMs in accordance with the analytical model and COMSOL simulations, as discussed above, can enhance their blocking capability for operating pressures beyond ≈146 kPa. The second failure mode involves the bursting of the TPU film at the inlet of the EPM valve. This phenomenon is attributed to the specific design choice of inserting tubings into the TPU flow channels, allowing for a tolerance in the fit that results in the formation of small air gaps. These gaps, in turn, facilitate the accumulation of pressurized air in the upstream areas when the EPM is in a magnetized state. The upstream pressurization leads to the gradual expansion of the TPU over time, eventually causing bursting of the inlet. To improve the system’s robustness, the valve edges are reinforced with nylon taffeta with single-side TPU coating. This effectively restricts the TPU film from exceeding its elastic limit, thereby enhancing the valve’s overall reliability and performance (see Fig. 1A).
We demonstrate the capability of the EPM valves as an electronic control system to facilitate the switching between different task modalities by regulating airflow for various actuation sequences. Notably, our focus is on showcasing the functionality of the system rather than optimizing specific tasks. To demonstrate the versatility of our EPMs in creating varied actuation sequences, we incorporate the valve design into a soft robot featuring appendages made of SBAs33. These SBAs, fabricated from TPU following a layered fabrication approach, exploit their collapsibility and dexterity to achieve multi-axis bending motions suitable for simple tasks like grasping and locomotion functionalities, all within a relatively compact form factor. Using multiple EPM valves in a soft robot allows the user to set pre-programmed actuation sequence by selecting the combination of valve states (i.e., choosing when the EPMs will be switched ON or OFF and the duration the state will be hold) to achieve oscillatory pressure inputs. Alternatively, a static input pressure can be set such that opening the valves inflate the actuators while closing the valves lock the actuators to a specified configuration by preventing pressurized air from further entering. Due to the inelastic expansion, SBAs must be actively vented using negative pressure33. As such, a vacuum can be supplied to deflate the actuators. Moreover, the thermoplastic elastomeric fluidic channels play a crucial role in distributing the pressurized air from the inlet tube to the rest of the interconnected tubings and TPU flow channels corresponding to each EPM. This approach reduces the complexity in designing an actuation sequence by simplifying the control scheme to a series of states and enables the operator to reprogram, or change the behavior of the robot by altering the system’s response to a controller input, without mechanically reconfiguring the robot (e.g., rewiring the interconnected tubings or reassembling the individual components).
We highlight the role of the EPM valve in facilitating timed actuation of SBAs. The demonstration can be viewed in Supplemental Movie 1 (S0.7.1). Each chamber’s inputs are connected to the three-DoF EPM valve, allowing for the distribution of input pressure (Fig. 4A). We present two examples of fluidic circuits to demonstrate how EPMs are utilized to achieve sequential actuation. In the first circuit (Fig. 4A), we supplied the system with a pressure supply that alternates between a constant 30 kPa and -100 kPa. Here, the actuator does not return to its rest state between commands, but instead simultaneously held pressure in one column while deflating its neighbor such that the SBA tip described a circle (Fig. 4B). This sequence was completed in only 14 seconds. In the second circuit (Fig. 4C), we use the regulator to set a constant pressure and the EPM valves then switch between a regulated, positive pressure of 50 kPa to inflate the actuators, and atmospheric pressure to exhausts the system, achieved by connecting each pressure supply to a single-DoF EPM valve. Figure 4D shows the SBA moving between each combination of column inflation, returning to the rest state between each (see Supplemental Movie 1). The EPMs are activated to either inflate or deflate the corresponding chamber at a preset time following the oscillatory input. We utilize a script that defines a preprogrammed actuation sequence, controlling the magnetization and demagnetization of the corresponding EPMs, thereby opening or closing the respective channels. The script also includes pauses between each command. During inflation phases, the regulator provided 30 kPa to the inlet for 1 s and 50 kPa to the inlet for 2 s in the second case. This is determined by the set pressure of the regulated air flow, the fluidic resistance of the valves and connecting tubings (thereby determining the flow rate), and the actuation volume of the SBAs. During deflation phases, the system exhausts to atmosphere for 5 s in the first case while the regulator provided— 100 kPa to the inlet for 5 s when deflating one chamber or 7 s when deflating two chambers in the second case. Full deflation was therefore the rate limiting aspect for this process, with multi-chamber combinations requiring more time to deflate due to the greater volume of air to exhaust.
A A schematic representation of the EPM valve architecture used to control a three-DoF SBA. B A time sequence of the SBA's movement in which the tip rotated through 360° without returning to the rest state between each command. C A schematic representation illustrating the EPMs as a switching mechanism to either pressurize or exhaust the actuator. D A time sequence of the SBA's movement in which each combination of chamber inflation states was reached.
The capability of soft robots to perform complex tasks requires the control and coordination of multiple actuators. Thus, it is critical that each actuator be individually addressed as each requires varying pressure for a specific time step in an actuation sequence to achieve periodic motions. We designed a soft robot with an integrated six-DoF EPM valve to electronically control two three-chamber SBAs, positioned side-by-side, acting as the limbs. Each DoF is connected to a channel controlled by an EPM. A pressure regulator was connected to the six-DoF splitter with the EPM valve to actuate each chamber as shown in Fig. 5A. This single regulated pressure switches between 30 kPa and − 100 kPa. This enables us to program different behaviors and motions in response to varying fluidic input conditions. We highlight the reprogrammability of the robot, or its ability to change behavior, by exploring various operational modes as well as the ability to reduce the number of external pressure inputs required to achieve a variety of actuation sequences.
A Electronically controlled pneumatic soft robot showing SBAs, TPE fluidic channels, and EPM valves mounted on a flexible acrylic plate with a single tube for positive and negative constant pressure supply. A side view of the multi-function six-DoF robot during its gait depicts the use of the SBAs for legged locomotion. B A schematic representation of the EPM valve architecture used to control two three-DoF SBAs. C The multi-function robot used as a grasper to pick up (from left to right) a pencil, a cup, a dried apricot, and a rubber duck. The top and bottom rows show the grasper in its OFF and ON states, respectively. D A time sequence showing the multi-function six-DoF robot locomoting across a table. E A time sequence of the two segment SBA arm showing a selection of dual-curvature positions.
First, the two SBAs were connected end-to-end to form a soft tentacle capable of reaching dual curvature states. Figure 5E shows the dual segment tentacle moving between selected positions. The EPM valves were used to select which actuators would change inflation state, and the regulator provided either 30 kPa or -100 kPa as before. This sequence can be viewed in Supplemental Movie 2. Next, the two SBAs positioned side-by-side to create a multi-function soft robot capable of grasping or locomotion. Figure 5C shows the robot mounted via an acrylic plate to a commercial robot arm (UR6, Universal Robotics) where it functioned as a grasping end effector. The EPMs were used to close valves 3 and 4 (Fig. 5A). Then the pressure regulator sent a pressure step of 30 kPa, inflating actuators 1, 2, 5, and 6. The robot arm moved between 2 cm and 5 cm upwards, and the regulator applied -100 kPa deflating the actuators and releasing the grasped object. Figure 5C shows the robot grasping a pencil, a rubber duck, a dried apricot, and a cup. The grasping mode can be viewed in Supplemental Movie 3.
The same robot was then demonstrated locomoting across a table. First, the EPMs were used to close valves 1 and 6. The pressure regulator then repeated a pressure sequence consisting of 30 kPa for 0.2 s followed by -100 kPa for 0.5 s. The resulting hopping motion allowed the robot to travel across the table at a rate of 61.81 mm min−1, equivalent to 0.72 body lengths per minute. Figure 5D shows a time sequence of the robot’s motion over the duration of 1 min. Figure 5E shows a side view of the robot with the selected actuators inflated. The locomotion mode can be viewed in Supplemental Movie 4. As our approach relies on electronic control, the method of reprogramming the robot for varied tasks is straightforward as the designer only needs to input the commands that switches the EPM states and the duration of inflation and deflation of the connecting soft actuators. The input supply pressure could be adjusted in magnitude and duration, as exemplified in our use of the regulators, to generate different responses. Finally, we underscore how electronically activated valves enable real-time control using a six-DoF soft robot capable of switching between two actuation sequences when sent a command from the attached computer. In the first, seen in Fig. 6A, a static pressure of 30 kPa was supplied to the input with the valves initially all in the closed state. The EPMs were then demagnetized in sequence to open the valves, inflating the chambers of the SBAs in a chosen order. -100 kPa was supplied to the input to reset the system. This sequence repeated indefinitely until a command was sent from the attached computer, immediately switching the robot to the second sequence. The second sequence, seen in Fig. 6B highlights the second operation mode. Here, a configuration of SBA actuation was chosen by opening valves before supplying pressure to the robot. A pressure wave consisting of 30 kPa for 0.2 s followed by -100 kPa for 0.5 s was then supplied to the robot. In this way, the pre-selected actuators inflated and deflated in an oscillatory manner, as in the locomotion demonstration detailed above. While in this demonstration the transition between actuation sequences was initiated by a manually sent instruction, this could instead be completed using stimulus to a sensor. This demonstration can be viewed in Supplemental Movie 5.
A With a constant positive pressure supplied to the robot’s inlet, the valves open in sequence, allowing the actuators to inflate. Once all actuators have inflated, the inlet switches to negative pressure, the valves close, and the sequence repeats indefinitely until a keyboard command is sent. B With the valves states fixed, the inlet pressure alternates between positive and negative pressure, inflating and deflating the chosen actuators in sequence for a preset duration. In both A, B, the schemes show, from top to bottom, the valve states, the input pressure, images from the experiment, and schematic representations of the robot state. The time axes are not shown to scale.
In this paper, we presented an EPM valve for controlling multi-DoF pneumatic soft robots. This valve uses the force exerted by an EPM on a steel keeper plate to hold closed a thin fluidic channel manufactured using thermoplastic films. COMSOL simulations and a theoretical model were used to optimize the dimensions of the EPM and keeper plate. The valve is electronically controlled, allowing the pressure to be distributed onboard the robot in real-time or using pre-programmed sequences of actuation. The valve does not contact the working fluid of the robot, in this case air, and has minimal moving parts. The valve is designed to work with soft robots manufactured from thin plastic film. To the best of the authors’ knowledge, it is the first electronically controlled valve designed for this purpose.
More importantly, the EPM valve demonstrates high pressure blocking ability, both in positive (up to 146 kPa) and negative pressure (up to -100 kPa) scenarios. The valve has an operation speed in the range of milliseconds with a response time of ≈ 0.147 s. It effectively blocks the channel even when already pressurized, ensuring precise control over fluid flow. The valve outperforms current electronically controlled valves, both the commercially available and those found in literature in terms of maximum pressure capability, size, and weight while requiring an operational voltage on par with off-the-shelf valves (see Table 1). Figure 3F showcases the EPM valve’s capability to control increasing pressures, underscoring its versatility in managing dynamic pressure conditions. The valve maintains pressure holding for extended periods, with tests indicating stability for over 15 min. This reliability is essential in applications requiring sustained pressure control. Cycle testing validates the EPM valve’s ability to work repeatedly without performance degradation. Despite its small form factor (working height of ≤1 cm) and lightweight construction (2.9 g), the EPM valve can effectively block high pressures, surpassing current valve designs found in the literature designed for centimeter-scale actuators (see Table 1). The EPM valve’s compact design allows for seamless integration into small-scale soft robotic systems. This feature is particularly advantageous for pneumatic actuated soft robots, as it ensures that the mobility and agility of the robot are not compromised while providing pressure control.
While we have demonstrated that the current scale of the EPMs in our proposed valve design is capable of effectively blocking pneumatic pressures up to ≈150 kPa, the EPMs can be scaled down for a more lightweight and compact system. However, the magnetic force and, consequently, the blocking capability of the valve decreases as the dimensions of the end caps decrease as illustrated in our analytical and finite element method (FEM) models of the EPMs (Figure S1). Thus smaller EPM valves would be suitable for soft robots operating at low pressures. In our prior work, we demonstrated the EPM’s functionality with smaller dimensions through the magnetically induced stiffening of magnetorheological (MR) fluid, with the permanent magnets measuring 6.35 mm in length and 1.6 mm in diameter and steel end caps sized at 3.6 × 2 × 1.6. The measured magnetic field strength was ≈ 30 mT in the ON state and approximately 5 mT in the OFF state46, with a reported power requirement of ≈50 mJ to switch states. In contrast, Tugwell et al. demonstrated the efficacy of utilizing larger-scale EPMs in surgical applications55. The study featured a cylindrical EPM measuring 40 mm in diameter and 60 mm in height, weighing approximately 1.5 kg, with a peak holding force of 2 N at a distance of 20 mm. The field experiment conducted reports a peak magnetic field strength of 69 mT in the ON state and 13.75 mT in the OFF state. Additionally, the EPM requires an electrical pulse of 20 A to a solenoid with wire resistance of 3.9 Ω, resulting in an instantaneous power consumption of 1560 W required for switching states – ≈ 17× greater than the EPM used in this work. With that being stated, to ensure proper valve performance across different applications, the dimensions of the EPM and thermoplastic channel must be scaled accordingly. For example, scaling up the EPM while maintaining the same channel size may enhance the valve’s blocking capability, assuming that the increased mass of the larger EPM in the demagnetized state does not impede flow sufficiently to prevent it from being lifted by the flow pressure. Conversely, scaling down the EPM while maintaining the same channel size could increase the actuation speed and responsiveness of the system due to the increased flow rate through the channel. However, this could potentially decrease the blocking capability of the valve due to the larger surface area of the channel that the EPM must contact to block flow.
We then demonstrated the application of the EPM valve in controlling three and six-DoF systems of pneumatic SBAs, thereby illustrating the wide-ranging capabilities enabled by EPM valves in soft robotics applications. In the first demonstration, the EPM valves controlled a robot consisting of three SBAs connected as a single dexterous arm. Using the EPM valves, any of the robot arm’s seven inflated position states could be reached using a single pressure inlet. The robot was first shown reaching each of these seven positions while returning to its initial rest state in between each. The robot was then shown reaching six of these positions without returning to its initial rest state, highlighting the EPM valve’s ability to hold positive pressures in an actuator while deflating another of its DoFs. The inflation of the chambers of the SBA can be controlled in less than 1 s as can be seen in Supplementary Movies S0.7. By employing EPM valves for timed actuation of SBAs, we were able to achieve precise control over fluid flow, facilitating various tasks with pre-programmed actuation sequences. The EPM valves not only enable the selection of which DoF to actuate within a soft robotic system but also act as switching flow controllers by facilitating the switching between constant pressure supplies.
The versatility and effectiveness of EPM valves are then further underscored by their reprogrammability, or the ability to transition between modalities, and regulation of airflow for different actuation sequences, crucial for the adaptability of soft robotic systems. The demonstrations with six-DoF valve where we achieve multi-axis bending motions, grasping functionalities, and locomotion behaviors through the controlled inflation and deflation of two three-DoF SBAs highlight such capability. Particularly, the first of the six-DoF demonstrations showed two three-DoF SBA arms connected in a serial configuration. The EPM valves were used to control the serial arm as it moved through a sequence of dual-curvature configurations. The two three-DoF SBAs were then combined with the six-DoF EPM valve splitter in a flexible acrylic frame to construct a multi-function soft robot to highlight the advantage of onboard electronic control for the purpose of reprogrammability. This robot was first used as an end effector on a commercially available robot arm. The robot grasped irregular objects: a pencil, a cup, a dried apricot, and a rubber duck. In this experiment, the inflation command and movement of the robot arm were controlled by an operator. The EPM valves were preset to allow the grasping motion, and the operator chose when the pressure regulator was engaged. The multi-function robot was then removed from the commercial robot arm and shown to locomote across a table using a hopping motion. The final experiment demonstrated the multi-function robot’s ability to switch between two primary operation modes. In the first, a constant 30 kPa was supplied to the inlet with the EPM valves initially in the closed configuration. The EPMs were then demagnetized in order, opening the valves, and allowing the SBAs to inflate one DoF at a time. A pressure of -100 kPa was then supplied to the robot’s inlet to deflate the actuators, the EPMs were remagnetized to close the valves, and the sequence was repeated. This sequence continued until a keyboard command was sent by the operator. In the second sequence, the EPM states were preset, and an oscillatory inflation pattern was provided by alternating the pressure at the inlet between 30 kPa and -100 kPa. This sequence continued for a set duration. This second sequence highlighted the same operation mode as in the locomotion experiment. An operator initiated the change between the two operation modes in this experiment, but the electronic nature of the valve’s operation would allow them to be integrated with other forms of electronic stimulus, such as from sensors. Together, these operation modes could be useful for a robot that must locomote autonomously until a stimulus is detected, and then complete some task.
In previous works, we demonstrated a fully modulated method of controlling multi-DoF soft robots with EPMs and magnetorheological fluid by using current pulses of different length to achieve intermediate magnetization states, thus altering the material properties of the flowing fluid in a continuous manner42. In this work we only considered the fully on and fully off states of the EPMs, yielding a binary valve. In future works, we will investigate the possibility of using partially magnetized EPMs to allow for proportional control, using ramped pressure inputs to open valves in order of increasing magnetization. This will allow for electronically reconfigurable actuator sequencing, akin to the method using fluidic resistance presented by Vasios et al.56. By modulating the EPM magnetization, it could be possible to use our pneumatic EPM valve not only for binary pressure routing, but also as variable resistors. A parallel configuration of identical EPMs modulated to different magnetizations could be used to affect the actuation sequence and rate of inflation among connected soft actuators. This could be used to create soft robots that perform more complex locomotion movements using a single fixed pressure source, unlike the alternating pressure source used in this paper for our locomoting robot. The result would be a combination of electronically controlled binary and continuous valve hardware that could be swapped in real time.
The solutions to many of the problems in traditional rigid robotics are generally a matter of software. Machine learning, computer vision, coordination of robot swarms, and the many other trending topics in the development of artificial intelligence all presuppose the availability of electronic sensing, actuation, and computation. These techniques are poised to dramatically alter the way humans and robots interact with the world and with each other. If soft robots are to be a part of this future, methods of bridging the gap between hardware and software must be explored. Intrinsically mechanical methods of control will always have their place in soft robots designed for extreme environments, but the bulk of applications will benefit from the real-time feedback afforded by electronics.
An EPM and keeper plate were modeled in 3D using the Magnetic Fields interface and analyzed with a stationary study. For this EPM, a = 8 mm, b = 1.6 mm, d = 3.175 mm, and l = 6.35 mm. Additionally, the height of the end caps was set to 8 mm and the keeper plate was 10 mm in length, 8 mm in width, and 1.6 mm thick. A sphere of air with a diameter of 50 mm, including a 5 mm infinite element domain, surrounded the EPM and keeper plate. The material for the neodymium magnet was set to BMN-40, the alnico magnet was set to BMAca-40/5, and the steel components were set to low carbon steel 1010. A parametric sweep was conducted to vary g from 0.001 mm to 1 mm. With g set to 0.1 mm, additional parametric sweeps were conducted of the end cap height, keeper plate width, keeper plate thickness, and b. In all cases the mesh was set to physics controlled with extra fine element size.
Each EPM was composed of an AlNiCo 5 permanent magnet (CA-0160, Magnet Kingdom), a grade N42 Neodymium magnet (D24, K& J Magnetics), steel end caps, and an electromagnet coil. The permanent magnets each had a diameter of 3.175 mm. The neodymium magnets were 6.35 mm in length, and the alnico magnets were purchased with a length of 12.7 mm. The alnico magnets were manually cut and filed to a length of 6.35 mm. The steel end caps and keeper plates were cut into 8 mm squares and 8 mm × 10 mm rectangles, respectively, from a 1.6 mm thick sheet of low-carbon steel (1388K144, McMaster-Carr) using wire EDM. For each EPM, one alnico magnet and one neodymium magnet were glued side by side to a pair of end caps using a 3D printed alignment jig and superglue (524540, Loctite). Once the glue set, a 150-turn coil electromagnet was wrapped around the permanent magnets using 36-gauge copper motor winding wire (7588K85, McMaster-Carr). The insulation was stripped from the ends of the coil and pin connectors were soldered on.
The flow channels and stacked balloon actuators were manufactured layerwise from 38.1 μm thick thermoplastic elastomer film (Strechlon 200, Fibre Glast) and masks made from 25.4 μm thick teflon film (8569K15, McMaster-Carr). Each layer was first cut using a laser cutter (VLS 6.60, Universal Laser Systems). Steel alignment pins were inserted into an aluminum assembly fixture. A layer of 2.38 mm thick rubber (2614T17, McMaster-Carr) for even pressure distribution and a blank layer of teflon to prevent undesired melting were placed on the fixture. The laser cut layers were then stacked and topped with an additional blank teflon layer, an additional rubber layer, and the top half of the aluminum assembly fixture. The assembly was then placed in a heat press (5420, Carver) and pressed at 90 kPa and 140 °C for 3 min in the case of the flow channels or 5 min for the actuators. The tubes (RPT040, Braintree Scientific) were inserted manually and glued in place using superglue (524540, Loctite) after the heat pressing process. Figure S5 shows an exploded view of the fabrication stack.
The EPM valve was supplied with compressed air regulated by an electronic proportional regulator (ITV0030-2N, SMC) that was controlled using analog output voltages from a data acquisition device (USB-6002, National Instruments) programmed using custom MATLAB scripts (version R2022a).
To obtain the transient responses by increasing step inputs, the EPM valve was fed with a constant pressure supply starting from approximately 50 kPa to 150 kPa in 50-kPa increments. Any fluctuations observed in the measured input and output pressures were attributed to sensor noise which were reduced using a moving average filter. Timed input pulses controlling the states of the EPMs were coordinated using Arduino code featuring built-in timer functions. In both tests, the code execution was synchronized with the initialization of the input voltage to the regulator responsible for supplying the input pressure. The experiment was set to run for 30 s, after which the EPM switches direction (i.e., switch from the initial demagnetized or magnetized states), followed by the recording of the input and output pressures for an additional 30 s. Three trials were conducted for each sequence.
As for the transient response by ramp-up staircase pressure input, the input pressure was gradually ramped up from 0 to 10, 10 to 25, and 25 to 50 kPa. Input pulses controlling the states of the EPMs were operated manually by user-input commands to Arduino Serial Monitor tool as described below.
For all tests, we used electronic pressure sensors (Nidec Copal Electronics P-7100-102GM5) positioned 10 cm upstream of the inlet and a second was positioned 10 cm downstream of the outlet, 10 cm before the connection to a stop valve (McMaster, 7033T21) and outputted electronic signals were captured using the data acquisition device (USB-6002, National Instruments) with a sampling rate of 10 Hz. To facilitate the release of the accumulated compressed air, a 30 gauge dispensing needle (McMaster, 75165A31) was connected at the end of the stop valve in the open position to function as a pull-down resistor due to its high fluidic resistance. All data acquisition and processing were completed using MATLAB scripts (version R2022a).
The magnetic field of the EPMs is modulated using full-bridge motor drivers, which are electronically controlled. The motor drivers are powered by a bench-top power supply (1671A, B&K Precision) set to 20 V. A commercial full-bridge motor driver (Handson Technology BTS7960) controlled by the digital output pins of a DAQ (USB-6353, National Instruments) was initially used. To facilitate control of multiple EPMs simultaneously, the commercial board was substituted with a custom PCB equipped with pairs of BTN8962TA half-bridge integrated circuits (Infineon Technologies) controlled by an Arduino Nano. The custom PCB is capable of achieving current pulse durations down to approximately 3 μs.
An experiment was conducted to validate the force between the EPM and steel keeper plate as predicted by COMSOL and equations (1) and (2). A tensile testing machine (5943, Instron) mounted with a 20 N load cell and a custom 3D printed fixture was used to pull the EPM away from a steel keeper plate at a rate of 1 mm min−1 over a distance of 1 mm. Before each trial, the EPM was placed in direct contact with the keeper plate under a small compressive load and fully magnetized. Three trials were conducted, and the data was compiled in MATLAB.
Data presented in the manuscript is available upon request to the corresponding author.
Whitesides, G. M. Soft Robotics. Angew. Chem. - Int. Ed. 57, 4258–4273 (2018).
Hauser, H., Nanayakkara, T. & Forni, F. Leveraging morphological computation for controlling soft robots: learning from nature to control soft robots. IEEE Control Syst. 43, 114–129 (2023).
Pfeifer, R., Iida, F. & Lungarella, M. Cognition from the bottom up: On biological inspiration, body morphology, and soft materials https://linkinghub.elsevier.com/retrieve/pii/S1364661314000862 (2014).
Kim, Y. J., Cheng, S., Kim, S. & Iagnemma, K. D. A Stiffness-Adjustable Hyperredundant Manipulator Using a Variable Neutral-Line Mechanism for Minimally Invasive Surgery. IEEE Trans. Robot. 30, 382–395 (2013).
Scimeca, L. & Iida, F.Sensory Systems for Robotic Applications. January 2023 (Institution of Engineering and Technology, 2022). https://digital-library.theiet.org/content/books/ce/pbce097e.
Giordano, G., Carlotti, M. & Mazzolai, B. A Perspective on Cephalopods Mimicry and Bioinspired Technologies toward Proprioceptive Autonomous Soft Robots. Adv. Mater. Technol. 2100437, 2100437 (2021).
Ranzani, T., Russo, S., Bartlett, N. W., Wehner, M. & Wood, R. J. Increasing the dimensionality of soft microstructures through injection-induced self-folding. Adv. Mater. 30, 1802739 (2018).
Mahon, S. et al. Capability by Stacking: The Current Design Heuristic for Soft Robots. Biomimetics 3, 16 (2018).
McDonald, K. & Ranzani, T. Hardware methods for onboard control of fluidically actuated soft robots. Front. Robot. AI 8, 1–19 (2021).
Yin, A., Lin, H.-C., Thelen, J., Mahner, B. & Ranzani, T. Combining Locomotion and Grasping Functionalities in Soft Robots. Adv. Intell. Syst. 1, 1900089 (2019).
Polygerinos, P. et al. Soft Robotics: Review of Fluid-Driven Intrinsically Soft Devices; Manufacturing, Sensing, Control, and Applications in Human-Robot Interaction. Adv. Eng. Mater. 19, 1700016 (2017).
Gorissen, B. et al. Elastic Inflatable Actuators for Soft Robotic Applications. Adv. Mater. 29, 1604977 (2017).
Joshi, S., Sonar, H. & Paik, J. Flow path optimization for soft pneumatic actuators: towards optimal performance and portability. IEEE Robot. Autom. Lett. 3766, 1–1 (2021).
Stanley, A. A. et al. Lumped-parameter response time models for pneumatic circuit dynamics. J. Dyn. Syst., Meas. Control, Trans. ASME 143, 1–11 (2021).
Mahon, S. T. et al. Capability by stacking: The current design heuristic for soft robots. Biomimetics 3, 1–16 (2018).
Unger, M. A., Chou, H. P., Thorsen, T., Scherer, A. & Quake, S. R. Monolithic microfabricated valves and pumps by multilayer soft lithography. Sci. (N. Y., N. Y.) 288, 113–116 (2000).
Rothemund, P. et al. A soft, bistable valve for autonomous control of soft actuators. Sci. Robot. 3, 1–10 (2018).
Wehner, M. et al. An integrated design and fabrication strategy for entirely soft, autonomous robots. Nature 536, 451–455 (2016).
Preston, D. J. et al. A soft ring oscillator. Sci. Robot. 4, 1–9 (2019).
Preston, D. J. et al. Digital logic for soft devices. Proc. Natl. Acad. Sci. 116, 7750–7759 (2019).
Mahon, S. T., Buchoux, A., Sayed, M. E., Teng, L. & Stokes, A. A. Soft robots for extreme environments: Removing electronic control. In RoboSoft 2019 - 2019 IEEE International Conference on Soft Robotics, 782–787 (2019).
Nemitz, M. P. et al. Soft non-volatile memory for non-electronic information storage in soft robots. In IEEE International Conference on Soft Robotics 2020, 7–12 (New Haven, 2020).
Decker, C. J. et al. Programmable soft valves for digital and analog control. Proc. Natl. Acad. Sci. 119, 1–10 (2022).
Conrad, S. et al. 3D-printed digital pneumatic logic for the control of soft robotic actuators. Sci. Robot. 9, eadh4060 (2024).
Bartlett, N. W., Becker, K. P. & Wood, R. J. A fluidic demultiplexer for controlling large arrays of soft actuators. Soft Matter 16, 5871–5877 (2020).
Lu, Q., Xu, H., Guo, Y., Wang, J. Y. & Yao, L. Fluidic computation kit: Towards electronic-free shape-changing interfaces. In Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems, 1–21 (2023).
Drotman, D., Jadhav, S., Sharp, D., Chan, C. & Tolley, M. T. Electronics-Free Pneumatic Circuits for Controlling Soft Legged Robots. Science Robotics 6, eaay2627 (2021).
Connolly, F., Wagner, D. A., Walsh, C. J. & Bertoldi, K. SI_Sew-free anisotropic textile composites for rapid design and manufacturing of soft wearable robots. Extrem. Mech. Lett. 27, 52–58 (2019).
Shveda, R. A. et al. A wearable textile-based pneumatic energy harvesting system for assistive robotics. Sci. Adv. 8, 1–11 (2022).
Rajappan, A. et al. Logic-enabled textiles. Proc. Natl. Acad. Sci. 119, 1–9 (2022).
Sanchez, V., Walsh, C. J. & Wood, R. J. Textile technology for soft robotic and autonomous garments. Adv. Funct. Mater. 31, 2008278 (2021).
Zhu, M. et al. Soft, Wearable Robotics and Haptics:Technologies, Trends, and Emerging Applications. Proceedings of the IEEE 110, 246–272 (2020).
Rogatinsky, J. et al. A Collapsible Soft Actuator Facilitates Performance in Constrained Environments. Adv. Intell. Syst. 2200085, 2200085 (2022).
Becker, S., Ranzani, T., Russo, S. & Wood, R. J. Pop-up tissue retraction mechanism for endoscopic surgery. In 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 920–927 (IEEE, 2017). http://ieeexplore.ieee.org/document/8202255/.
Ranzani, T., Russo, S., Schwab, F., Walsh, C. J. & Wood, R. J. Deployable stabilization mechanisms for endoscopic procedures. In 2017 IEEE International Conference on Robotics and Automation (ICRA), 1125–1131 (IEEE, 2017). http://ieeexplore.ieee.org/document/7989134/.
Runciman, M., Avery, J., Darzi, A. & Mylonas, G. Open loop position control of soft hydraulic actuators for minimally invasive surgery. Appl. Sci. 11, 7391 (2021).
Rogatinsky, J. et al. A multifunctional soft robot for cardiac interventions. Sci. Adv. 9, eadi5559 (2023).
Park, Y. J., Ko, M. G., Jamil, B., Shin, J. & Rodrigue, H. Simple and Scalable Soft Actuation Through Coupled Inflatable Tubes. IEEE Access 10, 41979–41989 (2022).
Xu, S., Chen, Y., Hyun, N.-sP., Becker, K. P. & Wood, R. J. A dynamic electrically driven soft valve for control of soft hydraulic actuators. Proc. Natl. Acad. Sci. 118, 1–9 (2021).
Poccard-Saudart, J. et al. Controlling soft fluidic actuators using soft DEA-based valves. IEEE Robot. Autom. Lett. 7, 8837–8844 (2022).
Wang, Z. et al. Self-Closing and Self-Healing Multi-Material Suction Cups for Energy-Efficient Vacuum Grippers. Advanced Intelligent Systems 5, (2023).
Mcdonald, K., Kinnicutt, L., Moran, A. M. & Ranzani, T. Modulation of Magnetorheological Fluid Flow in Soft Robots Using Electropermanent Magnets. IEEE Robot. Autom. Lett. 3766, 1–1 (2022).
Knaian, A.Electropermanent magnetic connectors and actuators: devices and their application in programmable matter. Doctor of philosophy in electrical engineering and computer science, Massachusetts Institute of Technology http://dspace.mit.edu/handle/1721.1/60151 (2010).
Marchese, A. D., Onal, C. D. & Rus, D. Soft robot actuators using energy-efficient valves controlled by electropermanent magnets. In 2011 IEEE/RSJ International Conference on Intelligent Robots and Systems, 1, 756–761 (IEEE, 2011). http://ieeexplore.ieee.org/document/6095064/.
McDonald, K., Rendos, A., Woodman, S., Brown, K. A. & Ranzani, T. Magnetorheological fluid based flow control for soft robots. Advanced Intelligent Systems 2000139 (2020).
Gaeta, L. T. et al. Magnetically induced stiffening for soft robotics. Soft Matter 19, 2623–2636 (2023).
Sadeghi, A., Beccai, L. & Mazzolai, B. Design and development of innovative adhesive suckers inspired by the tube feet of sea urchins. Proceedings of the IEEE RAS and EMBS International Conference on Biomedical Robotics and Biomechatronics 617–622 (2012).
Tonazzini, A. et al. Variable Stiffness Fiber with Self-Healing Capability. Adv. Mater. 28, 10142–10148 (2016).
Zatopa, A., Walker, S. & Menguc, Y. Fully Soft 3D-Printed Electroactive Fluidic Valve for Soft Hydraulic Robots. Soft Robot. 5, soro.2017.0019 (2018).
Bira, N., Menguc, Y. & Davidson, J. R. 3D-Printed Electroactive Hydraulic Valves for Use in Soft Robotic Applications. Proceedings - IEEE International Conference on Robotics and Automation 11200–11206 (2020).
Huang, T. et al. A Lightweight Flexible Semi-Cylindrical Valve for Seamless Integration in Soft Robots Based on the Giant Electrorheological Fluid. SSRN Electron. J. 347, 113905 (2022).
Leps, T. et al. A low-power, jamming, magnetorheological valve using electropermanent magnets suitable for distributed control in soft robots. Smart Mater. Struct. 29, 105025 (2020).
Connolly, F., Polygerinos, P., Walsh, C. J. & Bertoldi, K. Mechanical programming of soft actuators by varying fiber angle. Soft Robot. 2, 26–32 (2015).
Mosadegh, B. et al. Pneumatic networks for soft robotics that actuate rapidly. Adv. Funct. Mater. 24, 2163–2170 (2014).
Tugwell, J. et al. Electropermanent magnetic anchoring for surgery and endoscopy. IEEE Trans. Biomed. Eng. 62, 842–848 (2014).
Vasios, N., Gross, A. J., Soifer, S., Overvelde, J. T. & Bertoldi, K. Harnessing viscous flow to simplify the actuation of fluidic soft robots. Soft Robot. 7, 1–9 (2020).
Gerboni, G., Henselmans, P. W., Arkenbout, E. A., Furth, W. R. & Breedveld, P. HelixFlex : bioinspired maneuverable instrument for skull base surgery. Bioinspiration Biomim. 10, 66013 (2015).
Booth, J. W., Case, J. C., White, E. L., Shah, D. S. & Kramer-bottiglio, R. An Addressable Pneumatic Regulator for Distributed Control of Soft Robots 2–7 (2018).
Moers, A. J., De Volder, M. & Reynaerts, D. Integrated high pressure microhydraulic actuation and control for surgical instruments. Biomed. Microdev. 14, 699–708 (2012).
This work was supported by the Office of Naval Research (ONR) grant number N00014-22-1-2244. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the ONR.
These authors contributed equally: Anna Maria Moran, Vi T. Vo.
Department of Mechanical Engineering, Boston University, Boston, MA, USA
Anna Maria Moran, Vi T. Vo, Kevin J. McDonald, Pranav Sultania, Eva Langenbrunner, Jun Hong Vince Chong, Amartya Naik, Lorenzo Kinnicutt, Jingshuo Li & Tommaso Ranzani
Department of Biomedical Engineering, Boston University, Boston, MA, USA
Materials Science and Engineering Division, Boston University, Boston, MA, USA
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
A.M.M., V.T.V., K.J.M., P.S., T.R. conceived and designed the research; K.J.M., P.S. carried out the analytical model and simulations; A.M.M., V.T.V., K.J.M. conducted the experiments and performed the characterizations; A.M.M., V.T.V., K.J.M. designed and conducted the demonstrations; A.M.M., V.T.V., K.J.M., E.L., J.H.V.C., A.N., J.L. fabricated the samples; L.K. designed the electronics control; A.M.M., V.T.V., K.J.M. analyzed the data and prepared the figures; K.J.M., T.R. wrote the initial draft of the manuscript; A.M.M., V.T.V., P.S., T.R. wrote and revised the manuscript; T.R. supervised the study.
The authors declare no competing interests.
Communications Engineering thanks Chaoqun Xiang and the other, anonymous, reviewers for their contribution to the peer review of this work. Peer reviewer reports are available.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
Moran, A.M., Vo, V.T., McDonald, K.J. et al. An electropermanent magnet valve for the onboard control of multi-degree of freedom pneumatic soft robots. Commun Eng 3, 117 (2024). https://doi.org/10.1038/s44172-024-00251-y
DOI: https://doi.org/10.1038/s44172-024-00251-y
Anyone you share the following link with will be able to read this content:
Sorry, a shareable link is not currently available for this article.
Provided by the Springer Nature SharedIt content-sharing initiative
Communications Engineering (Commun Eng) ISSN 2731-3395 (online)
ball valve pressure Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.