Artificial muscle

Abstract

McKibben-type pneumatic rubber artificial muscles are known as soft actuators for use in robot devices. However, the McKibben-type pneumatic rubber artificial muscles require a pneumatic supply system, which poses a problem for miniaturization.　This artificial muscle includes a cylindrical braided tube that has side surfaces equidistant from a central axis formed by braiding fibers, end securing parts for securing both ends of the braided tube, respectively, a motor that has a drive shaft connected to one of the end securing parts in alignment with the central axis, and a drive control part for driving the motor, the artificial muscle being lightweight, not bulky, and can be suitably used in a wearable environment.

Description

Artificial Muscle

[0001] The present invention relates to an actuator used in a robot device, and more particularly to a soft actuator that is lightweight and flexible.

[0002] Variable stiffness actuators have conventionally been used in robot devices. Patent Document 1 discloses an actuator device that includes a rotating base that rotates by being directly or indirectly connected to a motor, a sliding base that is slidable on the rotation axis of the rotating base, one or more weight members that rotate together with the rotating base and are movable outward by centrifugal force, and a link mechanism that uses the centrifugal force to change the distance between the rotating base and the sliding base.

[0003] This involves connecting the rotating shafts with a link mechanism, attaching weights to the joints of the link mechanism, and utilizing the fact that when the rotating shafts are rotated at high speed, the centrifugal force causes the weights to expand in a direction perpendicular to the rotating shafts, thereby generating a force in the direction of the rotating shafts by bringing the rotating shafts closer together.

[0004] In recent years, pneumatic soft actuators with passive flexibility have been attracting attention in the field of robotics. Among these, McKibben (hereafter referred to as "McKibben")-type pneumatic rubber artificial muscles are lightweight, flexible, have a high power-to-weight ratio, and are highly compatible with humans. As a result, they are increasingly being used as driving devices for work assistance and medical and welfare robots. Another notable characteristic of McKibben-type pneumatic rubber artificial muscles is their high backdrivability, which is due to the compressibility of air. Here, backdrivability refers to the ease with which they move in the direction of an external force when the driving force is cut off.

[0005] Patent Document 2 discloses a composite McKibben artificial muscle that contracts uniformly and whose contraction rate and contraction force can be easily adjusted.

[0006] JP 2009-115173 A JP 2020-56437 A

[0007] McKibben pneumatic artificial rubber muscles contract when compressed air is supplied to the rubber tube. Therefore, an air pressure supply system, such as an air compressor or tank, is required to drive them, which reduces their wearability and limits the environments in which they can be used.

[0008] The present invention was conceived in view of the above problems, and provides an artificial muscle that is lighter than a McKibben type pneumatic rubber artificial muscle that includes a rigid actuator, a drive unit, and the like, and that exhibits backdrivability characteristics that are comparable to those of a McKibben type pneumatic rubber artificial muscle.

[0009] More specifically, the artificial muscle according to the present invention is characterized by having a cylindrical braided tube formed by weaving fibers on sides equidistant from a central axis; end fixing parts that fix both ends of the braided tube; a motor whose drive shaft is connected to one of the end fixing parts in alignment with the central axis; and a drive control part that drives the motor.

[0010] The drive system of the artificial muscle of the present invention is composed only of a braided tube and a motor that rotates it, and because it does not have a compressed air supply device like McKibben artificial muscle, it is possible to make the entire device lighter, and in particular, reduce its bulk, making it suitable for use in wearable devices.

[0011] Although it can be said to be an actuator that uses centrifugal force, the braided tube is softer than rigid actuators, making it highly suitable as a soft actuator.

[0012] Furthermore, because it uses a lightweight braided tube, it can withstand external forces in the opposite direction without breaking, meaning it has high backdrivability.

[0013] FIG. 1 is a diagram illustrating the configuration of an artificial muscle according to the present invention. FIG. 2 is a diagram explaining a method for evaluating a braided tube used in an artificial muscle. FIG. 3 is a graph showing the characteristics of a braided tube made from PET fiber. FIG. 4 is a graph showing the characteristics of a braided tube made from stainless steel fiber. FIG. 5 is a graph comparing the characteristics of a braided tube made from PET fiber and a braided tube made from stainless steel fiber. FIG. 6 is a graph showing the relationship between the contraction amount and angular velocity when an artificial muscle is operating. FIG. 7 is a graph showing the dynamic characteristics when an artificial muscle is contracted and extended, as the relationship between the contraction rate and the force obtained. FIG. 8 is a diagram explaining various specifications when considering an operating model of an artificial muscle. FIG. 9 is a graph comparing an artificial muscle with a model (calculation).

[0014] An artificial muscle according to the present invention will be described below with reference to the drawings. Note that the following description exemplifies one embodiment and one example of the present invention, and the present invention is not limited to the following description. The following description can be modified within the scope of the invention.

[0015] Furthermore, embodiments and examples obtained by appropriately combining the technical means disclosed in different embodiments and examples are also included in the technical scope of the present invention. Furthermore, all documents described in this specification are incorporated herein by reference. In this specification, when a numerical range is described as "A to B," the description means "A or more (including A or more) and B or less (including B or less)."

[0016] <Configuration of Artificial Muscle> Figure 1 shows the configuration of an artificial muscle according to the present invention. Referring to Figure 1(a), the artificial muscle 1 is composed of a braided tube 10, end fixing parts 12 that fix both ends of the braided tube 10, and a motor 14 with a drive shaft 14a connected to one of the end fixing parts 12a. The motor 14 is connected to a drive control part 16. The drive control part 16 has a built-in power supply (not shown). The power supply may be provided separately from the drive control part 16.

[0017] [Braided Tube 1] The braided tube 10 is a tubular (cylindrical) tube made of resin or metal fibers woven together with a diameter of 10 μm to several mm. The direction connecting the two ends is called the longitudinal direction. While resin or metal is preferred as the material for the fibers making up the braided tube 10, there are no particular limitations. Fibers made of glass, silicone, or carbon fiber can also be used. Commercially available polyester tubes for wire protection can also be used.

[0018] There are no particular limitations on the way the fibers are woven. For example, a single fiber can be woven, or multiple fibers can be bundled and woven. However, when used in artificial muscle 1, the braided tube must have elasticity. Therefore, weaving methods that increase the coefficient of friction between the fibers or that fix the fibers together are not desirable.

[0019] Furthermore, the braided tube 10 needs to follow the rotation of the motor 14 in a braided state. Therefore, a braiding method that weakens the braided tube 10 is not preferable. Also, it is preferable that the braided fibers slide against each other, so that the braided tube can easily deform (thicken) in the radial direction. There are also no particular limitations on the length or diameter. The weight of the braided tube 10 is 1.0 x 10 －４ kg to 5.0 kg, and more preferably 1.0 x 10 －３ kg to 0.1 kg, and most preferably 1.0 x 10 －３ kg ~ 0.5 x 10 －１ kg is good.

[0020] As described above, the specifications for the braided tube 10 used in the artificial muscle 1 of the present invention are quite broad, so what kind of braided tube 10 is suitable for the present invention will be described later, including measurement methods. The central axis of the braided tube 10 in the longitudinal direction is represented by the central axis 10c.

[0021] [End Fixing Portion] Next, the end fixing portion 12 will be described. The end fixing portion 12 includes an upper end fixing portion 12a (one of the end fixing portions) connected to the motor 14 and a lower end fixing portion 12b connected to a workpiece (not shown). These end fixing portions 12 are fixed so that their center of rotation is the central axis 10c of the braided tube 10.

[0022] There are no particular limitations on the shape of the end fixing portion 12 or the method of connecting it to the braided tube 10, as long as it can fix the end of the braided tube 10. For example, a flat round resin plate can be used as the end fixing portion 12, and the end of the braided tube 10 can be fixed thereto with an adhesive. The end fixing portion 12 is fixed to the motor 14 or a workpiece (not shown), and therefore functions as a flange itself. Therefore, it may be processed to be fixed to the motor 14 or a workpiece (not shown).

[0023] [Motor] The motor 14 includes a main body 14b and a drive shaft 14a. There are no particular limitations on the specifications of the motor 14, but it is preferable that the motor 14 be capable of providing sufficient rotation to the braided tube 10. The drive shaft 14a is fixed to the upper end fixing portion 12a so as to coincide with the central axis 10c of the braided tube 10.

[0024] [Drive Control Unit] A computer typically configured with a CPU and memory can be suitably used as the drive control unit 16. The drive control unit 16 may control the motor 14 independently, or may control the motor 14 according to instructions from another control device. Note that control of the motor 14 is control of the rotation of the braided tube 10, and is control of the artificial muscle 1.

[0025] The drive control unit 16 is connected to the motor 14 via a signal line 15. The signal line 15 supplies a controlled drive current to the motor 14. In other words, the drive control unit 16 can variably control the rotation speed of the motor 14. Of course, digital control may be used to rotate the motor 14 at a constant speed or not at all.

[0026] <Operation of Artificial Muscle> See Figure 1(b). Figure 1(b) shows the operating state of the artificial muscle 1. The drive shaft 14a of the motor 14 rotates due to the current Ic from the drive control unit 16 (not shown). The rotation of the drive shaft 14a of the motor 14 rotates the braided tube 10. When the braided tube 10 rotates, centrifugal force is applied to the fiber portion, which tries to move away from the central axis 10c. As a result, the braided tube 10 contracts in the longitudinal direction.

[0027] If the length of the braided tube 10 before rotation is L1 and the length after rotation is L2, the braided tube 10 contracts by L1 - L2 = ΔL1. At this time, a force Fa is generated in the lower end fixing part 12b. This force Fa is the force that the artificial muscle 1 can exert.

[0028] Here, if the braided tube 10 is pulled in the opposite direction to the force Fa with a force greater than the force Fa, the length L2 of the braided tube 10 will increase, thereby demonstrating backdrivability.

[0029] [Braided Tube 2] Next, a braided tube 10 that can be suitably used for the artificial muscle 1 according to the present invention will be described with reference to Figure 2. As explained above in <Operation of Artificial Muscle>, the braided tube 10 contracts in length when rotated around the central axis 10c, and as a result, the artificial muscle 1 obtains a force in the longitudinal direction.

[0030] Since the purpose of the artificial muscle 1 is to obtain force in the longitudinal direction, a braided tube 10 that cannot generate a predetermined force in the longitudinal direction is not worthy of use. Therefore, using a device such as that shown in Figure 2, it is possible to determine whether or not the braided tube 10 can be used from the perspective of power-to-weight ratio.

[0031] FIG. 2(a) shows a braided tube 10 with a mass m (kg) and a length La (m) and provided with end fixing portions 12 (12a and 12b). FIG. 2(b) shows the state when the braided tube 10 is displaced by ΔLd. The measurement system includes a displacement meter 40 that measures the length of the braided tube 10 and a stress meter 30 that measures the reaction force of the braided tube 10 at that time. The displacement meter 40 is composed of a detection unit 42 and a main body 44, and the stress meter 30 is composed of a force sensor 32 and a main body 34. These outputs are input into a computer (not displaying them) as needed and are plotted as a graph of shrinkage rate versus stress.

[0032] The data input into the computer is smoothed using a low-pass filter with a cutoff frequency of 1 Hz, because raw data is prone to noise, making it difficult to determine the indices described below.

[0033] Increasing the mass of the braided tube 10 can potentially produce a large force (contractile force). However, responsiveness will be lower (slower). On the other hand, if the mass of the braided tube 10 is small, responsiveness will be high, but it can be predicted that the force produced will be smaller. By statically squeezing this braided tube 10 to a certain length (Lb) and measuring the reaction force F of the braided tube 10 at that time, it is possible to determine an index for evaluating the performance of the artificial muscle 1 that is independent of the material or size. This index is called the index PWR and is defined as shown in equation (1).

[0034]

[0035] Here, m is the mass (kg) of the braided tube 10, ε is the shrinkage rate (unitless) of the braided tube 10, and F is the repulsive force (N) when the braided tube 10 is deformed. Repulsive force is also referred to as reactive force. Therefore, the unit of the indicator PWR is (kg·N). The shrinkage rate ε is calculated using equation (2), and decimal values ​​are used as is.

[0036]

[0037] The index PWR is the value obtained by multiplying the reaction force F obtained when the contraction rate is ε by the mass of the braided tube 10 at that time. In other words, if the contraction rate ε is fixed, the characteristics of the braided tube 10 can be specified regardless of its length, diameter, or material. As shown in the examples below, stable characteristics can be obtained by measuring the contraction rate ε between 10% and 20%. Here, the value measured at a contraction rate ε of 10% (= 0.1) is used as the index for estimating the performance of the artificial muscle 1. Therefore, the equation for the index PWR, which represents the performance of the braided tube 10 used in the artificial muscle 1 according to the present invention, can be obtained by rewriting equation (1) as equation (3).

[0038]

[0039] The artificial muscle 1 according to the present invention is intended to be a soft actuator with a total weight of 5 kg or less, so the index PWR is 1.0 × 10 －５ A range of 5.6 (kg / N) to 5.6 (kg / N) is considered to be the preferred range for the artificial muscle 1 of the present invention. Note that the mass m in equations (1) and (3) may include the mass of the weight attached to the braided tube 10. An example of the state in which the weight is attached to the braided tube 10 is shown in Figure 1(c). Weights 20 are attached to the side walls of the braided tube 10 so that the rotational moment is uniform.

[0040] (Example 1: Static properties of braided tube) The test shown in Figure 2 was carried out using a braided tube 10 made of PET resin fibers and a braided tube 10 made of stainless steel fibers. The specifications of the samples are shown in Table 1. Four samples were prepared. Samples 1 and 2 were made of PET fibers, and Samples 3 and 4 were made of stainless steel fibers (marked "SUS"). The length and diameter are the sizes when made into the braided tube 10.

[0041]

[0042] The results are shown in Figures 3 to 5. In Figures 3 to 5, the horizontal axis is the contraction ratio (%) and the vertical axis is the reaction force (N). Please refer to Figure 3. Figure 3 shows the results for Sample 1 and Sample 2. The difference between Sample 1 and Sample 2 is that Sample 2 is longer than Sample 1. Accordingly, Sample 2 is also heavier.

[0043] The reaction force decreased as the sample length increased. The shrinkage rate-reaction force curve for Sample 1 (hereinafter referred to as the "shrinkage rate-reaction force characteristic curve") showed a unique shape, with a rising section up to a shrinkage rate of 10%, followed by a constant reaction force section, after which the reaction force increased again from a shrinkage rate of about 30%.

[0044] On the other hand, the shrinkage rate-reaction force characteristic curve of Sample 2 was shifted downward compared to Sample 1. However, the shapes of the shrinkage rate-reaction force characteristic curves were similar.

[0045] Please refer to Figure 4. Figure 4 shows the results for Sample 3 and Sample 4. The difference between Sample 3 and Sample 4 is that Sample 3 is thick and short, while Sample 4 is thin and long. Both samples have the same weight. The characteristic curve of shrinkage rate vs. reaction force had a similar shape to that shown in Figure 3. The reaction force measured for Sample 3 was higher, and although it is expected that there is an upper limit, it suggests that a thick and short braided tube 10 is more likely to generate a reaction force.

[0046] Next, let's look at Figure 5. Figure 5 compares Sample 1 (PET fiber) and Sample 3 (stainless steel fiber). Because their rigidity differs by an order of magnitude, it is difficult to compare them on the same vertical axis, and Sample 1 shows the reaction force on the right vertical axis, while Sample 3 shows the reaction force on the left vertical axis. Although the magnitude of the reaction force differs by an order of magnitude, the shapes of the shrinkage rate-reaction force characteristic curves are similar.

[0047] From the above, by examining the relationship between the shrinkage rate and the reaction force measured as shown in Figure 2, it is possible to comprehensively evaluate the size, material, and weaving method of the braided tube 10. More specifically, since a constant range of reaction force is observed between the shrinkage rate of approximately 10% and 20%, it is reasonable to use the reaction force value relative to the shrinkage rate within this range as an index of the braided tube 10.

[0048] (Example 2: Dynamic characteristics of artificial muscle) <Experimental conditions> To confirm the contractile force characteristics of the artificial muscle 1, a brushless DC motor (BLDC motor) was rotated with the artificial muscle 1 fixed to a manual stage. The artificial muscle 1 was contracted or extended while being driven by the manual stage, and the contractile force at that time was measured with a load cell attached to the shaft side. In the experiment, the motor rotation speed was increased until a contractile force of 30 N and 20 N was generated at the initial length.

[0049] From there, the tube was contracted until the contraction rate reached approximately 30%, and then returned to its initial length (contraction rate 0% → 30% → 0%), at which point the contraction force was measured. The braided tube 10 used here was made of PET and had an initial length of 188 mm, an initial outer diameter of 18 mm, and a mass of 4.5 g. Furthermore, preliminary experiments confirmed that the rotation speed decreased as the artificial muscle 1 contracted. Therefore, an angular velocity sensor was attached to the manual stage, and the angular velocity at 5 mm intervals was also recorded for each contraction amount.

[0050] <Experimental results> In order to obtain a contraction force of 30 N at the initial length, the initial angular velocity ω ０ is 1.41 x 10 ３ rad / s, 1.11 x 10 for 20N ３ rad / s was required. Figure 6 shows the change in rotation speed relative to the contraction amount. Referring to Figure 6, the horizontal axis represents the contraction amount [m] and the vertical axis represents the angular velocity [rad / s]. Circles represent the case of 30N, and black squares represent the case of 20N.

[0051] Figure 6 shows that under both conditions, angular velocity decreased as the contraction amount increased, and the slope a was similar for both 30 N and 20 N. The reason for the decrease in angular velocity is thought to be air resistance acting on the fibers during rotation. The braided tube 10 expands radially as it contracts, so the radius of rotation also increases. This also increases the circumferential speed, and this air resistance acts as a load torque on the motor, reducing the rotation speed. These results confirm a displacement dependency whereby contractile force decreases as the contraction rate increases, similar to the output characteristics of McKibben artificial muscle.

[0052] Figure 7 shows the relationship between contraction rate and contractile force. Referring to Figure 7, the horizontal axis is contraction rate (%) and the vertical axis is contractile force (N). The upper line shows the results for 30 N, and the lower line shows the results for 20 N. Figure 7 confirms a displacement dependency whereby the contractile force decreases as the contraction rate increases, similar to the output characteristics of McKibben artificial muscle.

[0053] Furthermore, under the 30 N condition, hysteresis occurred between contraction rates of 16% and 20%. This was because the axis of the fiber sleeve vibrated during the experiment, creating an unstable state, and it was thought that this was due to misalignment of the rotation axis when the artificial muscle was manufactured.

[0054] Although suppressing vibrations remains an issue, it was confirmed that contractile force can be generated over a wide range of motion. The contraction rate of McKibben artificial muscle is approximately 25% to 30%, but the artificial muscle 1 according to the present invention exceeds 30%. This is thought to be because the artificial muscle 1 according to the present invention does not use rubber materials that suppress expansion.

[0055] <Contraction Force Estimation Model> A mathematical model for estimating the contraction force of the artificial muscle according to the present invention will be considered. The variables used are summarized in Table 2. The centrifugal force is determined by the mass m of the braided tube, the outer diameter D ｏｕｔ , which can be calculated from the angular velocity ω using equation (4).

[0056]

[0057]

[0058] The fiber refers to the fibers that make up the braided tube. ０ , L ｚ０ , D ｏｕｔ０ are the initial angular velocity, the initial length of the artificial muscle (initial length of the braided tube), and the initial diameter of the sleeve (initial diameter of the braided tube), respectively. ｚ０ , D ｏｕｔ０ is the initial angular velocity ω ０ are the length and diameter at time.

[0059] D in the above formula ｏｕｔis a function of the contraction amount dL of the artificial muscle. The geometric relationship of the braided tube is shown in Figures 8(a) and 8(b). Assuming that the braided tube contracts while maintaining a perfect cylindrical shape when the artificial muscle contracts, D ｏｕｔ is calculated from the following equation (5).

[0060]

[0061] The fiber length Lf can be calculated from equation (6) assuming that the fiber does not expand or contract during driving.

[0062]

[0063] The fiber crossing angle α can be calculated from the total length Lz of the artificial muscle using the following formula (7).

[0064]

[0065] The number of rotations of the motor corresponding to the amount of contraction is calculated from the above experimental results using the following formula (8) and substituted into formula (4).

[0066]

[0067] In addition, ω ０ is the initial angular velocity, and "a" is the slope obtained from the graph of contraction amount vs. angular velocity in Figure 6. Finally, the centrifugal force is converted into the contraction force in the axial direction. From Figure 8(c), the contraction force F is calculated from equation (9) using the fiber crossing angle α. ｅ At this time, the centrifugal force is multiplied by the efficiency η. This is because both ends of the braided tube are fixed with resin parts, so it is taken into consideration that not all of the mass of the braided tube is converted into centrifugal force.

[0068]

[0069] The contractile force estimated from the above equation is compared with the experimental value. The resulting graph is shown in Figure 9. Referring to Figure 9, the horizontal axis is the contraction rate (%) and the vertical axis is force [N]. The force here corresponds to the contractile force that the artificial muscle can exert. The upper line is for 30 N, and the lower line is for 20 N.

[0070] The efficiency η in this case was determined to be 0.1 so that the calculated maximum contraction force would be 30 N and 20 N. From the results in Figure 9, it can be seen that the estimated contraction force also tends to decrease as the contraction amount increases. In addition, the modeling error could be confirmed, and the error increased as the contraction amount increased.

[0071] One reason for this is the difference in the shape of the braided tube when it contracts. For simplicity, the model constructed here assumed that it contracted into a perfect cylindrical shape. However, when an artificial muscle actually contracts, it is not a perfect cylinder, with the center bulging out significantly and the outer diameter becoming smaller towards the ends. This is thought to be the reason why the radius of rotation when calculating centrifugal force differs. We also surmise that the reason the effect of centrifugal force is small when η is 0.1 is related to the shape when it contracts.

[0072] The artificial muscle according to the present invention is small and not bulky, and can be suitably used in environments such as small robots and wearable devices.

[0073] REFERENCE SIGNS LIST 1 artificial muscle 10 braided tube 10c central axis 12 end fixing portion 12a upper end fixing portion 12b lower end fixing portion 14 motor 14b main body 14a drive shaft 15 signal line 16 drive control unit 20 weight 30 stress meter 32 force sensor 34 main body 40 displacement meter 42 detection unit 44 main body Ic current

Claims

1. An artificial muscle comprising: a cylindrical braided tube formed by weaving fibers on sides equidistant from a central axis; end fixing parts that fix both ends of the braided tube; a motor whose drive shaft is connected to one of the end fixing parts in alignment with the central axis; and a drive control part that drives the motor.

2. The braided tube has a mass of the braided tube and the end fixing portion in m [kg], and the reaction force of the braided tube when the braided tube is shrunk in the full length direction along the central axis of the braided tube so that the shrinkage rate is 10% is F [N]. In this case, the index PWR expressed by the formula (3) is 1.0×10 －５ The artificial muscle according to claim 1, wherein the Mn content is equal to or greater than 5.

6.

3. The artificial muscle according to claim 1 or 2, wherein a weight is attached to the braided tube.

Images