Bionic artificial muscle, control system and method therefor and robot

Abstract

A bionic artificial muscle, a control system and method therefor and a robot. The bionic artificial muscle comprises: wires (101), each wire (101) comprising a fixed end (1012) and a stretching end (1011), the fixed end (1012) being fixedly connected to an anchor point on a bone (3011), and the stretching end (1011) being movably connected to a connecting member (103) on the bone (3011); and a muscle part (102), the muscle part (102) being constituted by an elastic material (1021), and the elastic material (1021) wrapping the outer sides of the wires (101) between the anchor point and the connecting member (103). When the stretching ends (1011) move the wires (101) in a direction away from the connecting member (103), the elastic material (1021) of the muscle part (102) is squeezed and deformed; when the stretching ends (1011) move the wires (101) in a direction close to the connecting member (103), the elastic material (1021) of the muscle part (102) is restored to an initial state.

Description

Bionic artificial muscle and control system and method thereof, and robot TECHNICAL FIELD

[0001] The present application relates to the field of bionic muscle and robot, and particularly relates to a bionic artificial muscle and control system and method thereof, and robot. BACKGROUND

[0002] A bionic robot is a robot that imitates a living being and performs work with biological characteristics. For a long time, how to efficiently reproduce the work capacity and appearance form of the referenced object (such as a human or an animal) is a big problem in the field of bionic robots. In the traditional design mode, it is often inevitable to make a large number of adjustments or modifications on the relevant structure or form according to the function to be achieved, which not only increases the design difficulty and manufacturing cost, but also often sacrifices the universality and adaptability of the robot due to trade-off. Moreover, such customized design will lead to a large robot body and a fragile structure, and once the key parts exposed outside are damaged, the overall running ability will be easily lost, and the reliability is low. In addition, the bionic robot is difficult to achieve high similarity in appearance design with the referenced living being, and relatively lacks aesthetic sense.

[0003] Currently, the main driving methods of bionic robots include hydraulic drive and motor drive. The core principle of hydraulic drive is to generate high-pressure liquid through a liquid compression pump, and to work on the output mechanism. Due to the incompressibility of liquid, the work of the system on the liquid makes the liquid produce very high pressure. When this pressure acts on the cylinder, it will produce a strong thrust. At the same time, the liquid compression pump can be set away from the execution end, and only one is needed. The potential energy of the liquid can be transmitted to multiple remote ends through the pipeline. The output force of the hydraulic drive method is huge, and it can realize the output of super large force at very low energy cost. Moreover, the expansion cost of the actuator is relatively low, only one liquid compression pump is needed as the power source, and multiple actuators can be easily increased to maximize marginal benefits. The weight of the actuator is relatively low, which helps to reduce the inertia of the robot limb at the end, and improves the agility of the movement. Hydraulic drive is naturally suitable for low-speed execution mechanism, without the need for complex mechanical transmission devices, it can directly drive the load. This feature is particularly suitable for application scenarios such as humanoid robots that require fine control of speed. However, its disadvantages are also obvious: for example, low energy efficiency, high system complexity, involving many parts, leading to difficult control and maintenance problems such as cylinder explosion and oil leakage; at the same time, it is complex to realize high-response control of hydraulic pressure, and it needs additional independent motor servo pump and force sensor support; in addition, the overall integration of the hydraulic system is too high, and the processing precision requirement is strict, and traditional machining cutting technology is difficult to meet the demand, often need to use metal 3D printing and other high-end manufacturing technology, thus increasing the development and production cost. Therefore, although the driving method using hydraulic pressure basically meets the related ideas of bionic robots, the high design and manufacturing cost and various problems make the robot using this method currently only stay in the laboratory stage, and it is difficult to commercialize and productize.

[0004] In contrast, the core of motor-driven is mainly focused on motor type, control method and sensor technology. On the motor type, usually use DC motor or stepper motor. DC motor is suitable for applications that require high speed and variable speed control, while stepper motor is suitable for applications that require high precision positioning and controllable step angle. Control methods include open-loop control and closed-loop control. Open-loop control is based on pre-set motor input signal control, but cannot feedback adjustment to the actual state of the motor. Closed-loop control feedbacks the actual motor state through sensors to achieve more accurate control. Sensor technology includes encoder, position sensor and force / torque sensor. Encoder is used to measure the rotation angle and speed of the motor, position sensor is used to measure the position of the motor, and force / torque sensor is used to measure the force or torque applied by the motor. Motor drive completes the functional requirements of robots by selecting and debugging these components. Compared with hydraulic drive, motor drive is mature and reliable, has long service life and good robustness; natural torque loop feedback, no need for special force sensor; force response speed is extremely high, high nature servo drive can reach 10 kHz level response; energy conversion efficiency is higher; motor power density is high, which can easily achieve high speed. However, the disadvantages cannot be ignored: the torque density of motor is one order of magnitude lower than that of hydraulic actuator; because of the structure of the motor, it is difficult to realize all-directional movement at a joint, sometimes it is necessary to completely deviate from the original movement mode of the reference object and artificially design a new one, which is difficult to design and the universality is low; once major improvements are made to the structure, a lot of time will be spent on retraining the algorithm, and there is no universal reference mode; the number and types of parts are too many, the manufacturing requirements of some parts are high, it is difficult to reduce costs quickly and form an industrial scale; some parts cannot withstand impact interference and some scenes are destined to be unable to use.

[0005] Although various factors are considered, motor-driven is more suitable than hydraulic-driven in the field of robots, but due to the complex research and development, slow iteration, insufficient scale and other problems, a lot of trial and error is still needed, so it is still difficult to popularize in the short term.

[0006] In addition, there is another conceptual driving method, which is to use artificial muscles. Artificial muscles, also known as electroactive polymers (EPAs) or artificial muscles, are a new type of intelligent polymer material that can stretch, bend, tighten or expand by changing the internal structure of the material under an external electric field, which is very similar to biological muscles. Currently, there are many types of materials used for artificial muscles, such as plastic, polymer similar to rubber, gel and metal, but these materials still face many challenges in practical applications. For example, they consume a lot of energy, are prone to failure and cannot be repaired like real muscles, which makes the large-scale application of artificial muscles in robotics technology still not realized.

[0007] SUMMARY

[0008] To solve the above technical problems, the present application provides a bionic artificial muscle, a control system and method thereof, and a robot.

[0009] The technical problems solved by the present application can be implemented by the following technical solutions:

[0010] The first aspect of the present application is to provide a bionic artificial muscle, comprising:

[0011] a wire, the wire comprising a fixed end and a stretchable end, the fixed end being fixedly connected with an anchor point on a bone, and the stretchable end being movably connected with a connecting member on the bone;

[0012] a muscle part, the muscle part being made of an elastic material, the elastic material being wrapped outside the wire between the anchor point and the connecting member; wherein when the stretchable end moves the wire in a direction away from the connecting member, the elastic material of the muscle part is deformed by being squeezed; and when the stretchable end moves the wire in a direction close to the connecting member, the elastic material of the muscle part returns to the initial state.

[0013] Preferably, the wire between the anchor point and the connecting member comprises an extendable part away from the fixed end and an inextendable part close to the fixed end.

[0014] When the stretchable end moves in a direction away from the connecting member, the extendable part extends from the connecting member; and when the stretchable end moves in a direction close to the connecting member, the extendable part is retracted from the connecting member.

[0015] Preferably, the muscle part is made of the same elastic material.

[0016] Preferably, the muscle part is made of different elastic materials, and the elasticity of the elastic material gradually decreases from the middle to the two ends of the muscle part.

[0017] The second aspect of the present application is to provide a control system of a bionic artificial muscle, comprising at least one bionic artificial muscle as described above, the control method comprising:

[0018] at least one driving unit, respectively connected with the stretchable end of the wire of each bionic artificial muscle, for driving the stretchable end of the wire of the corresponding bionic artificial muscle to move the wire in a direction away from or close to the connecting member under the action of a control signal.

[0019] The control unit is connected to the at least one driving unit, and is configured to convert the deformation amount of the muscle part in the bionic artificial muscle into the corresponding wire winding / unwinding amount, and output the control signal according to the wire winding / unwinding amount.

[0020] Preferably, the robot further comprises:

[0021] The expected deformation determination unit is configured to determine the expected value of the deformation of the muscle part of each bionic artificial muscle when performing a preset action.

[0022] The deformation calculation unit is connected to the expected deformation determination unit and the control unit, respectively, and is configured to determine the deformation amount of the muscle part in each bionic artificial muscle according to the current actual value and the expected value of the muscle part of each bionic artificial muscle, and output the determined deformation amount to the control unit.

[0023] The third aspect of the present application provides a control method of a bionic artificial muscle, which is applied to the bionic artificial muscle as described above, and the control method comprises:

[0024] Step S1: converting the deformation amount of the muscle part in the bionic artificial muscle into the corresponding wire winding / unwinding amount, and outputting the control signal according to the wire winding / unwinding amount.

[0025] Step S2: driving the stretching end of the wire of the corresponding bionic artificial muscle to move the wire in the direction away from or close to the connecting piece according to the control signal.

[0026] The fourth aspect of the present application provides a robot, which comprises:

[0027] The robot body comprises a skeleton, and at least one bionic artificial muscle as described above is assembled on the skeleton.

[0028] The processor is configured to convert the deformation amount of the muscle part in the bionic artificial muscle into the corresponding wire winding / unwinding amount according to the deformation amount of the muscle part in the bionic artificial muscle when performing a preset action, and output the control signal according to the wire winding / unwinding amount.

[0029] The controller is connected to the skeleton and the bionic artificial muscle, respectively, and is configured to control the skeleton to perform a preset action, and drive the stretching end of the wire of the corresponding bionic artificial muscle to move the wire in the direction away from or close to the connecting piece according to the control signal.

[0030] Preferably, the robot further comprises:

[0031] A visual system is configured to collect observation data, wherein the observation data at least includes an environmental space where the robot is located, objects in the environmental space, and action demonstration information of a preset action that is critical when the objects perform a preset task;

[0032] The processor is further connected to the visual system, and the processor further includes:

[0033] A body dynamic determination module is configured to output all target body dynamic parameters of a target object at all given time points during completion of the preset task based on a preset body dynamic model according to the action demonstration information;

[0034] An expected deformation determination module is connected to the body dynamic determination module and is configured to determine expected values of muscle part deformations of each of the biomimetic artificial muscles when performing the preset action according to the target body dynamic parameters at each given time point;

[0035] A deformation calculation module is connected to the expected deformation determination module and is configured to determine deformation amounts of the muscle parts in each of the biomimetic artificial muscles according to current actual values and the expected values of the muscle parts of each of the biomimetic artificial muscles.

[0036] Preferably, the application further includes:

[0037] A multi-perception sensing system is configured to collect multi-perception sensing data in real time during performance of the preset action by the robot and feed the multi-perception sensing data collected in real time to the processor;

[0038] The processor is further connected to the multi-perception sensing system and is configured to adjust or re-define the retraction and release amounts according to the fed multi-perception sensing data and output the control signals according to the adjusted or re-defined retraction and release amounts.

[0039] The application has the following advantages or beneficial effects:

[0040] The biomimetic artificial muscle of the application abandons the traditional mode of relying on current control of material state changes and instead adopts a purely mechanical device principle, accurately controls the elastic material of the muscle part by precisely controlling the retraction or release of the wire, realizes the change effect on the whole level while keeping the original material size unchanged, realizes the replication of all activity modes and functions of the referenced object, greatly reduces the cost of additional artificial design and debugging, and is convenient for subsequent upgrading and maintenance; meanwhile, the robot adopting the biomimetic artificial muscle of the application can avoid the problems in the aspects of design, cost, control, and maintenance, etc., and injects new vitality and possibility into the development of robot technology. BRIEF DESCRIPTION OF DRAWINGS

[0041] Fig. 1 is a schematic diagram of a bionic artificial muscle in a preferred embodiment of the present application;

[0042] In Fig. 1, 1a and 1b are biological anatomical structures before and after elbow flexion, and 1c and 1d are bionic artificial muscles simulating the bones before and after elbow flexion.

[0043] Fig. 2 is a schematic diagram of a bionic artificial muscle without elastic material in a preferred embodiment of the present application;

[0044] In Fig. 2, 2a and 2b are schematic diagrams of wires before and after elbow flexion.

[0045] Fig. 3 is a schematic diagram of elastic material of a muscle part in a preferred embodiment of the present application;

[0046] In Fig. 3, 3a and 3b are schematic diagrams of elastic material of a muscle part before and after elbow flexion.

[0047] Fig. 4 is a structural block diagram of a control system of a bionic artificial muscle in a preferred embodiment of the present application;

[0048] Fig. 5 is a flowchart of a control method of a bionic artificial muscle in a preferred embodiment of the present application;

[0049] Fig. 6 is a structural block diagram of a robot in a preferred embodiment of the present application. DETAILED DESCRIPTION

[0050] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0051] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict.

[0052] The present application will be further described below with reference to the drawings and specific embodiments, but is not limited by the embodiments.

[0053] Referring to Figs. 1, 2 and 3, in a preferred embodiment of the present application, based on the above-mentioned problems existing in the prior art, a bionic artificial muscle is provided, which comprises:

[0054] The wire 101 comprises a fixed end 1012 and a stretching end 1011. The fixed end 1012 is fixedly connected with an anchor point (not shown in the figure) on the bone 3011, and the stretching end 1011 is movably connected with the connecting piece 103 on the bone 3011.

[0055] The muscle part 102 is composed of an elastic material 1021 wrapped outside the wire 101 between the anchor point and the connecting piece 103; wherein the elastic material 1021 of the muscle part 102 is extruded and deformed when the stretching end 1011 moves the wire 101 in the direction away from the connecting piece 103; and the elastic material 1021 of the muscle part 102 returns to the initial state when the stretching end 1011 moves the wire 101 in the direction close to the connecting piece 103.

[0056] Specifically, considering that the existing artificial muscle changes the internal structure of the material through an external electric field to achieve the effects of stretching, bending, tightening or swelling, there are problems such as high energy consumption, easy failure and inability to self-repair.

[0057] The bionic artificial muscle proposed in the embodiment mainly consists of a wire 101 and a muscle part 102, and the muscle part 102 is composed of an elastic material 1021. The material has good elasticity and recovery, and can tightly wrap the wire 101 between the anchor point and the connecting piece 103. Through simple mechanical operation, the wire can be bundled or released, thereby realizing the simulation movement of the contraction and relaxation function of the real muscle, and avoiding the problems such as high energy consumption, easy failure and inability to self-repair caused by the change of the internal structure of the material controlled by the external electric field in the traditional artificial muscle.

[0058] In the embodiment, the wire is made of linear material. The linear material has sufficient strength to ensure that it can withstand the tension from all directions without being pulled apart during the simulation of muscle contraction and relaxation. The selection of the linear material can be based on the specific application scene and demand, and the most suitable linear material can be selected, which is not limited to a certain specific material. Under the premise of meeting the strength requirement, the linear material with low cost, easy processing and acquisition can be selected, which helps to reduce the overall manufacturing cost of the bionic artificial muscle and improve its market competitiveness.

[0059] In the embodiment, the anchor point is mainly used to fix the position of the artificial muscle or other related parts on the bone. The position of the anchor point is selected based on the anatomical structure of the bone and the movement demand of the artificial muscle, and is usually located in the area that can provide the best support and stability on the bone. The anchor point provides a stable connection point to ensure that the artificial muscle does not fall off or shift from the bone during movement. The anchor point usually has sufficient strength and rigidity to withstand the tension and stress generated by the artificial muscle during work, ensuring the stability of the entire system.

[0060] In this embodiment, the connecting piece is used to connect the wire wrapped with elastic material of the muscle part and the anchor point, forming a complete system to ensure that the wire can be bundled or released from the connecting piece. The anchor point and the connecting piece are complementary. The anchor point provides a fixed point on the bone, and the connecting piece connects these fixed points with the wire or artificial muscle.

[0061] Preferably, the connecting piece itself can also act as an anchor point. The connecting piece is arranged at the connection between the muscle and the bone, and is directly fixed on the bone and connected with the artificial muscle or wire.

[0062] In this embodiment, the connecting piece 103 can adopt a limiting bayonet design. By setting the limiting bayonet, the wire 101 is limited. When a part of the middle wire 101 (i.e. the extendable part 1013) is pulled out, the elastic material wrapped on the wire 101 cannot pass through the bayonet, and the elastic material that cannot pass through the bayonet will be squeezed, thereby forming a change effect similar to "muscle bulge".

[0063] Further, the connecting piece can also contain sensors or other monitoring devices to real-time feedback the state and performance of the artificial muscle.

[0064] As a preferred embodiment, as shown in FIG. 2, the wire 101 between the anchor point and the connecting piece 103 includes: an extendable part 1013 away from the fixed end 1012 and an unextendable part 1014 close to the fixed end 1012.

[0065] When the stretching end 1011 moves away from the connecting piece 103, the extendable part 1013 extends from the connecting piece 103; and when the stretching end 1011 moves close to the connecting piece 103, the extendable part 1013 is retracted from the connecting piece 103.

[0066] Specifically, the wire 101 between the anchor point and the connecting piece 103 can be divided into an extendable part 1013 and an unextendable part 1014. When the wire 101 is bundled, i.e. the stretching end 1011 moves away from the connecting piece 103, the extendable part 1013 of the wire 101 extends, and the elastic material 1021 of the muscle part 102 cannot pass through the connecting piece (i.e. the bayonet) and is squeezed to deform, thereby simulating the bulging change of the muscle.

[0067] Conversely, when the wire 101 is released, i.e. the stretching end 1011 moves close to the connecting piece 103, the extendable part 1013 is retracted, and the elastic material 1021 gradually returns to the initial state, simulating the relaxation process of the muscle.

[0068] Further, different colors are used in Figure 2 to visually distinguish the various parts of the wire 101. Blue is used to indicate the non-extendable part 1014, i.e. the part of the wire that cannot be pulled out; red is used to indicate the extendable part 1013, i.e. the part of the wire that can be pulled out; orange is used to indicate the part of the wire where the stretching end 1011 is located, which is responsible for performing the stretching or winding action, thereby achieving the squeezing and releasing of the muscle part 102, simulating the process of muscle bulging and relaxation.

[0069] Further, the elastic material 1021 can be a single material or a composite material. Depending on specific needs, materials with different properties are selected to meet specific effects.

[0070] As a preferred embodiment, the muscle part 102 is composed of the same elastic material 1021.

[0071] Specifically, the elastic material 1021 can be a single material, and the muscle part 102 is uniformly composed of the same elastic material 1021 to simplify the manufacturing process and ensure the consistency of overall elasticity.

[0072] As a preferred embodiment, the muscle part 102 is composed of different elastic materials 1021, as shown in Figure 3.

[0073] Specifically, in order to improve the simulation effect, the elastic material 1021 can be composite. The muscle part 102 can be designed in segments or layers using a variety of elastic materials 1021, and through the combination of different materials, a more delicate and realistic muscle change effect can be achieved. Depending on specific needs, materials with different elastic coefficients are selected to achieve a more realistic muscle change effect.

[0074] As a preferred embodiment, the elastic material 1021 used in the muscle part 102 gradually decreases in elasticity from the middle to the two ends.

[0075] Specifically, in order to further improve the accuracy of simulation, an elastic material 1021 with an elastic coefficient that gradually decreases from the middle to the two ends can be used. This design can more accurately simulate the elastic differences exhibited by real muscles in different areas, thereby achieving higher simulation in both appearance and function. The flexibility of this material selection not only gives the bionic artificial muscle a realistic appearance, but also enables it to achieve high simulation in function, ensuring that the bionic artificial muscle can achieve optimal performance matching and effect presentation for different application scenarios, providing strong support for the development of bionic robots and other related fields.

[0076] Further, in this embodiment, one bionic artificial muscle is composed of a muscle part 102 composed of a wire 101 and an elastic material 1021 tightly wrapped outside. In order to enhance the flexibility and adaptability of the system, one or more bionic artificial muscles are combined into a group to form a group of bionic artificial muscles.

[0077] In order to meet the needs of different application scenarios, the present application introduces the concept of "bionic artificial muscle" system. The system is composed of one or more groups of bionic artificial muscles, and by adjusting the number, layout and control mode of the muscles, the system can accurately reproduce the muscle system of real organisms, realize various motion forms from simple to complex and from static to dynamic, and bring revolutionary breakthroughs to bionic robots, medical rehabilitation, motion simulation and other fields.

[0078] Further, the limbs composed of one or more muscles can be further wrapped with a layer of "skin" to maintain integrity and facilitate force transmission. The structures between the muscles can also be designed or the corresponding materials can be used to reduce friction and facilitate independent movement.

[0079] The "bionic artificial muscle" system of the present application replaces the traditional hydraulic drive or motor drive for controlling the limb part, thereby maximizing the precise control of each group or each "muscle" to reproduce as much as possible the entire activity mode and function of the reference object, while greatly reducing the cost of additional artificial design and debugging. The "bionic artificial muscle" system does not need to drive the structure connected to it by controlling the shape change of the material itself like a real artificial muscle, but by controlling the contraction or release of the linear material to achieve the overall change effect without changing the size of the original material.

[0080] Further, the "bionic artificial muscle" system of the present application can be regarded as a purely mechanical device, and the stretching control of the wire end can be realized by the existing driving mode. Preferably, by concentrating the protection of each part of the existing driving mode, the control is facilitated; at the same time, by quickly replacing the peripheral inexpensive components, subsequent upgrading and maintenance are facilitated.

[0081] As an example, but not limited, take the human arm as an example: one or more groups of linear materials can be used, and an elastic material such as rubber 1021 can be flexibly wrapped outside the wire 101 to replace the functions of the biceps and triceps in the arm. By bundling one end or middle part of the wire 101 to the inside of the chest or bone, when the middle wire 101 is bundled, the simulated muscle spacing will be shortened accordingly, and the elastic material 1021 will be squeezed, thereby reproducing the muscle bulge change, achieving the effect of simulating muscle contraction. At the same time of reproducing the muscle bulge change, through the cooperation of each muscle, the force is facilitated in the whole structure to achieve a more outstanding effect.

[0082] Referring to FIG. 4, the second aspect of the present application is to provide a control system 200 of the bionic artificial muscle, including at least one bionic artificial muscle (1001, 100n) as described above, and the control method includes:

[0083] At least one driving unit (2011, 201n) is connected to the stretching end 1011 of the wire 101 of each bionic artificial muscle (1001, 100n), respectively, for driving the stretching end 1011 of the wire 101 of the corresponding bionic artificial muscle (1001, 100n) to move the wire 101 in the direction away from or close to the connecting piece 103 under the action of the control signal;

[0084] The control unit 202 is connected to the at least one driving unit (2011, 201n), for converting the deformation amount of the muscle part 102 in the bionic artificial muscle (1001, 100n) into the amount of winding and unwinding of the corresponding wire 101, and outputting the control signal according to the amount of winding and unwinding.

[0085] Specifically, the control system 200 of the embodiment of the present application mainly consists of a driving unit, a control unit and a plurality of bionic artificial muscles, and through the cooperative work of the components, the operation of one or more bionic artificial muscles in the above-mentioned “bionic artificial muscle” system can be accurately controlled, and the high-efficiency and precise muscle simulation effect can be realized to simulate the complex dynamic behavior of the real biological muscle.

[0086] The bionic artificial muscle is equipped with one or more, and each bionic artificial muscle (1001 to 100n) contains a wire 101, a muscle part 102 and a connecting piece 103 and other key components. The wire 101 adopts linear material, and the stretching end 1011 is used for receiving power input from the driving unit to realize winding or releasing of the wire. The muscle part 102 is tightly wrapped outside the wire 101 and is composed of an elastic material 1021, which can deform with the stretching and contraction of the wire 101 to simulate the bulging and relaxing of the muscle.

[0087] The number of driving units corresponds to the number of bionic artificial muscles. Each driving unit (e.g. 2011) is connected to the tensile end 1011 of the wire 101 of one bionic artificial muscle (e.g. 1001). The driving unit (e.g. 2011) controls the extension and retraction of the wire 101 of each bionic artificial muscle (e.g. 1001). Preferably, the driving units (2011,..., 201n) are controlled by conventional hydraulic or motor-driven units.

[0088] The control unit 202 is responsible for calculating the required wire 101 extension and retraction amount according to the deformation data and the current deformation amount, and sending accurate control signals to each driving unit to ensure that the bionic artificial muscle can perform extension and retraction movement according to the predetermined mode.

[0089] As a preferred embodiment, it further comprises:

[0090] The desired deformation determination unit 203 is used to determine the desired value (i.e. target deformation amount) of the muscle part 102 of each bionic artificial muscle (1001,..., 100n) after deformation when performing a preset action;

[0091] The deformation calculation unit 204 is connected to the desired deformation determination unit 204 and the control unit 202, respectively, and is used to determine the deformation amount of the muscle part 102 of each bionic artificial muscle (1001,..., 100n) according to the current actual value (i.e. current deformation amount) and the desired value of the muscle part 102 of each bionic artificial muscle (1001,..., 100n), and output the determined deformation amount to the control unit.

[0092] In this embodiment, the deformation amount or deformation data can be the difference between the current deformation amount and the desired target deformation amount of the muscle part 102. This deformation difference represents the degree of further deformation required by the muscle part.

[0093] The current deformation amount can be obtained by real-time monitoring of sensors. The sensors can be displacement sensors, strain sensors or other devices that can accurately measure deformation.

[0094] The target deformation amount refers to the specific deformation amount required by the muscle part 102 of the bionic artificial muscle to achieve the desired shape or position when performing a specific action. This deformation amount can be calculated based on preset action requirements, biomechanical models, muscle anatomy characteristics or experimental data, etc.

[0095] Further, it is desirable to preset an artificial intelligence (AI) model in the deformation determination unit 203. The target deformation amount can also be achieved through the preset artificial intelligence (AI) model. The AI model can adopt existing limb action deduction models, including but not limited to the first and second adversarial network models proposed in the prior application patent CN202410136427.8. Through the limb action deduction model, the action to be performed and the body dynamic parameters of all muscle blocks when performing the action are deduced; then, by comparing the detailed muscle dissection data in the preset database, all muscle blocks to be used in the execution of the action are determined; then, according to the body dynamic parameters of all muscle blocks when performing the action, the deformation amount of all muscle blocks used is measured, thereby finally converting the wire 101 into the amount of extension and retraction.

[0096] Further, the mechanical model of the bionic artificial muscle can be used in the control unit 202 to convert the deformation difference into the amount of extension and retraction of the wire 101, and then generate a corresponding control signal to drive the driving unit (such as a motor, a hydraulic cylinder, etc.) to perform the extension and retraction action of the wire 101.

[0097] The mechanical model can be obtained based on experimental data, theoretical derivation or simulation. In the conversion process, factors such as the elastic modulus of the muscle part material, the geometric shape, the coupling relationship between the wire and the muscle part, etc. need to be considered. If the system needs to handle dynamic actions or fast response, the dynamic effects such as inertia, damping and friction during the extension and retraction of the wire 101 also need to be considered.

[0098] Further, during the driving process, the system will monitor the actual deformation of the muscle part in real time through the sensor and compare it with the target deformation amount. If there is a deviation, the system will feedback and adjust the control signal to further reduce the deviation until it reaches or approaches the target deformation amount.

[0099] It should be noted that, since the mechanical properties and behaviors of the bionic artificial muscle can be affected by various factors (such as temperature, humidity, material aging, etc.), in actual application, the above calculation process may need to be modified and optimized to ensure the accuracy and stability of the system.

[0100] Referring to FIG. 5, the third aspect of the present application is to provide a control method of a bionic artificial muscle, characterized in that it is applied to the bionic artificial muscle as described above, and the control method comprises:

[0101] Step S1, converting the deformation amount of the muscle part 102 in the bionic artificial muscle (1001,..., 100n) into the amount of extension and retraction of the corresponding wire 101, and outputting a control signal according to the amount of extension and retraction;

[0102] Step S2, according to the control signal, drive the corresponding bionic artificial muscle (1001,..., 100n) wire 101 of the stretching end 1011 along the direction away from or close to the connector 103 to move the wire 101.

[0103] Specifically, first, the actual deformation amount of the muscle part 102 is monitored by the sensor, and compared with the target deformation amount; based on the comparison result, the control unit calculates the amount of winding and unwinding that the wire 101 should perform, and generates a corresponding control signal to send to the corresponding drive unit, such as a motor, a hydraulic cylinder, etc., to drive the stretching end 1011 of the wire 101 to move, so that the wire 101 is wound or released, and then drives the muscle part 102 to deform accordingly, to simulate the contraction and relaxation process of the real muscle.

[0104] Further, in step S2, it also includes:

[0105] determining the expected value (i.e. target deformation amount) of the muscle part 102 of each bionic artificial muscle (1001,..., 100n) after deformation when performing the preset action;

[0106] According to the current actual value (i.e. current deformation amount) and the expected value of the muscle part 102 of each bionic artificial muscle (1001,..., 100n), determine the deformation amount of the muscle part 102 in each bionic artificial muscle (1001,..., 100n);

[0107] Then, according to the determined deformation amount, convert it into the winding and unwinding amount of the corresponding wire 101, and output the control signal according to the winding and unwinding amount.

[0108] Further, the target deformation amount can be obtained by the above-mentioned pre-set AI model (such as the limb action deduction model).

[0109] Referring to FIG. 6, the fourth aspect of the present application provides a robot 300, comprising:

[0110] The machine body 301 comprises a skeleton 3011, and at least one bionic artificial muscle (1001,..., 100n) as described above is assembled on the skeleton 3011;

[0111] The processor 302 is used to convert the deformation amount of the muscle part 102 in the bionic artificial muscle into the winding and unwinding amount of the corresponding wire 101 according to the deformation amount of the muscle part 102 in the bionic artificial muscle when performing the preset action, and output the control signal according to the winding and unwinding amount;

[0112] The controller 303 is connected to the skeleton 3011 and the bionic artificial muscle (1001, 100n) respectively, and is used to control the skeleton 3011 to perform a preset action; meanwhile, the controller 303 drives the stretching end 1011 of the wire 101 of the corresponding bionic artificial muscle (1001, 100n) to move the wire 101 in a direction away from or close to the connecting piece 103 according to a control signal.

[0113] Specifically, the embodiment of the present application provides a novel robot, which is attached with the above-mentioned bionic artificial muscle on a skeleton structure of the robot, and the robot realizes complex actions and tasks by utilizing the deformation of the above-mentioned bionic artificial muscle, so as to avoid the problems in the aspects of design, cost, control, maintenance and the like in the prior art.

[0114] As a preferred implementation, the robot further comprises:

[0115] The vision system 304 is used to collect observation data; wherein the observation data at least includes an environment space where the robot 300 is located, objects in the environment space and action demonstration information of a key preset action when the objects perform a preset task;

[0116] The processor 302 is further connected to the vision system 304, and the processor 302 further comprises:

[0117] The body dynamic determination module 3021 is used to output all target body dynamic parameters of a target object at all given time points in the process of completing a preset task based on a preset body dynamic model according to the action demonstration information;

[0118] The expected deformation determination module 3022 is connected to the body dynamic determination module 3021, and is used to determine an expected value of the muscle part 102 of each bionic artificial muscle after deformation when performing a preset action according to the target body dynamic parameters at each given time point;

[0119] The deformation calculation module 3023 is connected to the expected deformation determination module 3022, and is used to determine a deformation amount of the muscle part 102 in each bionic artificial muscle according to a current actual value and an expected value of the muscle part 102 of each bionic artificial muscle.

[0120] Specifically, the preset body dynamic model is the above-mentioned AI model (such as a limb action deduction model). The limb action deduction model is used to determine a change amount of a muscle when performing a related action, so as to control the winding and unwinding operation of the wire 101.

[0121] In practical applications, the robot first analyzes the current environment through the vision system, formulates the purpose of the behavior; then searches for a behavior template that roughly matches the current task in the built-in behavior library, and intelligently corrects the behavior parameters in the template according to the actual environment, for example, the template in the library only walks 5 meters, but the actual application needs to walk 10 meters, so one or more parameters need to be corrected according to the actual situation.

[0122] During task execution, the robot continuously collects real-time feedback data from various sensors (such as skin pressure sensors and overall horizontal center of gravity sensors, and other types as needed), which are sent to the processor for comprehensive analysis to further optimize muscle control strategies under each action, ensuring that the robot can accurately and stably complete the preset task.

[0123] After the task is successfully executed, the robot returns the relevant data and experience accumulated during the execution process to the behavior library for reference and optimization for subsequent tasks. In addition, by using advanced algorithms such as reinforcement learning, the robot can also continuously optimize its AI model and algorithm, improving the overall task execution efficiency and intelligence level.

[0124] As a preferred embodiment, it further comprises:

[0125] The multi-sensing sensing system 305 is used to collect multi-sensing sensing data during the execution of the preset action, and feed back the multi-sensing sensing data collected in real time to the processor;

[0126] The processor 302 is also connected to the multi-sensing sensing system 305, for adjusting or re-formulating the amount of collection and release according to the feedback multi-sensing sensing data, and outputting a control signal according to the adjusted or re-formulated amount of collection and release.

[0127] Specifically, considering that the vision system 304 may have a blind area in environmental perception, a multi-sensing sensing system 305 can also be provided to use multiple sensing methods for optimization to achieve comprehensive and high-precision monitoring of the surrounding environment.

[0128] When the multi-sensing sensing system 305 detects a physical environmental change related to the execution of the preset task, the real-time data generated will be given higher priority, taking precedence over the pre-execution action formulated based on the vision system or other information sources, ensuring that the robot can fully evaluate the actual status of the current environment before taking action, ensuring that the action taken meets the task requirements and does not cause any damage to the robot itself, improving its adaptability and safety in complex and variable environments.

[0129] Further, the bionic artificial muscle of the application can create a highly simulated robot, and on this basis, the elastic material of the muscle part can also be selected from some materials with strength far exceeding human muscles, so that the limb segment is made more slender under the premise of maintaining the function level of the robot, the appearance aesthetics of the robot is improved, and sufficient space is provided for installing more solid armor. The installation of the armor not only enhances the physical protection ability of the robot, but also enables the robot to cope with more complex and harsh working environments, thereby expanding the application field of the robot.

[0130] In the above preferred embodiment, the bionic artificial muscle of the application can also be applied to the field of exoskeleton. The exoskeleton is essentially a wearable robot (such as two humanoid robots closely attached to a person in front and behind), which can move following the movement of the human body. In this embodiment, by attaching corresponding muscles, i.e. bionic artificial muscles, to the key parts of the exoskeleton, the exoskeleton and the attached bionic artificial muscles will move flexibly and closely following the movement of the human body when the human body moves, avoiding the inconvenience factors such as exposed hydraulic pipes and motor wheels commonly seen in traditional exoskeletons, and improving the overall aesthetics and practicality.

[0131] The mounting of the robot equipment will be borne by the skeletal structure of the robot, and the exoskeleton can bear heavier equipment mounting by using the self-supporting and high-strength characteristics of the bionic artificial muscle. In addition, various protective equipment can be installed on the exoskeleton to further improve the safety protection ability of the wearer.

[0132] In the above preferred embodiment, the method for controlling the movement of the robot limb comprises the following steps:

[0133] I. Robot hardware design:

[0134] Firstly, determine the bionic object. The bionic object can be a number of organisms in nature and their characteristics, such as humans or animals. For objects that rely on fluid motion, high-performance fluid power devices can be integrated into the design to simulate and enhance their natural motion capabilities.

[0135] Then, according to the anatomical structure of the object, design the various components of the robot, such as the skeleton, the "bionic artificial muscle" system, and the driving and control unit. The load hanging point can also be provided on the skeletal structure to facilitate the subsequent installation of armor or functional modules, thereby improving the overall protection and functionality of the robot.

[0136] II. Model building:

[0137] Firstly, conduct comprehensive motion capture on the bionic object to record the muscle movement pattern and spatial positioning information of the bionic object.

[0138] Subsequently, in a computer simulation environment, the designed robot model is used to reproduce the action, and the extension value change of each wire 101 during the action execution process is recorded in detail. Among them, the reproduced action is determined by motion capture, for example, what action is captured, the robot executes what action, and records the action that will appear under a series of extension value changes of the wire 101.

[0139] Subsequently, after assembling the bionic muscle part on the skeleton structure of the robot, the action is simulated by controlling the extension of the wire 101, and the force value change of each wire 101 during the action execution process is monitored in real time by using a force meter at the control end of the driving unit, so as to accurately calculate the control parameters of the driving unit.

[0140] Then, these motion capture results are used as a database to build a corresponding AI model, such as the body dynamic model (such as the body action deduction model) described above, which uses a stacked adversarial network to formulate a body expression, and the motion capture results are used as real samples.

[0141] Three, task execution

[0142] Firstly, the robot's visual system, such as laser radar, optical perspective system, etc., is used to construct the environment space, and the three-dimensional space and the approximate object are determined;

[0143] Then, the task target is determined, and the corresponding task action required by the real person demonstration is performed; among them, multiple demonstrations can be repeated during the demonstration process to ensure the accuracy and repeatability of the demonstration, and to provide high-quality samples for repeated learning of the robot.

[0144] Next, the key actions in the real person demonstration are screened and analyzed by the robot's visual system to obtain their body features in space; among them, these key actions are screened, and the screenshot image can be used as the input of AI, that is, the body feature information of the original picture in stage1 of the first adversarial network model (SeqGAN) of the AI model.

[0145] Next, search for related behaviors in the database as a reference, and refine the body feature information of the key actions to be executed according to the current task requirements. Among them, the searched reference can be similar behaviors or a combination of multiple behaviors, and the refined body information is the local adversarial of stage1.

[0146] Next, a complete plan from start to task completion is designed, and an action behavior highly similar to the real person demonstration action is simulated. Among them, the current body dynamic parameters of the robot are added to the front end of the output of stage1, and the global adversarial of stage2 is performed, which can be optimized multiple times.

[0147] Then, during the task execution process, the feedback data of various sensors are continuously collected, including but not limited to environmental perception, action execution state and the like, and are corrected in real time to achieve the target effect. For example, once the sensor receives a danger signal, the state of the robot is immediately adjusted in priority to ensure safety, and then the plan is adjusted or redefined as needed.

[0148] Finally, the data of the successfully completed task is incorporated into the database; at the same time, a reasonable reward mechanism is set by using reinforcement learning technology, and the AI algorithm and control process are continuously optimized, so that the robot can continuously optimize itself and improve the task execution efficiency and success rate.

[0149] The robot of the present application only needs to make moderate and subtle adjustments on the anatomical structure during design, so as to realize diversified design and aesthetic improvement, and simplify the design complexity; the robot has high bionic characteristics, has many large-angle joint structures, and gives the robot more excellent degrees of freedom and flexibility, and simulates biological movement more naturally and smoothly. In the robot mounting design, only protection and specific functions need to be considered for external mounting, and important key motion structures do not need to be externally mounted, which theoretically enhances the load-bearing capacity.

[0150] After the design of the robot is changed, the personalized control system of the robot does not need to be designed by humans, but directly uses biological patterns by means of motion capture technology and AI algorithm model, which reduces the burden of human design and is more efficient; in addition, the built-in reinforcement learning mechanism enables the robot to continuously optimize itself to adapt to complex and variable task environments.

[0151] The advantages or beneficial effects of the above technical solutions are that the bionic artificial muscle of the present application discards the traditional way of relying on current control material state change, and instead adopts pure mechanical device principles, accurately controls the elastic material of the muscle part by precisely controlling the contraction or release of the wire, realizes the change effect on the whole level under the condition that the original material size is unchanged, realizes the replication of all activity patterns and functions of the referenced object, greatly reduces the cost of additional human design and debugging, and is convenient for subsequent upgrade maintenance; at the same time, the robot using the bionic artificial muscle of the present application can avoid the problems in the aspects of design, cost, control and maintenance, and injects new vitality and possibility into the development of robot technology.

[0152] The above is only the preferred embodiment of the present application, and does not limit the implementation and protection scope of the present application. For those skilled in the art, it should be realized that any equivalent replacement and obvious change made by applying the contents of the present application and the drawings should be included in the protection scope of the present application.

Claims

1. A biomimetic artificial muscle, characterized by, Comprising: a wire, the wire comprising a fixed end and a stretching end, the fixed end being fixedly connected with an anchor point on a bone, the stretching end being movably connected with a connecting member on the bone; a muscle part, the muscle part being made of elastic material, the elastic material being wrapped outside the wire between the anchor point and the connecting member; wherein, when the stretching end moves the wire in a direction away from the connecting member, the elastic material of the muscle part is squeezed and deformed; and when the stretching end moves the wire in a direction towards the connecting member, the elastic material of the muscle part returns to the initial state.

2. The biomimetic artificial muscle of claim 1, wherein, The wire between the anchor point and the connecting member comprises: an extendable part away from the fixed end and an inextendable part close to the fixed end; When the stretching end moves in a direction away from the connecting member, the extendable part extends from the connecting member; and when the stretching end moves in a direction towards the connecting member, the extendable part is retracted from the connecting member.

3. The biomimetic artificial muscle of claim 1, wherein, The muscle part is made of the same elastic material.

4. The biomimetic artificial muscle of claim 1, wherein, The muscle part is made of different elastic materials, and the elasticity of the elastic material gradually decreases from the middle to the two ends of the muscle part.

5. A control system for a biomimetic artificial muscle, characterized by The control method comprises: at least one driving unit, respectively connected with the stretching end of the wire of each bionic artificial muscle, for driving the stretching end of the wire of the corresponding bionic artificial muscle to move the wire in a direction away from or towards the connecting member under the action of a control signal; a control unit connected with the at least one driving unit, for converting the deformation amount of the muscle part in the bionic artificial muscle into the extension and retraction amount of the corresponding wire according to the deformation amount of the muscle part, and outputting the control signal according to the extension and retraction amount. Further comprising:

6. The control system of the biomimetic artificial muscle according to claim 5, wherein, an expected deformation determination unit for determining the expected value of the deformation of the muscle part of each bionic artificial muscle when performing a preset action; a deformation calculation unit respectively connected with the expected deformation determination unit and the control unit, for determining the deformation amount of the muscle part in each bionic artificial muscle according to the current actual value and the expected value of the muscle part of each bionic artificial muscle, and outputting the determined deformation amount to the control unit. The control method comprises:

7. A control method of a biomimetic artificial muscle, characterized by, Step S1, converting the deformation amount of the muscle part in the bionic artificial muscle into the extension and retraction amount of the corresponding wire according to the deformation amount of the muscle part, and outputting a control signal according to the extension and retraction amount; Step S2, driving the stretching end of the wire of the corresponding bionic artificial muscle to move the wire in a direction away from or towards the connecting member according to the control signal. Comprising:

8. A robot, characterized in that a machine body, the machine body comprising a skeleton, the skeleton being assembled with at least one bionic artificial muscle according to any one of claims 1-4; a processor for converting the deformation amount of the muscle part in the bionic artificial muscle into the extension and retraction amount of the corresponding wire according to the deformation amount of the muscle part when performing a preset action, and outputting a control signal according to the extension and retraction amount. ​ A controller is connected to the skeleton and the bionic artificial muscle respectively, and used to control the skeleton to perform a preset action; meanwhile, according to the control signal, the stretching end of the wire of the corresponding bionic artificial muscle is driven to move the wire in the direction away from or close to the connecting piece.

9. The robot of claim 8, wherein, Further comprising: a visual system used to collect observation data; wherein the observation data at least includes an environmental space where the robot is located, objects in the environmental space, and action demonstration information of a preset action which is critical when the objects perform a preset task; the processor is further connected to the visual system, and the processor further comprises: a body dynamic determination module used to output all target body dynamic parameters of a target object at all given time points in the process of completing the preset task based on a preset body dynamic model according to the action demonstration information; an expected deformation determination module connected to the body dynamic determination module, and used to determine expected values of muscle part deformations of each bionic artificial muscle when performing the preset action according to the target body dynamic parameters at each given time point; a deformation calculation module connected to the expected deformation determination module, and used to determine deformation amounts of the muscle parts of each bionic artificial muscle according to current actual values and the expected values of the muscle parts of each bionic artificial muscle.

10. The robot of claim 8, wherein, Further comprising: a multi-perception sensing system used to collect multi-perception sensing data in real time during the robot performing the preset action, and feed back the multi-perception sensing data collected in real time to the processor; the processor is further connected to the multi-perception sensing system, and used to adjust or re-define the retraction and extension amounts according to the feedback multi-perception sensing data, and output the control signal according to the adjusted or re-defined retraction and extension amounts.

Images