A bionic biped robot and its stable gait planning method
By adopting a stable gait planning method combining buffered hoof foot and Poincaré mapping in bipedal robots, the problems of large walking resistance and poor stability in the existing technology are solved, faster and more stable movement is achieved, and the robot is improved exercise efficiency and robustness.
Patent Information
- Application Number
- CN202210397792.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-04-15
- Publication Date
- 2024-12-31
- Estimated Expiration
- 2042-04-15
AI Technical Summary
Existing bipedal robots have great resistance, limited speed, and poor stability when walking. They lack hoof-line bionic design, which affects movement efficiency and robustness.
A bionic bipedal robot was designed, using a stable gait planning method combining buffered hoof foot structure and Poincaré mapping. The motion state was simulated through the spring inverted pendulum model, optimized the layout and structure of the drive motor, and used lightweight carbon fiber material and buffering. The device shock absorption and designs a walking stability evaluation model based on the attractive domain characteristics.
It achieves faster motion speed and lighter structure, improves motion stability and robustness, reduces energy loss and shaking, and improves energy efficiency and gait performance.
Smart Images

Figure CN114872814B_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the field of biped robots, and in particular relates to a bionic biped robot and a stable gait planning method thereof. Background Art
[0002] All terrestrial higher animals in nature have three main ways of walking on their hind limbs: plantigrade (walking on the soles of the feet), digitigrade (walking on the toes), and ungulate (walking on the toenails).
[0003] Plantar walking means walking on the soles of the feet. This way of walking requires the metatarsals to touch the ground, forming a flat sole. When walking, the toes, metatarsals, and tarsals will touch the ground in turn. Humans, monkeys, gorillas, bears, rabbits, red pandas, and rodents such as mice and rats all walk in this way. This way of walking has too many parts in the part where the foot contacts the ground, and the load on the distal end of the appendage is too heavy, resulting in greater resistance when walking and running very slowly. But it also has many advantages. First of all, it has strong stability. The flat soles of the feet make the lower body more stable. The two hind legs are enough to support the weight of the entire body, and it is quite stable when walking.
[0004] Toe-walking means walking on the toes. The metatarsals of toe-walking animals are not turned into the instep of the foot, but are lifted up, and only a few toes (phalanges) are used to support the body and walk, similar to humans walking in high heels, and the heels are particularly high. Compared with plantigrade animals, toe-walking animals have another huge advantage due to the lifting of metatarsals: long legs and large strides (metatarsals are lifted, upgraded, and become part of the leg: a thin and long new leg). Because only toes are used for standing, the "foot" structure is simpler and lighter, with a smaller contact area with the ground, less resistance, and a larger stride. Toe-walking animals run more lightly and move faster than plantigrade animals. For example, a cheetah can run at a speed of up to 115 kilometers per hour.
[0005] Hoofwalking is the walking style of the odd-toed ungulate and even-toed ungulate mammals, such as sheep, cattle and pigs. From the perspective of bones, hoofwalking is more special. Even the metatarsals no longer touch the ground, but the "toenails" touch the ground. In order to meet this need, the toenails of hoofwalkers have become strong and thick during the evolution, forming what we call hooves. Although walking on tiptoes (toewalking) is faster than walking on the soles of the feet (plantarwalking), they are still inferior to hoofwalkers who walk on toenails in long-distance racing. Hoofwalking is not only faster, but also has longer legs. This structure that maximizes the transformation of all parts below the buttocks into legs makes a typical hoofed herbivore (such as horses, deer, etc.) often have a plump and strong buttocks and light and tough lower limbs, high burst and high endurance, and is simply born for sports. The maximum speed of gazelles can reach 100 kilometers per hour, and they have super endurance.
[0006] like Figure 1 As shown in the figure, the hind leg structure of hoofed animals, such as deer, is that the protruding joints are tarsal bones, the hoof is a specialized toenail, and the hoof is a small section that gradually becomes thinner (highly fused metatarsal bones), followed by thicker joints, and then a long "calf" (metatarsal bones) above. This structure is used in the bionic design of the legs of bipedal robots, that is, to design a hoofed bipedal robot, which provides design ideas for faster robot movement.
[0007] In the field of bipedal robots, the prior art includes humanoid robots (plantigrade) and ostrich-like robots (toe-walking), but no hoof-walking robots. Patent 110254553A discloses a full-morph ostrich-like high-speed running and jumping robot, which adopts a toe-walking structure. Patent CN109606492A discloses a bipedal jumping robot based on a ducted propulsion system. The robot's jumping legs are equipped with servo motors for controlling movement at the ankle, knee and hip joints, and the feet use a large contact area sole (plantigrade). Although this robot has achieved large-scale jumps and smooth landing on complex terrain, there are too many parts in the part where the foot contacts the ground, the load at the distal end of the robot's appendages is too heavy, and the robot's own weight is large, resulting in greater resistance when walking and limited running speed. CN215155133U discloses a bipedal robot, the legs of which also adopt a parallelogram series structure, but only the leg drive parts are set at the thigh root (hip joint), and the feet adopt a wheeled structure, and a flywheel located at the waist with a rotation axis perpendicular to the legs is used to maintain the robot's movement balance. This structure only sets one motor drive at the hip joint, which is not conducive to force buffering. The robot shakes and trembles greatly when walking, and the power loss is large, which is not conducive to efficiency improvement. Patent CN111516773A discloses a reconfigurable bipedal robot with multiple movement modes, which can realize the switching between bipedal and quadrupedal movement modes, but still adopts a plantigrade bipedal structure.
[0008] In short, hoof-walking bionic bipedal robots are still a technological blank. Summary of the invention
[0009] In view of the shortcomings of the prior art, the present invention provides a bionic bipedal robot and a stable gait planning method thereof; to make up for the gaps in the prior art, while having better force output, faster movement speed, and a lighter structure, its motion state is simulated through a spring inverted pendulum model to avoid its poor stability problem; at the same time, based on the Poincare mapping and combined with the characteristics of the attraction domain, a "limit cycle walking" periodic walking gait stability evaluation model is designed to ensure its robustness in the motion state.
[0010] To achieve the above object, the present invention provides the following technical solutions:
[0011] A bionic biped robot has two mechanical legs and is a bilaterally symmetrical structure, characterized in that the mechanical leg comprises a thigh, a calf and a buffer hoof; wherein the buffer hoof comprises three structures with different cross sections connected up and down, the average cross-sectional area of the first structure at the proximal end of the ground and the average cross-sectional area of the third structure at the distal end of the ground are both smaller than the cross-sectional area of the second structure in the middle; the third structure is an elastic flexible unit; the thigh is connected to the calf by a planar rotation pair; and the calf is fixedly connected to the third structure of the buffer hoof.
[0012] Preferably, the thigh part and the calf part are connected through a knee joint assembly; the knee joint assembly includes a knee joint rotating ring, a first knee joint connecting part fixedly connected to the calf part, and a second knee joint connecting part fixedly connected to the thigh part; the knee joint rotating ring is respectively connected to the first knee joint connecting part and the second knee joint connecting part with pins to form a plane rotation pair.
[0013] Further preferably, the calf portion adopts a lightweight calf carbon fiber tube; the third structure of the buffer hoof (33) adopts a cylindrical buffer; the calf carbon fiber tube (501) and the buffer are connected by a buffer sleeve threaded nesting.
[0014] Preferably, the bionic biped robot further includes: an electromechanical control box and a joint yaw assembly; the electromechanical control box is used to place a robot control device and a power device; the control device includes an industrial computer, a motion controller, and a bottom circuit board; the power device includes a battery; the joint yaw assembly is used to control the robot's steering movement, including a yaw motor and a connector for connecting the electromechanical control box and the mechanical leg; the electromechanical control box is installed on the top of the bionic biped robot.
[0015] Further preferably, an inertial measurement unit is also installed on the electromechanical control box; and a pressure sensor is also installed on the bottom surface of the buffer hoof.
[0016] Further preferably, the number of yaw motors is one; the rotor of the yaw motor and the electromechanical control box are fixedly connected to the horizontal plate of the L-shaped first connecting plate, and the left mechanical leg of the robot is fixedly connected to the vertical plate of the L-shaped first connecting plate; the stator of the yaw motor is fixedly connected to the horizontal plate of the L-shaped second connecting plate, and the right mechanical leg of the robot is fixedly connected to the vertical plate of the L-shaped second connecting plate; when the left mechanical leg of the robot touches the ground as a supporting leg, the yaw motor rotates, the rotor of the yaw motor fixedly connected to the left mechanical leg remains stationary, and the right mechanical leg fixedly connected to the stator of the yaw motor rotates; when the right mechanical leg of the robot touches the ground as a supporting leg, the yaw motor rotates, the stator of the yaw motor fixedly connected to the right mechanical leg remains stationary, and the left mechanical leg and the control box fixedly connected to the yaw motor rotor rotate at the same time.
[0017] Further preferably, the number of the yaw motors is two; one end of the rotor and stator of the first yaw motor is fixedly connected to the electromechanical control box, and the corresponding other end is fixedly connected to the left mechanical leg of the robot through an L-shaped connecting plate; one end of the rotor and stator of the second yaw motor is fixedly connected to the electromechanical control box, and the corresponding other end is fixedly connected to the right mechanical leg of the robot through an L-shaped connecting plate;
[0018] Still further preferably, three motors are arranged at the hip joint of the thigh, the hip joint rolling motor controls the lateral swing of the leg, the hip joint pitching motor controls the forward and backward swing of the leg, and the knee joint pitching motor controls the movement of the knee joint; the rotation axes of the hip joint rolling motor and the hip joint pitching motor are not parallel.
[0019] Further preferably, the thigh part includes a four-bar linkage; the four-bar linkage is composed of a planar fourth connecting plate, a fisheye bearing rod, a fifth connecting plate, a thigh carbon fiber tube and a sixth connecting plate fixedly connected thereto; the stator of the hip joint rolling motor is connected to the joint yaw assembly; the hip joint pitch motor is fixed to the rotor of the hip joint rolling motor through an L-shaped third connecting plate; the fourth connecting plate is fixedly connected to the rotor of the knee joint pitch motor and the stator of the hip joint pitch motor through a first flange and a second flange respectively; the fisheye bearing rod is rotatably connected to the fourth connecting plate and the fifth connecting plate through fisheye bearings at both ends respectively; the sixth connecting plate is fixedly connected to the rotor of the knee joint pitch motor.
[0020] Preferably, a stable gait planning method for a bionic biped robot is characterized in that it includes the following steps: S1. Simplifying the bionic biped robot into a spring-loaded inverted pendulum and establishing a dynamic model; S2. Designing the parameter variables and constraints and initial parameters of the robot's periodic walking gait; the parameter variables include the robot's joint swing angle and angular velocity, the walking gait cycle, the torque received by each joint during walking, the walking speed, and the walking step length; S3. Simulating the walking gait, and evaluating the stability of the robot's periodic walking gait by using a method combining the Poincare section with the attraction domain feature; the gait stability includes local stability and global stability; the method for judging the local stability is: making a slight adjustment to the initial state to form a given error, and observing whether the walking state after walking can eliminate the error and converge to a stable state; if the given error can be eliminated, the walking gait trajectory becomes a fixed point on the Poincare section, then the walking gait is locally stable, otherwise, the walking gait is locally unstable; the method for judging the global stability is: analyzing the machine The attraction domain characteristics of the human periodic walking process in the state space. If the area of the attraction domain exceeds the threshold, the global stability of the walking gait is good, otherwise, the global stability of the walking gait is poor; S4. Based on the dynamic model described in S1, the parameter variables and their constraints and initial parameters of the robot's periodic walking gait described in S2 are used as feedback inputs to design the objective function of the robot's walking stability and the optimization algorithm for solving the optimal solution of the objective function to obtain the optimal parameters of the walking gait; the optimization algorithm is a sequential quadratic programming algorithm and / or a convex optimization algorithm; the objective function is: min[αF(x, J(j))+(1-α)G(x, S(q))], wherein α is the local stability weight, 1-α represents the global stability weight, x is the parameter variable of the periodic walking gait, J(j) represents the Jacobian matrix of the gait solution equation, S(q) represents the attraction domain area when judging the global stability, F is a mapping function of x and the Jacobian matrix J that reflects the local stability, and G is a mapping function of x and the attraction domain area that reflects the global stability.
[0021] Compared with the prior art, the present invention has the following beneficial effects:
[0022] 1. The use of a buffer hoof structure provides a new mode for achieving faster movement of bipedal robots;
[0023] 2. All drive motors are placed near the robot's hip joint to move the robot's overall center of gravity upward, thereby increasing the robot's movement speed and control response speed;
[0024] 3. Optimize the structure of setting a joint yaw motor for each of the two mechanical legs at the waist of the robot, and use one joint yaw motor to intermittently control the two mechanical legs, further reducing the weight of the robot and improving energy efficiency;
[0025] 4. The robot leg structure is constructed using a traditional four-bar linkage, but two hip joint motors for pitch and roll are set up. The flexibility of the two separate motors can provide better motion buffering, reduce the shaking of the robot, make the walking gait more stable, and avoid the high peak power loss caused by the coaxial axis of the joint motor, thereby improving energy efficiency.
[0026] 5. The robot does not have a drive motor at the knee joint, but relies on passive dynamic walking to effectively reduce the inertia of the moving part and make the robot's movement more stable;
[0027] 6. The thighs and calves are made of lightweight and elastic carbon fiber material to further reduce hardware loss and energy loss during exercise;
[0028] 7. Set up a buffer flexible unit to make the robot have good shock absorption performance, which can better buffer the foot-ground collision during walking, improve the robot's gait performance, and better protect the robot's mechanical and power systems;
[0029] 8. Use the spring-load inverted pendulum model to establish the robot's single-leg dynamics model, plan the robot's gait, and improve the robot's motion stability;
[0030] 9. Based on the Poincare mapping and combined with the characteristics of the attraction domain, a "limit cycle walking" periodic walking gait stability evaluation model is designed, and the objective function of the robot's walking stability and the optimization algorithm for solving the optimal solution of the objective function are designed to obtain the optimal parameters of the stable gait and ensure its robustness in the motion state. BRIEF DESCRIPTION OF THE DRAWINGS
[0031] Figure 1 This is a diagram of the three main hind limb skeletal structures of higher terrestrial animals. From left to right, the hind limb skeletal structures of plantigrade, toe-walking, and hoof-walking animals are shown in this figure.
[0032] Figure 2 This is a schematic structural diagram of a single yaw motor bionic biped robot according to an embodiment of the present invention;
[0033] Figure 3 This is a schematic structural diagram of a dual yaw motor bionic biped robot according to an embodiment of the present invention;
[0034] Figure 4 A three-dimensional model of a single yaw motor bionic biped robot according to an embodiment of the present invention Figure 1 ;
[0035] Figure 5 A three-dimensional model of a single yaw motor bionic biped robot according to an embodiment of the present invention Figure 2 ;
[0036] Figure 6 A schematic diagram of an electromechanical control structure of a bionic biped robot according to an embodiment of the present invention;
[0037] Figure 7 A simplified model diagram of a single leg of a bionic biped robot according to an embodiment of the present invention;
[0038] Figure 8 Schematic diagram of the motion process of a bionic biped robot according to an embodiment of the present invention; in the figure, when the left leg is supported in motion state P01, the yaw motor rotates and the right leg undergoes yaw motion; when the right leg is supported in motion state P02, the yaw motor rotates and the left leg undergoes yaw motion;
[0039] In the figure: 1, electromechanical control box, 2, joint yaw assembly, 3, mechanical leg, 4, stairs, 31, thigh, 32, calf, 33, buffer hoof, 34, thigh four-bar linkage, 201, L-shaped first connecting plate, 203, L-shaped second connecting plate, 204, L-shaped third connecting plate, 205, fourth connecting plate, 301, sixth connecting plate, 303, thigh carbon fiber tube, 305, fisheye bearing rod, 403, fifth connecting plate, 2021, first flange, 2022, second flange Lan, 404, knee joint rotating ring, 405, knee joint first connecting piece, 406, knee joint second connecting piece, 501, calf carbon fiber tube, 502, buffer sleeve, 601, yaw motor, 6021, left leg hip joint rolling motor, 6022, right leg hip joint rolling motor, 6031, left leg knee joint pitch motor, 6041, left leg hip joint pitch motor, 331, foot third structure, 332, foot second structure, 333, foot first structure, 334, pressure sensor. DETAILED DESCRIPTION
[0040] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the following embodiments and the accompanying drawings specifically illustrate a bionic biped robot and a stable gait planning method thereof of the present invention. It should be noted that the description of these implementation methods is used to help understand the present invention, but does not constitute a limitation of the present invention.
[0041] It should be noted that when two elements are "fixedly connected" or "rotatably connected", the two elements can be directly connected or there can be an intermediate element. On the contrary, when an element is said to be "directly on" another element, there is no intermediate element. The fixed connection or fixed connection method can be screw connection, welding, riveting, plug connection, or connection through a third component.
[0042] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention pertains. The terms used herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.
[0043] A bionic biped robot of the present invention has a bilaterally symmetrical structure, and comprises an electromechanical control box 1, a joint yaw assembly 2 and two mechanical legs 3.
[0044] The joint yaw assembly 2 is used to control the yaw movement (i.e., turning movement) of the robot legs, i.e., the rotation of the robot legs relative to the waist (Z axis), and includes one or two yaw motors. A bionic biped robot of the present invention can use two motors to independently control the yaw movement of the left and right legs of the robot, such as Figure 3 As shown; a motor can also be used to control the intermittent yaw movement of the left and right legs, such as Figure 2 As shown. The use of a yaw motor is conducive to further reducing the weight of the robot and improving energy efficiency; this embodiment preferably uses a yaw motor to control the yaw movement of the left and right legs as an example to illustrate the robot structure.
[0045] like Figure 4 As shown, the joint yaw assembly 2 is composed of a yaw motor 601 and a connecting member. The rotor of the yaw motor 601 is fixedly connected to the horizontal plate of the L-shaped first connecting plate 201 fixed on the electromechanical control box 1 through a threaded structure; the vertical plate of the L-shaped first connecting plate 201 is fixedly connected to the stator of the robot's left leg hip joint rolling motor 6021; the horizontal plate of the L-shaped second connecting plate 203 is fixedly connected to the stator of the yaw motor 601, and at the same time, its vertical plate is fixedly connected to the stator of the robot's right leg hip joint rolling motor 6022. This structure can realize the yaw movement of the robot controlled by the yaw motor 601, that is, the left and right turning movement of the robot (rotation around the z-axis).
[0046] When the robot turns, the industrial computer sends a signal to the controller to control the rotation of the yaw motor 601 to complete the yaw movement; when the robot's left mechanical leg touches the ground as a supporting leg, the yaw motor rotates, and the rotor of the yaw motor 601 fixedly connected to the left mechanical leg does not move, while the right mechanical leg fixedly connected to the stator of the yaw motor 601 rotates; when the robot's right mechanical leg touches the ground as a supporting leg, the yaw motor rotates, and the stator of the yaw motor 601 fixedly connected to the right mechanical leg does not move, while the left mechanical leg fixedly connected to the rotor of the yaw motor 601 and the control box rotate at the same time. Figure 8As shown, the movement is the relative movement of the left and right legs, that is: when the left leg touches the ground, the right leg rotates (the L-shaped second connecting plate 203 rotates); when the right leg touches the ground, the left leg rotates (the L-shaped first connecting plate 201 rotates); therefore, one yaw motor can realize the yaw movement of the robot.
[0047] It should be noted that when dual yaw motors are used to control the yaw motion of the robot, the connection structure that the robot can choose is: one end of the rotor and stator of the first yaw motor is fixedly connected to the electromechanical control box 1, and the corresponding other end is fixedly connected to the stator of the robot's left leg hip joint rolling motor 6021 through an L-shaped connecting plate; one end of the rotor and stator of the second yaw motor is fixedly connected to the electromechanical control box 1, and the corresponding other end is fixedly connected to the stator of the robot's right leg hip joint rolling motor 6022 through an L-shaped connecting plate. That is: the two yaw motor stators are respectively connected to the left and right leg hip joint rolling motors through connecting pieces, and their rotors are respectively connected to the electromechanical control box 1; or the two yaw motor rotors are respectively connected to the stators of the left and right leg hip joint rolling motors, and the two yaw motor stators are respectively connected to the electromechanical control box 1. When the left mechanical leg of the robot touches the ground as a supporting leg, the first yaw motor does not move, and the second yaw motor rotates, then the electromechanical control box does not move, and the right mechanical leg rotates; when the right mechanical leg of the robot touches the ground as a supporting leg, the second yaw motor does not move, and the first yaw motor rotates, then the electromechanical control box does not move, and the left mechanical leg rotates; in this way, the two motors control the yaw movement of the left and right legs respectively.
[0048] A bionic biped robot of the present invention has two symmetrical mechanical legs, each of which has three degrees of freedom. This embodiment takes the left leg as an example to illustrate the structure of the mechanical leg. Figure 2 As shown, the mechanical leg 3 includes a thigh portion 31 , a calf portion 32 and a cushioning hoof 33 .
[0049] like Figure 4 , Figure 5 As shown, the thigh 31 adopts a four-bar structure; the hip joint is provided with three motors, the hip joint rolling motor 6021 controls the lateral swing of the leg (rotation around the x-axis), the hip joint pitching motor 6041 controls the leg swing (rotation around the y-axis), and the knee joint pitching motor 6031 controls the knee joint movement. The leg swing of the robot mainly realizes the forward and backward movement of the robot, and the lateral swing of the leg is mainly used to adjust the center of gravity stability of the robot during movement.
[0050] The fourth connecting plate 205, the fisheye bearing rod 305, the fifth connecting plate 403, the thigh carbon fiber tube 303 and the sixth connecting plate 301 fixedly connected thereto constitute a four-bar series mechanism; the stator of the hip joint pitch motor 6041 is fixed to the rotor of the hip joint rolling motor 6021 through the L-shaped third connecting plate 204; the fourth connecting plate 205 is fixedly connected to the rotor of the knee joint pitch motor 6031 and the stator of the hip joint pitch motor 6041 through the first flange 2021 and the second flange 2022 respectively; the fisheye bearing rod 305 is rotatably connected to the fourth connecting plate 205 and the fifth connecting plate 403 through the fisheye bearings at both ends respectively; the sixth connecting plate 301 is fixedly connected to the rotor of the knee joint pitch motor 6031.
[0051] The calf part 32 comprises a calf carbon fiber tube 501 and a connecting piece; it is connected to the thigh part (31) through a knee joint assembly (knee joint rotating ring 404, a knee joint first connecting piece 405, and a knee joint second connecting piece 406); it is fixedly connected to the buffer hoof third structure 331 through a buffer sleeve 502. The knee joint second connecting piece 406 is fixedly connected to the thigh carbon fiber tube 303; the knee joint first connecting piece 405 is fixedly connected to the calf carbon fiber tube 501; the knee joint second connecting piece 406 and the knee joint first connecting piece 405 are respectively connected to the knee joint rotating ring 404 pin shaft to form a plane rotation pair; the rotation of the knee joint and the power transmission from the thigh to the calf are realized.
[0052] The buffer hoof 33 includes three structures connected up and down, a bionic hoof near the ground, a small structure that gradually tapers upward from the bionic hoof, and a joint that becomes thicker upward and is fixedly connected to the calf bone above; the bionic hoof includes bionic structures (parts or components) of various shapes, such as a bionic horse hoof structure, a bionic sheep hoof structure, and a simplified structure of a bionic hoof, such as a cylindrical shape, a conical shape, etc.; the two structures of the bionic hoof facing upward are preferably similar to the cylindrical shape of animal bones, wherein the joint fixedly connected to the calf is elastic; the three structures can be designed and processed as the same part, or they can be fixedly connected from different parts. In summary, the structural features of the buffer hoof 33 are: it includes three structures with different cross-sections connected up and down, the average cross-sectional area of the first structure near the ground and the average cross-sectional area of the third structure far from the ground are both larger than the cross-sectional area of the second structure in the middle, and the third structure far from the ground is an elastic flexible unit. This embodiment uses the simplified structure of a cylindrical bionic hoof to illustrate the structure of the buffer hoof, such as Figure 4 As shown, the cross-sectional diameter of the first structure 333 at the proximal end of the hoof-shaped foot and the cross-sectional diameter of the third structure 331 at the farthest end are both larger than the cross-sectional diameter of the second structure 332 in the middle, and the three structures are fixedly connected up and down to form a buffer hoof, wherein the third structure 331 at the farthest end of the buffer hoof adopts a cylindrical buffer, which is threadedly nested with the calf carbon fiber tube 501 through the buffer sleeve 502.
[0053] A bionic biped robot of the present invention has thighs and calves made of lightweight and elastic carbon fiber materials (303, 505), and a buffer hoof foot designed based on the characteristics of hoofed biological feet. A buffer flexible unit is provided, which can not only reduce the robot's own weight, but also make the robot have good shock absorption performance, can better buffer the foot-ground collision during walking, improve the robot's gait performance, and can better protect the robot's mechanical and power systems.
[0054] The electromechanical control box 1 is used to place the robot control device and the power device. In order to increase the physical center of gravity of the robot and facilitate control and motion performance improvement, the electromechanical control box is placed on the top of the robot. The control device includes a core control device that completes the robot's gait planning and walking control, such as a main controller (industrial computer), a microcontroller (motion controller), a low-level circuit board (including a power controller, a motor driver), and various sensors; the power device includes a battery, a power adapter, etc. A bionic bipedal robot of the present invention can use a battery to provide power for the robot's movement, or a tow-wire external power supply; this embodiment uses a battery as a power source as an example to illustrate the robot structure.
[0055] like Figure 6 , as shown in the schematic diagram of the electromechanical control structure of a bionic biped robot, the host computer includes a main controller and a microcontroller; the sensors include pressure sensors, speed sensors, and acceleration sensors; the actuators include hip joints, thighs, and calves. The actuators are driven by motors; the motors, sensor modules, and power supplies are all connected to the bottom circuit board; the bottom circuit board communicates with the host computer to transmit control signals and feedback signals. The battery powers the system. Among them, the host computer, the power supply device (battery), the bottom circuit board; the IMU sensor (inertial measurement unit) is arranged in the electromechanical control box 1; the pressure sensor 334 is distributed on the bottom of the robot, such as Figure 4 shown.
[0056] The robot controls the rotation of each motor through the host computer to achieve forward and backward and steering movements of the robot, and adjusts the output of each motor by coordinating and processing the feedback signals of sensors and motors to keep the robot's posture and movement stable. In order to facilitate unified management and control, the drive motors at four different positions of the robot: yaw motor, hip joint roll motor, hip joint pitch motor, knee joint pitch motor are preferably of the same model.
[0057] For a bionic biped robot of the present invention, it is crucial to solve the problem of maintaining the robot's own stability during walking so that it will not overturn or fall during dynamic walking. Gait planning is to plan each joint of the robot based on walking stability judgment conditions and other conditions and environmental factors. Accurate gait planning directly determines the walking stability and coordination of the biped robot in all aspects, so gait planning is a crucial part in the development of biped robots.
[0058] The invention discloses a stable gait planning method for a bionic biped robot. The method adopts a dynamic analysis method to plan a periodic walking gait of "limit cycle walking" for the robot, and adopts a method combining the Poincare section with the attraction domain characteristics to evaluate the stability of the robot's periodic walking gait. That is, the system state of each gait of the robot on the Poincare section will be mapped to the Poincare section of the next step. Each step of the walking gait constitutes a walking cycle. The system state of the next step on the Poincare section is the same as that of the previous step, but the system state at other times during the walking process does not need to be strictly tracked.
[0059] In the process of gait planning, we first need to design the gait process of the supporting leg; secondly, establish a dynamic model of walking gait, and describe and analyze the key links of the model; then, plan the initial gait parameters (including constraints on the angles, angular velocities and torque of each joint, as well as the speed and step length during walking), constraints, initial parameters, and establish a gait stability evaluation model; finally, optimize the gait parameters to obtain the ideal gait.
[0060] (1) Kinetic model
[0061] A bionic biped robot of the present invention can be equivalent to a spring-loaded inverted pendulum model (SLIP, Spring Loaded Inverted Pendulum Model) formed by a virtual spring connecting the hip joint and the foot end due to the compression and extension of the buffer hoof foot buffer, such as Figure 7 As shown; at the same time, due to the small spacing, the four motors at the hip can be approximated as a mass point, and the center of mass is located at the geometric center of the motor group; therefore, the whole robot can be controlled using the spring-load inverted pendulum model SLIP, as shown Figure 8 shown.
[0062] like Figure 8 As shown, the motion process can be divided into a support phase and a swing phase. In the support phase, the legs of the bipedal robot move around the foot end pivot to perform an approximate spring inverted pendulum motion. In the swing phase, the bipedal robot performs a free fall motion. The robot single-leg dynamics model can be expressed as follows:
[0063] ① Plane Lagrangian equation:
[0064]
[0065] Among them, m i is the mass of each member, J i is the moment of inertia of each member, x i ,y i 、j i is the centroid coordinate of each member, are the acceleration and angular acceleration of the mass center of the ith component when the component rotates uniformly at unit angular velocity. q is the angle of the crank.
[0066] ② Calculation of motion potential energy:
[0067]
[0068]
[0069] ③ Generalized force calculation:
[0070]
[0071] Among them, M is the comprehensive driving torque, δθ 1 is the comprehensive swing angle of the virtual spring of the leg.
[0072] The driving comprehensive torque M (i.e. the relationship between the driving force and the reaction force at the foot end) is obtained as:
[0073]
[0074] (2) Method for judging the stability of periodic walking gait
[0075] Since the bionic biped robot of the present invention has obvious periodic characteristics in bipedal walking, a periodic limit cycle walking gait can be constructed based on the above dynamic model. According to the limit cycle theory, if the system starts in a certain state and can return to the original state after a certain period of evolution, then in the next time period, the system will repeat the evolution process, so as to always remain on a closed-loop trajectory in the state space. The repeated trajectory is called a limit cycle. Take a certain characteristic motion state of the robot as the Poincare section (such as the moment after the foot hits the ground) to construct the Poincare mapping, design the walking gait of the robot, and take the intersection of the walking gait and the Poincare section as the initial state. After one cycle, the walking gait and the Poincare section still intersect at this point.
[0076] q n+1 =f(q n )
[0077] The state point with a periodic walking gait that satisfies the above equation is called a Poincare mapping fixed point.
[0078] The designed periodic walking gait includes a continuous swing phase and a discrete stance phase, such as Figure 8 shown.
[0079] The walking gait of the model is a periodic sequence of footsteps, which is stable overall, but not always stable locally. Therefore, the study of walking gait stability needs to consider: local stability and global stability. Local stability reflects whether the robot has the ability to walk stably, and global stability reflects the robustness of the robot's stable walking, that is, its anti-interference ability.
[0080] By making slight adjustments to the initial state to form a given error, observe whether the walking state can eliminate the error and converge to a stable state after walking, and verify its local stability. If the given error can be eliminated and the walking gait trajectory becomes a fixed point on the Poincare section, the walking gait is locally stable, otherwise, the walking gait is locally unstable.
[0081] By analyzing the characteristics of the attraction domain in the state space during the robot's periodic walking process, its global stability is verified. If the area of the attraction domain exceeds the threshold, it proves that the walking gait has good anti-interference ability and strong stability.
[0082] ① The ability of the fixed point of the Poincare section to attract adjacent points in the system state space is studied through local stability. The local stability of the walking gait is determined by the Jacobian matrix of the model. The error in walking is transmitted to the next step through the Jacobian matrix. When all the eigenvalues (Floquet multipliers) of the Jacobian matrix (J) are within the unit circle of the complex plane, the error in walking will become smaller and smaller in subsequent steps, and the periodic walking will be stable. The solution to local stability is proposed to adopt a numerical method, by adding small perturbations to the initial conditions of the walking gait, and calculating the local stability of the walking gait.
[0083] Assume that the robot starts walking from a point adjacent to the fixed point, that is, there is a small error between the initial state of the walking gait and the fixed point. After a few steps, if the error can be eliminated and the walking gait trajectory becomes a fixed point on the Poincare section, the walking gait is stable, otherwise, the walking gait is unstable. Apply a small disturbance to the initial state of the robot's walking, and the following relationship is derived:
[0084] q c +Δq n+1 =q n+1 =f(q n )=f(q c +Δq n )≈f(q c )+JΔq n =q c +JΔq n
[0085] in,
[0086] The error in the walk is passed to the next step through the multiplier J. If the maximum eigenvalue λ of the Jacobian matrix max If it is less than 1 or all eigenvalues are within the unit circle of the complex plane, the error in walking will become smaller and smaller in subsequent steps, and the periodic walking is stable. Otherwise, the error in walking will be magnified and the periodic walking will be unstable.
[0087] ② The ability of the closed-loop trajectory of periodic walking to attract adjacent trajectories in the state space is studied through global stability. The set of all system state points that eventually converge to the closed-loop trajectory after finite walking in the neighborhood of the closed-loop trajectory in the state space is called the attraction domain. The area of the attraction domain indicates the stability of the walking gait under interference. The larger the area of the attraction domain, the stronger the anti-interference ability of the gait. By analyzing the characteristics of the attraction domain of the walking gait, the global stability of the walking gait is studied to improve the anti-interference ability of the walking gait. The solution of the attraction domain is proposed to adopt the cell mapping method: the continuous system state space is discretized into cell units, and the cell unit mapping sequence and the attraction domain of the system are obtained by defining periodic cells and traversing the cell units.
[0088] (3) Optimization of walking gait parameters
[0089] Based on the robot spring-load inverted pendulum model, the parameter variables of the robot's periodic walking gait, its constraints and initial parameters are used as feedback input. The objective function of the robot's walking stability and the optimization algorithm for solving the optimal solution of the objective function are designed to obtain the optimal parameters of the walking gait.
[0090] ① Design of walking gait parameter variables. Walking gait parameter variables include the initial values of the robot's joint angles and angular velocities, the walking gait cycle, and the torque applied to each joint during walking.
[0091] ② Design of walking gait constraints. The walking gait constraints include constraints on the angles, angular velocities and torques of each joint, as well as the speed and step length during walking.
[0092] ③ Design of objective function. Design local stability and global stability as the objective function for walking gait solution (the process of finding the optimal solution of local stability and global stability is to find the minimum value of the objective function), as shown in the following formula:
[0093] min[αF(x,J(j))+(1-α)G(x,S(q))]
[0094] Among them, α represents the local stability weight, and correspondingly, 1-α represents the global stability weight; x refers to the gait parameters during walking (including the initial values of the robot's joint angles and angular velocities, the walking gait cycle, and the torque applied to each joint during walking); J(j) represents the Jacobian matrix of the gait solution equation; S(q) represents the area of the attraction domain when judging global stability; F is the mapping function between x and the Jacobian matrix J that reflects local stability; G is the mapping function between x and the area of the attraction domain that reflects global stability.
[0095] ④ Design of walking gait solver. The walking problem is regarded as a nonlinear constraint programming problem, and the minimum value of the objective function is solved by the optimization method, so as to achieve the solution of the optimal gait with stability.
[0096] The initial walking gait parameters and the joint torques during walking are substituted into the spring-inverted pendulum model for iteration until the foot-ground collision occurs after the mechanical leg pushes off the ground. The state after the collision will continue to be the initial state for the next time. The torque of each joint with respect to time during walking is calculated by the linear interpolation method. The optimal algorithm (sequential quadratic programming (SQP) algorithm and / or convex optimization algorithm) is applied to solve the minimum value of the objective function (i.e., the optimal state of stability) and solve the nonlinear programming problem under the walking gait constraints, thereby obtaining a stable walking gait.
[0097] The above-mentioned implementation modes are preferred cases of the present invention and are not used to limit the protection scope of the present invention. Various deformations or modifications that can be made by ordinary technicians in this field without creative work within the scope of the attached claims are still within the protection scope of this patent.
Claims
1. A bionic biped robot having two mechanical legs and a bilaterally symmetrical structure, characterized in that: The mechanical leg (3) comprises a thigh portion (31), a calf portion (32) and a buffer hoof (33), as well as an electromechanical control box (1) and a joint yaw assembly (2); The buffer hoof (33) comprises three structures with different cross sections connected in an upper and lower manner, wherein the average cross-sectional area of the first structure (333) near the ground and the average cross-sectional area of the third structure (331) far from the ground are both greater than the cross-sectional area of the second structure (332) in the middle; the third structure (331) is an elastic flexible unit; The thigh part (31) and the calf part (32) are connected via a planar rotation pair; The calf portion (32) is fixedly connected to the third structure (331) of the buffer hoof (33); The electromechanical control box (1) is used to house a robot control device and a power device; the control device comprises an industrial computer, a motion controller, and a bottom circuit board; the power device comprises a battery; the electromechanical control box (1) is installed on the top of the bionic biped robot; The joint yaw assembly (2) is used to control the steering movement of the robot, and comprises a yaw motor and a connecting piece for connecting the electromechanical control box (1) and the mechanical leg (3); Three motors are arranged at the hip joint of the thigh (31), the hip joint rolling motor (6021) controls the lateral swing of the leg, the hip joint pitching motor (6041) controls the forward and backward swing of the leg, and the knee joint pitching motor (6031) controls the movement of the knee joint; the rotation axes of the hip joint rolling motor (6021) and the hip joint pitching motor (6041) are not parallel; The thigh part (31) comprises a four-bar linkage; the four-bar linkage is composed of a planar fourth connecting plate (205), a fisheye bearing rod (305), a fifth connecting plate (403), a thigh carbon fiber tube (303), and a sixth connecting plate (301) fixedly connected thereto; The stator of the hip joint rolling motor (6021) is connected to the joint yaw component (2); The hip joint pitch motor (6041) is fixedly connected to the rotor of the hip joint rolling motor (6021) via an L-shaped third connecting plate (204); the fourth connecting plate (205) is fixedly connected to the rotor of the knee joint pitch motor (6031) and the stator of the hip joint pitch motor (6041) via a first flange (2021) and a second flange (2022) respectively; the fisheye bearing rod (305) is rotatably connected to the fourth connecting plate (205) and the fifth connecting plate (403) via fisheye bearings at both ends respectively; and the sixth connecting plate (301) is fixedly connected to the rotor of the knee joint pitch motor (6031).
2. A bionic biped robot as claimed in claim 1, characterized in that: The thigh part (31) and the calf part (32) are connected via a knee joint assembly; the knee joint assembly comprises a knee joint rotating ring (404), a first knee joint connecting member (405) fixedly connected to the calf part (32), and a second knee joint connecting member (406) fixedly connected to the thigh part (31); the knee joint rotating ring (404) is pin-connected to the first knee joint connecting member (405) and the second knee joint connecting member (406) respectively, to form a plane rotating pair.
3. A bionic biped robot as claimed in claim 2, characterized in that: The calf portion (32) adopts a lightweight calf carbon fiber tube (501); The third structure (331) of the buffer hoof (33) adopts a cylindrical buffer; The calf carbon fiber tube (501) and the buffer are connected by threaded nesting of the buffer sleeve (502).
4. A bionic biped robot as claimed in claim 1, characterized in that: The number of the yaw motor is one; The rotor of the yaw motor (601) and the electromechanical control box (1) are fixedly connected to the horizontal plate of the L-shaped first connecting plate (201), and the left mechanical leg of the robot is fixedly connected to the vertical plate of the L-shaped first connecting plate (201); the stator of the yaw motor (601) is fixedly connected to the horizontal plate of the L-shaped second connecting plate (203), and the right mechanical leg of the robot is fixedly connected to the vertical plate of the L-shaped second connecting plate (203); When the left mechanical leg of the robot touches the ground as a supporting leg, the yaw motor rotates, the rotor of the yaw motor (601) fixedly connected to the left mechanical leg remains stationary, and the right mechanical leg fixedly connected to the stator of the yaw motor (601) rotates; When the right mechanical leg of the robot touches the ground as a supporting leg, the yaw motor rotates, the stator of the yaw motor (601) fixedly connected to the right mechanical leg remains stationary, while the left mechanical leg and the control box fixedly connected to the rotor of the yaw motor (601) rotate simultaneously.
5. The bionic biped robot according to claim 1, characterized in that: The number of the yaw motors is two; One end of the rotor and stator of the first yaw motor is fixedly connected to the electromechanical control box (1), and the other corresponding end is fixedly connected to the left mechanical leg of the robot through an L-shaped connecting plate; One end of the rotor and stator of the second yaw motor is fixedly connected to the electromechanical control box (1), and the corresponding other end is fixedly connected to the right mechanical leg of the robot through an L-shaped connecting plate.
6. The bionic biped robot according to claim 1, characterized in that: An inertial measurement unit is also installed on the electromechanical control box (1); and a pressure sensor (334) is also installed on the bottom surface of the buffer hoof (33).
7. A stable gait planning method for a bionic biped robot as claimed in claim 1, characterized in that: The steps include: S1. simplifying the bionic bipedal robot into a spring-loaded inverted pendulum and establishing a dynamic model; S2. Design the parameter variables and constraints and initial parameters of the robot's periodic walking gait; the parameter variables include the robot's joint swing angle and angular velocity, the walking gait cycle, the torque of each joint during walking, the walking speed, and the walking step length; S3. Simulate walking gait and evaluate the stability of the robot's periodic walking gait by combining the Poincare section with the attraction domain feature; the gait stability includes local stability and global stability; The method for judging the local stability is: slightly adjusting the initial state to form a given error, and observing whether the walking state after walking can eliminate the error and converge to a stable state; if the given error can be eliminated and the walking gait trajectory becomes a fixed point on the Poincare section, the walking gait is locally stable, otherwise, the walking gait is locally unstable; The method for judging the global stability is: analyzing the attraction domain characteristics of the robot's periodic walking process in the state space, if the area of the attraction domain exceeds a threshold, the global stability of the walking gait is good, otherwise, the global stability of the walking gait is poor; S4. Based on the dynamic model described in S1, the parameter variables of the robot's periodic walking gait described in S2, its constraints and initial parameters are used as feedback input, an objective function of the robot's walking stability and an optimization algorithm for solving the optimal solution of the objective function are designed to obtain the optimal parameters of the walking gait; The optimization algorithm is a sequential quadratic programming algorithm and / or a convex optimization algorithm; The objective function is: ,in, is the local stability weight, represents the global stability weight, x is the parameter variable of the periodic walking gait, represents the Jacobian matrix of the gait solution equation, represents the area of the attraction domain when judging global stability, F is the mapping function of x and the Jacobian matrix J that reflects local stability, and G is the mapping function of x and the area of the attraction domain that reflects global stability.
