CN103407513A - Flat ground power type biped robot walking method adopting spring coupling - Google Patents

Abstract

The invention discloses a walking method of a ground dynamic biped robot adopting spring coupling. The method comprises the following steps: setting the initial state parameters of the robot; setting the control mode of the motor and the swing angle amplitude of the pendulum; The control method drives the motor to rotate, and then controls the movement of the swing rod on the inner and outer legs of the robot to make the robot walk; during the walking process, judge whether the walking state of the robot reaches the preset goal: if the walking state of the robot reaches the preset goal , then the walking adjustment of the robot is completed; otherwise, the predetermined setting of the robot is changed, and the robot is continuously driven to walk until the walking adjustment of the robot is completed. The invention controls the motors installed on the two legs to drive the swing rod, and adjusts the elastic potential energy of the swing rod through the coupling spring to realize the dynamic walking of the biped robot on the flat ground.

Description

Walking method of a dynamic biped robot on flat ground using spring coupling

technical field

The invention relates to the field of robot walking control, in particular to a walking method for a ground-powered biped robot using spring coupling. the

Background technique

Many researchers hope to achieve biped robots with human-like walking performance: natural, stable and efficient. In order to reduce the complexity of dynamic walking motion control, some constraints are added to the robot. According to the number of constraints, the walking modes of biped robots can be divided into three categories: static walking, ZMP walking, and limit cycle walking. Among them, static walking is the earliest and most constrained walking method, which requires that the center of mass of the robot is always kept within the polygon formed by the feet on the ground during the walking process. This walking mode requires precise control of the robot's joints and angles. Although it is easy to maintain the stability of the robot, it consumes too much energy due to excessive control requirements and greatly limits the walking speed of the robot. Therefore, most humanoid robots currently adopt a walking mode with less constraints, the Zero Moment Point (ZMP) walking mode. The zero moment point is a concept proposed by Yugoslavian scholar M.Vukobratovic in his theory of dynamic balance of walking robots in 1968. It refers to the point where the reaction force of the robot leg in contact with the ground does not generate torque around the horizontal axis. The ZMP theory requires that the supporting legs of the robot maintain horizontal contact with the ground, and its zero-moment point always remains within the polygon formed by the feet. Compared with the static walking mode, this mode does not necessarily constrain the center of mass of the robot within the polygon, which reduces artificial constraints. It has achieved great success in prototype applications, including Honda's ASIMO, Japan's AIST Research Institute's HRP4, and Sony's Qrio, all of which use this walking mode. However, robots designed according to the ZMP theory are still far behind human walking in terms of gait naturalness and energy efficiency. the

The third type of walking mode is limit cycle walking. This walking mode further reduces the artificial constraints on the robot's walking and effectively improves the performance of the robot. The gait of limit cycle walking presents periodic characteristics. The gait sequence is orbitally stable, that is, the gait sequence can form a stable limit cycle in the state space, but it does not have local stability at any instant in the gait cycle. sex. This walking mode has less artificial constraints on the robot, so it has a larger space to improve the energy efficiency and speed of the robot's walking. the

Passive walking is an example of limit cycle walking. McGeer's research found that passive biped robots can walk stably along small slopes only by gravity without any control, and the gait generated by their walking is very natural. Since then, many researchers have tried to realize the power walking of passive biped robots on flat ground based on this principle. For this reason, it is necessary to provide a new power source for the robot to replace the gravity on the slope. Researchers at Cornell University made a powered walking robot based on the principle of passive walking. They used the method of increasing the power at the ankle of the robot, and the soles of the feet kicked off the ground after each swinging leg collided with the ground to provide energy for walking. Its energy efficiency can reach the level of human walking, which is about one tenth of that of ZMP walking robot ASIMO. Researchers at Deflt University used the method of clamping the hip joint before the swing leg hits the ground, which also achieved the purpose of replenishing energy. the

Contents of the invention

In view of the above problems, the object of the present invention is to provide a method for controlling the motor-driven pendulum installed on the two legs, and adjusting its elastic potential energy through the coupling spring to realize the dynamic walking method of the biped robot on flat ground. the

In order to achieve the above object, a spring-coupled ground dynamic biped robot walking method according to the present invention is used to control the walking of the robot model, wherein the robot model includes at least a pair of inner legs and a pair of outer legs , the inner leg and the outer leg are hinged to form a hip structure;

The feet at the lower ends of the inner legs and the outer legs are provided with collision sensors that sense the collision between the feet and the ground; a controller is provided on the upper part of the hip structure, and the collision sensors are connected to the controller signal;

Motors are respectively installed on the opposite sides of the adjacent inner and outer legs, and the rotating shafts of the motors are connected to the swing rods; the ends of the two swing rods on the adjacent inner and outer legs are connected by springs;

The method includes the following steps:

Set the initial state parameters of the robot;

Set the control mode of the motor and the amplitude of the swing angle of the pendulum;

The controller drives the motor to rotate according to the control mode of the motor, and then controls the movement of the pendulum on the inner and outer legs of the robot to make the robot walk;

During the walking process, judge whether the walking state of the robot reaches the preset goal:

If the walking state of the robot reaches the preset target, the walking adjustment of the robot is completed;

Otherwise, change the preset settings of the robot and continue to drive the robot to walk until the walking adjustment of the robot is completed;

Wherein, the control method of the motor is:

In the process of walking, the motor rotates when the swinging legs of the inner and outer legs collide with the ground, and the two motors on the two inner legs rotate in the same way each time, and the two motors on the two outer legs rotate each time. The rotation action is the same; the rotation direction of the motors on the inner and outer legs is opposite, and the amplitude of the swing angle of the swing rod is the same. the

Preferably, during the walking process, when the walking state of the robot does not reach the preset target, the predetermined settings of the robot are changed as follows:

Change the magnitude of the swing angle of the pendulum, then continue to drive the robot to walk, and judge whether the walking state of the robot has reached the preset target; if the preset target has not been reached, repeat the process of changing the swing angle of the pendulum , until the walking state of the robot reaches the preset target. the

Preferably, the specific situation that needs to change the amplitude of the swing angle of the pendulum is as follows:

When the robot moves slowly after starting to walk, the stride is too small, and falls down without reaching the preset goal, it is necessary to increase the amplitude of the swing angle of the pendulum;

When the robot moves quickly after starting to walk, the stride is too large, and falls down without reaching the preset goal, it is necessary to reduce the swing angle amplitude of the pendulum. the

Preferably, when the magnitude of the swing angle of the pendulum is changed no matter what, the walking state of the robot cannot reach the preset target, and the changes made are as follows:

When the robot starts to walk slowly, the stride is too small, and falls down without reaching the preset goal, it is necessary to increase the length of the pendulum or replace the spring with a large elastic coefficient, and repeat the changing process until the robot's walking state achieve preset goals;

When the robot moves quickly after starting to walk, the stride is too large, and falls down without reaching the preset goal, it is necessary to shorten the length of the pendulum or replace the spring with a small elastic coefficient, and repeat the changing process until the robot walks The status reaches the preset target. the

The beneficial effects of the present invention are:

The walking method of the ground dynamic biped robot using spring coupling in the present invention is based on the principle of passive walking. During the walking process of the robot, the motors mounted on the two legs are controlled to drive the pendulum, and then the elasticity of the robot system is adjusted through the coupling spring. Potential energy supplements the energy of the robot to realize the open-loop walking of the robot on flat ground. This method well inherits the good walking performance of the passive walking robot, the control method is simple, and the walking speed of the robot can be adjusted. And because each time the pendulum has a cosine function relationship between the stretching amount of the spring and the angle between the two legs at the time of collision, the biped robot of the present invention has the advantage of progressive energy balance. the

Description of drawings

Fig. 1 is the side view of robot model described in the embodiment of the present invention;

Fig. 2 is the front view of robot model described in the embodiment of the present invention;

Fig. 3 is the principle diagram of the energy replenishment of walking of the robot model described in the embodiment of the present invention;

Fig. 4 is a schematic diagram of the evolution of the gait of the robot model according to the embodiment of the present invention along with the evolution of the motor rotation angle amplitude θ;

Fig. 5 is a schematic diagram of the evolution of the gait of the robot model according to the embodiment of the present invention with the spring coefficient K;

Fig. 6 is a flow chart of the implementation of the walking method according to the embodiment of the present invention. the

Detailed ways

The present invention will be further described below in conjunction with the accompanying drawings. the

As shown in Fig. 1-Fig. 2, the structure schematic diagram of the robot model is composed of a pair of inner legs 8, 9 and a pair of outer legs 2, 15, and the inner and outer legs are connected by hinges to form a hip structure 1. Both the inner and outer legs are constructed as a pair of legs to remove sideways movement. the

The lower end feet of the inner legs and outer legs are all provided with collision sensors 17, 18, 19, 20 that sense the collision of the feet with the ground, and the signal output terminals of the sensors 17, 18, 19, 20 are connected with the controller 16 signals. connected to the input. the

Motors 5, 10, 3, 12 are installed at the centers of mass of the inner legs and outer legs, and the rotating shafts of the motors 5, 10, 3, 12 are connected to the swing rods 6, 11, 4, 13 respectively, and the rotation of the motor drives the swing The bar rotates in the moving plane of the robot, the motors on the adjacent inner and outer legs are relatively installed, and the ends of the swing bars on the adjacent inner and outer legs are connected with linear springs 7,14. the

Aiming at the above-mentioned robot model, a walking method of a ground powered biped robot using spring coupling described in the embodiment of the present invention, as shown in Figure 6, the method includes the following steps:

Set the initial state parameters of the robot, wherein the initial state parameters include: θ _sw , θ _st , La, K, the units of θ _sw , θ _st are angles, the units of La are m, and the units of K are N/m, defining motion Alternate the support leg and the swing leg with the middle inner leg and outer leg:

θ _sw is the angle between the midline of the swing leg and the vertical line of the horizontal plane. When the swing leg is in front of the vertical line of the horizontal plane (the forward direction of the robot is forward), θ _sw >0, and after that, θ _sw <0 ,

θ _st is the angle between the centerline of the support leg and the vertical line of the horizontal plane. When the support leg is located in front of the vertical line of the horizontal plane (the forward direction of the robot is the front), θ _st >0, and after that, θ _st <0 ,

La is the length of the pendulum, which can be adjusted by changing the pendulum; K is the elastic coefficient of the spring, which can be adjusted by changing the spring. the

Set the control mode of the motor and the swing angle amplitude of the pendulum, wherein the control mode of the motor is: during walking, the motor rotates when the legs that play a swing role in the inner and outer legs collide with the ground, and the two legs The two motors on the inner legs have the same rotation action every time, and the two motors on the two outer legs have the same rotation action each time; the motors on the inner and outer legs rotate in opposite directions, and the swing angle amplitude of the swing rod is the same. the

The controller drives the motor to rotate according to the control mode of the motor, and then controls the movement of the pendulum on the inner and outer legs of the robot to make the robot walk;

During the walking process, judge whether the walking state of the robot reaches the preset goal (the preset goal can be measured by how many steps the robot walks smoothly):

If the walking state of the robot reaches the preset target, the walking adjustment of the robot is completed;

On the contrary, change the preset settings of the robot and continue to drive the robot to walk until the walking adjustment of the robot is completed. the

Among them, the gait cycle is defined as the time elapsed from the start of one step to the collision of the swing leg, where the start time refers to the moment when the swing leg leaves the ground, and the collision time refers to the collision between the swing leg and the ground, that is, the end of a gait cycle , the instant at which the next gait cycle begins, at which the supporting leg becomes the swinging leg, and the previous swinging leg becomes the supporting leg. In one gait cycle, the controller controls the motor to rotate at a fixed angle after each collision, and the amplitude of the motor rotation angle on the inner leg and the motor rotation angle on the outer leg are both θ, and the rotation The angles of the front and rear motors are symmetrical about the midline of the leg, the position of the swing leg swing bar is from behind the midline of the leg to before the midline of the leg, and the position of the support leg swing is from before the midline of the leg to behind the midline of the leg. the

This method is used to explain the regulation of robot walking:

During the walking process, when the walking state of the robot does not reach the preset goal, the predetermined settings of the robot are changed as follows:

Change the magnitude of the swing angle of the pendulum, then continue to drive the robot to walk, and judge whether the walking state of the robot has reached the preset target; if the preset target has not been reached, repeat the process of changing the swing angle of the pendulum , until the walking state of the robot reaches the preset target. the

Among them, the specific situation that needs to change the amplitude of the swing angle of the pendulum is as follows:

The robot moves slowly after starting to walk, the stride is too small, and falls down when the preset goal is not reached, indicating that θ is too small, and the added energy is insufficient, so increase θ by 1°, and repeat this process until the The robot can walk up to 100 steps. the

The robot walks fast after starting to walk, the stride is too large, and falls down when the preset goal is not reached, indicating that θ is too large, and too much energy is added, so that θ is reduced by 1°, and this process is repeated. Until the robot can walk 100 steps. the

It is worth noting that in the actual operation process, it is judged whether the robot can walk stably by observing the 100-step walking of the robot, and the number of observed walking steps can be reduced or increased according to actual requirements. the

There is also such a situation, when the magnitude of the swing angle of the pendulum is changed no matter what, the walking state of the robot cannot reach the preset target, and the specific changes are as follows:

When the robot moves slowly after starting to walk, the stride is too small, and falls down without reaching the preset goal, it is necessary to increase the length La of the pendulum by increasing La by 0.01L each time (L is the leg length of the robot), or Replace the spring with a larger elastic coefficient, increase the energy supplied by the system, and repeat this process until the robot can walk 100 steps;

When the robot moves quickly after starting to walk, the stride is too large, and falls down without reaching the preset goal, it is necessary to increase the length La of the pendulum by increasing La by 0.01L each time (L is the leg length of the robot), or Replace the spring with a larger elastic coefficient, increase the energy supplied by the system, and repeat this process until the robot can walk 100 steps. the

It is worth noting that in the actual operation process, it is judged whether the robot can walk stably by observing the 100-step walking of the robot, and the number of observed walking steps can be reduced or increased according to actual requirements. the

No matter how the operator adjusts, the robot cannot walk 100 steps, or can walk 100 steps, but needs to change the pace, then increase or decrease θ by 1°, or increase or decrease La by 0.01L, or replace Springs with larger or smaller spring coefficients can obtain different strides and gait cycles, and realize walking at different speeds. the

The following explains the above robot model and the principle of realizing the walking control of the robot model:

A complete one-step walk of the robot consists of a swing process and a collision. The swing process refers to the end of the supporting leg of the robot touching the ground and swinging forward with the end as the axis. At the same time, the swing leg swings from the back of the support leg to the front of the support leg in the air; the collision refers to the end of the swing process. The end of the swinging leg collides with the ground momentarily while the supporting leg lifts off the ground. After a collision the swing leg converts to a support leg and the support leg converts to a swing leg. A complete one-step walk of the robot starts immediately after the collision of the previous step, and ends after the next collision through the swing process. the

As shown in Figure 3, the robot starts at time A, with a certain initial velocity at this time, and the swing leg is about to leave the ground. Time B is the situation when the swing leg is behind the support leg, and time C is the situation when the swing leg swings past the support leg. Time D is the situation where the swing leg collides with the ground. Immediately after the swing leg collides, the motor rotates to make the swing rod swing and stretch the spring, and the state of the robot returns to time A. The motor rotates to make the swing rod swing and stretch the spring, so as to increase the elastic potential energy of the system and replenish energy for the system. Start a new cycle of walking. The pendulum swings and stretches the spring to add energy to the system, and the increased elastic potential energy increases the system energy by E1. The collision between the swing leg and the ground and other factors cause the system energy to decrease by E2. In order to make the system walk in a stable limit cycle, E1 should be roughly equal to E2. If E1 is greater than E2, the machine may converge to another limit cycle to walk faster, or walk with a period doubled gait, or walk with a chaotic gait, or fall forward too fast. If E1 is smaller than E2, the machine may converge to another limit cycle and walk at a slower speed, or the increase in system energy cannot make up for the loss of system energy and the machine will walk more and more slowly, eventually falling down. the

The biped robot described in the present invention has the characteristics of progressive energy balance, which is manifested as follows: during one-step walking, when the added energy is the same as the energy lost in the collision of the swinging legs, the walking of the robot enters the limit cycle, The initial velocity after the collision is the velocity at the fixed point. If the added energy is greater than the energy lost by the collision of the system, the robot will walk with a larger stride and faster speed, so the energy loss of each collision will increase. The angle between the legs has a cosine function relationship. The larger the stride, the larger the angle between the two legs at the time of collision, resulting in a smaller amount of stretching of the swing rod to the spring than the previous stretch of the swing rod to the spring. Comparing the elastic potential energy Since the amount of replenishment was reduced last time, the energy replenished and the energy lost in the collision of the swinging leg finally reached a balance, and the gait of the robot converged to the limit cycle with a larger stride and faster walking speed. If the added energy is less than the energy lost by the collision of the system, the robot will walk with a smaller stride and slower speed, so the energy loss of each collision is reduced, and because the stretching amount of the spring by the pendulum is different from the collision time The angle between the legs has a cosine function relationship. The smaller the stride, the smaller the angle between the two legs at the time of collision, resulting in an increase in the stretching of the swing rod to the spring compared with the previous stretch of the swing rod to the spring, and the elastic potential energy is added. The amount of replenishment increased last time, and finally the energy replenished and the energy lost in the collision of the swinging leg reached a balance, and the robot gait converged to the limit cycle with a smaller stride and slower walking speed. the

The biped robot of the present invention has the characteristic of gait evolution, which is manifested as follows. As the pendulum swing angle amplitude θ continues to change, the gait of the robot will show three stages of evolution, that is, from a stable and symmetrical single-period gait to an asymmetric gait with double-period, and finally into each step For different chaotic gaits, the gait evolution diagram is shown in Figure 4, which shows the change of θ _st value of the robot with θ at the moment of collision. Changes in the pendulum length La and the spring constant K can also cause gait evolution. Under certain parameters, the gait of the robot will also show the characteristics of four-stage evolution, that is, the chaotic gait returns to the asymmetric gait with doubled period. Figure 5 shows the change of the θ _st value of the robot with the spring coefficient K at the moment of collision Four stages of evolution emerge. The characteristic of gait evolution embodies the diversity of gait of the biped robot described in the present invention, and also shows that the robot described in the present invention maintains the characteristics of a passive walking robot.

The walking method of the ground dynamic biped robot using spring coupling in the present invention is based on the principle of passive walking. During the walking process of the robot, the motors mounted on the two legs are controlled to drive the pendulum, and then the elasticity of the robot system is adjusted through the coupling spring. Potential energy supplements the energy of the robot to realize the open-loop walking of the robot on flat ground. This method well inherits the good walking performance of the passive walking robot, the control method is simple, and the walking speed of the robot can be adjusted. And because each time the pendulum has a cosine function relationship between the stretching amount of the spring and the angle between the two legs at the time of collision, the biped robot of the present invention has the advantage of progressive energy balance. the

The above are only preferred embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention are all Should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be defined by the claims. the

Claims (4)

1. A walking method for a flat ground power type biped robot that adopts spring coupling, is characterized in that the method is used to control the walking of a robot model, wherein the robot model includes at least a pair of inner legs and a pair of outer legs, and the The inner leg and the outer leg are connected by hinges to form the hip structure; The feet at the lower ends of the inner legs and the outer legs are equipped with collision sensors that sense the collision between the feet and the ground; a controller is arranged on the upper part of the hip structure, and the collision sensors are connected to the controller signal; Motors are respectively installed on the opposite sides of the adjacent inner and outer legs, and the rotating shafts of the motors are connected to swing rods; the ends of the two swing rods on the adjacent inner and outer legs are connected by springs; The method includes the following steps: Set the initial state parameters of the robot; Set the control mode of the motor and the amplitude of the swing angle of the pendulum; The controller drives the motor to rotate according to the control mode of the motor, and then controls the movement of the pendulum on the inner and outer legs of the robot to make the robot walk; During the walking process, judge whether the walking state of the robot reaches the preset target: If the walking state of the robot reaches the preset target, the walking adjustment of the robot is completed; Otherwise, change the preset settings of the robot and continue to drive the robot to walk until the walking adjustment of the robot is completed; Wherein, the control method of the motor is: In the process of walking, the motor rotates when the swinging legs of the inner and outer legs collide with the ground, and the two motors on the two inner legs rotate in the same way each time, and the two motors on the two outer legs rotate each time. The rotation action is the same; the rotation direction of the motors on the inner and outer legs is opposite, and the amplitude of the swing angle of the swing rod is the same. 2. The walking method of the ground dynamic biped robot adopting spring coupling according to claim 1, characterized in that, in the walking process, when the walking state of the robot does not reach the preset target, the preset setting of the robot is changed as follows: : Change the magnitude of the swing angle of the pendulum, then continue to drive the robot to walk, and judge whether the walking state of the robot has reached the preset target; if the preset target has not been reached, repeat the process of changing the swing angle of the pendulum , until the walking state of the robot reaches the preset target. 3. The ground-level power-type biped robot walking method adopting spring coupling according to claim 2 is characterized in that, the specific circumstances that need to change the amplitude of the swing angle of the swing rod are: When the robot moves slowly after starting to walk, the stride is too small, and falls down without reaching the preset goal, it is necessary to increase the swing angle amplitude of the pendulum; When the robot moves quickly after starting to walk, the stride is too large, and falls down without reaching the preset goal, it is necessary to reduce the swing angle amplitude of the pendulum. 4. The walking method of the ground dynamic biped robot adopting spring coupling according to claim 2, wherein the walking state of the robot cannot be reached when the magnitude of the swing angle of the fork is changed anyway. If the target is set, the changes to be made are as follows: When the robot moves slowly after starting to walk, the stride is too small, and falls down without reaching the preset goal, it is necessary to increase the length of the pendulum or replace the spring with a large elastic coefficient, and repeat the changing process until the robot's walking state achieve preset goals; When the robot moves quickly after starting to walk, the stride is too large, and falls down without reaching the preset goal, it is necessary to shorten the length of the pendulum or replace the spring with a small elastic coefficient, and repeat the changing process until the robot walks The status reaches the preset target.

Images