Movement control signal generation method and device for tensegrity robot

Abstract

The invention discloses a method and device for generating a movement control signal of a tensegrity robot, and belongs to the technical field of robot control, the method comprises the following steps: selecting a part of cables of the tensegrity robot as an active cable, and taking the remaining part as a passive cable; inputting the initial value of the gait variable of each active cable into an optimization algorithm for iteration to obtain an updated value of the gait variable of each active cable, and inputting the updated value of the gait variable of each active cable into the optimization algorithm for repeated iteration until a specified number of iterations is reached; carrying out fine tuning and optimization on the gait variable through multi-round iteration; calculating an active cable update control signal corresponding to the update value of the gait variable of each active cable in each round of iteration and an evaluation parameter of the active cable update control signal so as to determine an active cable target control signal; the gait variable determined in the mode and the corresponding active cable target control signal are used for controlling the active cable and driving the passive cable to execute the gait action, and the stretch-draw whole robot can be accurately controlled to rapidly walk and move.

Description

Technical Field

[0001] The present invention belongs to the technical field of robot control, and more specifically, relates to a method and device for generating movement control signals of a tensegrity robot. Background Art

[0002] Tensegrity robots are a type of robot inspired by tensegrity structures. Tensegrity structures are structural systems that achieve stability and balance through the interaction of separate rigid components (such as rods or beams) and continuous flexible components (such as ropes or cables). Tensegrity structures are lightweight and strong. Tensegrity structures are usually light in weight, but can withstand large external loads, are easy to fold, flexible and adaptable (due to the characteristics of their flexible components, this type of structure can adapt to complex environments and deformations) and have efficient force transmission (force is transmitted through flexible components, so that the entire structure can effectively disperse and withstand external forces). Therefore, tensegrity robots can effectively save launch costs when exploring complex environments in outer space, and have high adaptability and impact resistance. They are effective robots for exploring extraterrestrial bodies. The robot usually produces deformation and movement by controlling the length change of the flexible cable or the length change of the rod.

[0003] At present, the motion control method of the tensegrity robot on the ground mainly utilizes deformation to change the position of the center of gravity and then generate torque for rolling, which has the problems of low movement efficiency and poor control accuracy. Summary of the invention

[0004] In view of the above defects or improvement needs of the prior art, the present invention provides a method and device for generating movement control signals of a tensegrity robot, the purpose of which is to solve the technical problems of low movement efficiency and poor control accuracy of the existing tensegrity robot.

[0005] To achieve the above object, according to one aspect of the present invention, a method for generating a movement control signal of a tensegrity robot is provided, comprising:

[0006] S1: Selecting some of the cables of the tensegrity robot as active cables and the remaining cables as passive cables;

[0007] S2: Determine each gait action in the movement behavior of the tensegrity robot; set the initial range of the gait variable corresponding to each gait action; generate the initial value of each gait variable of the active cable within the initial range; the gait variable includes: a time variable t representing the time consumed to complete each gait action, an active cable length variable X representing the change of the current length of the active cable compared to the original length, an active cable extension variable H representing the extension or shortening of the active cable, and a correction variable G for improving the trajectory deviation of the tensegrity robot when it moves straight;

[0008] S3: inputting the initial value of the gait variable of each active cable into the optimization algorithm for iteration to obtain the updated value of the gait variable of each active cable, and then inputting the updated value of the gait variable of the active cable into the optimization algorithm for repeated iteration until a specified number of iterations is reached;

[0009] S4: calculating each active cable update control signal and its corresponding evaluation parameter corresponding to the update value of the gait variable of each active cable in each iteration;

[0010] S5: Using the active cable update control signal corresponding to the evaluation parameter satisfying the preset condition as the active cable target control signal to control the active cable and drive the passive cable to perform the gait action.

[0011] In one of the embodiments, in S1: the selection principle of the active cable is: select a waist cable that can change the tilt posture of the tensegrity robot and a bottom cable that is necessary to move the tensegrity robot to contact the ground as the active cable.

[0012] In one embodiment, the active cable length variable X, the active cable expansion and contraction variable H and the correction variable G are all percentages; the product of the active cable length variable X and the corresponding original length of the active cable represents the change in the length of the active cable compared to the original length of the active cable in the current round of iteration; the active cable expansion and contraction variable H has the same amplitude as the active cable length variable X, and the sign of the active cable expansion and contraction variable H is positive or negative, and when it is positive, it indicates that the length is elongated relative to the initial moment, and when it is negative, it indicates that the length of the active cable is shortened relative to the initial moment.

[0013] In one embodiment, the S4 includes:

[0014] Interpolating the time series of the length changes of each active cable corresponding to the updated value of the gait variable of each active cable in each iteration, to obtain an active cable update control signal corresponding to the length change of the active cable over time;

[0015] The active cable update control signal is input into a simulation program to obtain the coordinates after the movement, and the coordinates after the movement and the coordinates before the movement are input into a fitness function for calculation to obtain the fitness representing the evaluation parameter corresponding to the active cable update control signal.

[0016] In one embodiment, the fitness function is fit=(X(end)-X(0))-abs(Y(end)-Y(0)), where X(end) and Y(end) respectively represent the coordinates of the horizontal and vertical axes at the end of the robot's movement, and X(0) and Y(0) respectively represent the coordinates of the horizontal and vertical axes at the beginning of the robot's movement; abs represents the absolute value function.

[0017] In one embodiment, S5 includes: selecting the active search update control signal corresponding to the maximum fitness as the active search target control signal.

[0018] In one of the embodiments, the optimization algorithm is a genetic algorithm, a particle swarm algorithm or an ant colony algorithm.

[0019] According to another aspect of the present invention, there is provided a device for generating a movement control signal of a tensegrity robot, comprising:

[0020] A selection module, for selecting a portion of all cables of the tensegrity robot as active cables and the remaining portion as passive cables;

[0021] A setting module is used to determine each gait action in the movement behavior of the tensegrity robot; set the initial range of the gait variable corresponding to each gait action; generate the initial value of each gait variable of the active cable within the initial range; the gait variable includes: a time variable t representing the time consumed to complete each gait action, an active cable length variable X representing the change of the current length of the active cable compared to the original length, an active cable extension variable H representing that the active cable is extended or shortened, and a correction variable G for improving the trajectory deviation of the tensegrity robot when it moves straight;

[0022] An optimization module, used for inputting the initial value of the gait variable of each active cable into the optimization algorithm for iteration, obtaining the updated value of the gait variable of each active cable, and then inputting the updated value of the gait variable of the active cable into the optimization algorithm for repeated iteration until a specified number of iterations is reached;

[0023] A calculation module, used for calculating each active cable update control signal and its corresponding evaluation parameter corresponding to the update value of each gait variable of the active cable in each iteration;

[0024] The determination module is used to use the active cable update control signal corresponding to the evaluation parameter satisfying the preset condition as the active cable target control signal to control the active cable and drive the passive cable to perform the gait action.

[0025] According to another aspect of the present invention, there is provided a movement control method of a tensegrity robot, comprising:

[0026] Based on the mobile control signal generation method of the tensegrity robot, an active cable target control signal is obtained;

[0027] The active cable target control signal is used to control the active cable of the tensegrity robot to extend and retract and drive the passive cable to move, so as to perform corresponding gait movements.

[0028] According to another aspect of the present invention, there is provided an electronic device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method for generating a movement control signal of a tensegrity robot or the movement control method of the tensegrity robot when executing the computer program.

[0029] According to another aspect of the present invention, there is provided a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the method for generating a movement control signal of a tensegrity robot or the movement control method of the tensegrity robot.

[0030] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:

[0031] (1) The present invention provides a method for generating movement control signals of a tensegrity robot, wherein some of the cables of the tensegrity robot are selected as active cables, and only some of the cables need to be controlled, thereby reducing the control complexity; the initial values ​​of the gait variables of each active cable are input into an optimization algorithm for iteration to obtain updated values ​​of the gait variables of each active cable, and then the updated values ​​of the gait variables of the active cables are input into the optimization algorithm for repeated iteration until a specified number of iterations is reached; the gait variables are fine-tuned and optimized through multiple rounds of iterations; the active cable update control signals and their evaluation parameters corresponding to the updated values ​​of the gait variables of each active cable in each round of iteration are calculated; and the active cable target control signals are selected according to the evaluation parameters; the gait variables determined in this way and the active cable target control signals corresponding thereto can accurately control the tensegrity robot to walk quickly, thereby solving the technical problems of low movement efficiency and poor control accuracy of the existing tensegrity robots.

[0032] (2) The principle for selecting the active cables in this scheme is: select the waist cable that can change the tilt posture of the tensegrity robot and the bottom cable that is necessary to move the tensegrity robot to contact the ground as the active cables; the advantage is that the number of active cables can be reduced, and only the above-mentioned active cables need to be selected to complete the movement control of the tensegrity robot, thereby reducing the calculation amount of the optimization algorithm and the control difficulty.

[0033] (3) The active cable length variable X, active cable extension variable H and correction variable G described in this scheme are all percentage numbers. The advantage is that they can clearly and simply represent the length state of each active cable in the motion gait. The percentage method makes it applicable to the original length of different active cables, which is convenient for unifying the range of length variables.

[0034] (4) In this scheme, the time series of the length changes of each active cable corresponding to the updated value of the gait variable of each active cable in each iteration are interpolated. The advantage is that a continuous control signal can be obtained. Different control signals can be obtained by using different interpolation methods. For example, using quadratic function interpolation will make the control signal smoother. Furthermore, the active cable update control signal is input into the simulation program to obtain the coordinates after movement, and the coordinates after movement and the coordinates before movement are input into the fitness function to calculate the fitness representing the evaluation parameter. The advantage of this method is that the control result of the control signal is evaluated by using the simulation program, so that the evaluation of the control signal is more reliable without the need for actual prototype verification. At the same time, it can speed up the optimization iteration of the entire control signal and improve the fitness of the control signal.

[0035] (5) In this scheme, the fitness function is set to fit = (X(end)-X(0))-abs(Y(end)-Y(0)). The advantage is that it allows the robot to move straight in the specified direction and reduce the deviation from the specified direction.

[0036] (6) In this scheme, the active search update control signal corresponding to the maximum fitness is selected as the active search target control signal. The advantage is that the calculation complexity is low and the optimal active search target control signal can be quickly found.

[0037] (7) The optimization algorithm described in this scheme is a genetic algorithm, a particle swarm algorithm or an ant colony algorithm. The above algorithms are all mature and stable optimization algorithms that can improve the stability of the mobile control signal generation method of the entire tensegrity robot. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 is a flow chart of a method for generating a movement control signal of a tensegrity robot provided in Example 1 of the present invention;

[0039] Figure 2 is a schematic structural diagram of a three-rod and nine-cable prismatic tensegrity robot provided in Example 1 of the present invention;

[0040] Figure 3 This is a simulation diagram of the movement effect of the three-rod and nine-cable prismatic tensegrity robot provided in Example 1 of the present invention moving straight in accordance with the first movement mode;

[0041] Figure 4 This is a simulation diagram of the movement effect corresponding to the straight movement of the three-rod and nine-cable prismatic tensegrity robot provided in Example 1 of the present invention according to the second movement mode. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0043] Example 1

[0044] The present embodiments are all described by taking a three-rod and nine-cable prismatic tensegrity robot as an example. It should be noted that the number of rods and cables of the tensegrity robot of the present application is not limited, and tensegrity robots with other numbers of rods and cables can also generate movement control signals according to the following method.

[0045] This embodiment provides a method for generating a movement control signal of a tensegrity robot, such as Figure 1 As shown, the following steps S1-S5 are included.

[0046] S1: Select some cables from all the cables of the tensegrity robot as active cables and the rest as passive cables.

[0047] Among them, according to the structural characteristics and force analysis of the three-rod and nine-cable prismatic robot, the robot's moving gait is designed, and some cables are selected from all the cables of the tensegrity robot as active cables, which can be controlled by motors and can actively change their lengths; the rest are passive cables, that is, cables that have no motor control and can only passively change their lengths.

[0048] As an optional implementation, in S1: the principle of selecting the active cable is: selecting the waist cable that can change the tilt posture of the tensegrity robot and the bottom cable necessary for moving the tensegrity robot to contact the ground as the active cable.

[0049] Take the three-rod and nine-cable tensegrity robot as an example. Figure 2As shown, it includes 3 rods and 9 cables. Numbers 1, 2, and 3 represent bottom cables, numbers 4, 5, and 6 represent top cables, numbers 7, 8, and 9 represent waist cables, numbers B1, B2, and B3 represent rod structures, and numbers F1, F2, and F3 represent the feet of the rods that are in contact with the ground, where F3 is the front foot, and F1 and F2 are the back feet. Each rod has two rod ends, where the waist cable is a cable connecting the bottom rod end (the rod end in contact with the ground) and the top rod end (the rod end not in contact with the ground), the bottom cable is a cable connecting the bottom rod end, and the top cable is a cable connecting the top rod end. When the three-rod and nine-cable tensegrity robot is placed along the forward direction, that is, the midline of the equilateral triangle formed by the three bottom cables on the bottom coincides with the forward direction, the rod ends and waist cables are divided into two parts, the rod end relatively forward is the front rod end, and the rod end relatively backward is the rear rod end. Similarly, the waist cable relatively forward is the front waist cable, and the waist cable relatively backward is the rear waist cable. The principle for selecting the active and passive cables is that the necessary active cables include the waist cables (e.g. Figure 2 7, 8, 9) and the necessary bottom cables (e.g. Figure 2 1, 2, 3) in , and the rest can be passive cables. The more active cables are selected, the better the control effect may be, but at the same time it will increase the calculation amount of the optimization algorithm.

[0050] S2: Determine each gait action in the movement behavior of the tensegrity robot; set the initial range of the gait variable corresponding to each gait action; wherein, the gait variable is set according to the movement gait of the robot and the selected active cable, and the initial value of the gait variable of each active cable can be generated within the initial range by making a robot movement gait table and setting the initial range of the gait variable; the gait variables include: a time variable t representing the time consumed to complete each gait action, an active cable length variable X representing the change of the current length of the active cable compared to the original length, an active cable extension variable H representing whether the active cable is extended or shortened, and a correction variable G for improving the trajectory deviation of the tensegrity robot when it moves straight.

[0051] The idea of ​​designing the gait action corresponding to straight walking is as follows: first, the waist cable is changed so that the three-rod and nine-rod prismatic tensegrity robot tilts as a whole in the forward direction, thereby generating a pressure difference between the front and rear rod ends of the three-rod and nine-rod prismatic tensegrity robot that are in contact with the ground; then, the active cables at the front and rear rod ends are contracted so that the rear rod end moves forward; at this time, the tensegrity robot tilts backward as a whole again, and the active cables between the front and rear rod ends are restored to their initial length so that the front rod end moves forward; then, the robot is restored to the state of being tilted toward the front as a whole, completing a gait cycle.

[0052] As an optional implementation, the active cable length variable X, the active cable expansion and contraction variable H and the correction variable G are all percentages; the product of the active cable length variable X and the corresponding original length of the active cable represents the length of the active cable in the current round of iteration; the active cable expansion and contraction variable H has the same amplitude as the active cable length variable X, and the sign of the active cable expansion and contraction variable H is positive or negative, and when it is positive, it indicates that the length is elongated relative to the initial moment, and when it is negative, it indicates that the length of the active cable is shortened relative to the initial moment.

[0053] Among them, the gait variables are X i , H i , G i ,t i Four different phase variables, X i , H i , G i The product of the gait variable and the initial length of the corresponding active cable can represent the change in the length of the active cable in the current gait stage compared with the initial length. A positive value indicates that the length is elongated relative to the initial moment, and a negative value indicates that the length of the active cable is shortened relative to the initial moment. i In order to improve the trajectory deviation of the three-rod and nine-cable prismatic tensegrity robot when it moves straight, t i The time spent on the action in this gait phase, where X i >0,H i Not sure whether it is positive or negative, 0<G i <1,t i >0. G i The purpose is to correct the behavior of the three-rod and nine-cable prismatic tensegrity robot deviating from the predetermined trajectory to the left when the robot is leaning forward or moving forward due to the leftward deviation of the center of gravity. During the iteration process, you can take a result to observe whether the robot is deviating to the left or right, or increase the G of the corresponding waist cable according to the direction in which the robot's center of gravity deviates from the predetermined trajectory. i Correct the variables so that the robot can move in the intended direction.

[0054] The length state variables in the gait variables are divided into X i , H i , G i , where X i (X 1 ,X 2 ,X 3 ...) is used to indicate the state parameters of the active cables, whether they are contracted or stretched. For example, when a three-bar, nine-cable prismatic tensegrity robot leans forward, the states of the three waist cables are clear, and the front waist cable contracts (contraction is indicated by -X i Indicated by X), the back waist cable is extended (extension is indicated by X i express).

[0055] Length status parameter H i (H 1 ,H 2 ,H 3 …) is a state parameter used to indicate whether the unknown active cable is in a contraction or extension state. Its value is uncertain. In some cases where it is difficult to determine the length change state of the active cable, it can be used to indicate the length state of the active cable.

[0056] Length status parameter G i (G 1 ,G 2 ,G 3 …) is used to represent the adjustment parameters to correct the problem of the three-rod nine-cable prismatic tensegrity robot deviating from the straight trajectory when it moves straight. It is mainly aimed at the waist cable. Since the center of gravity of the three-rod nine-cable prismatic tensegrity robot is not actually at the center of the bottom triangle, for example, when the robot leans forward, it is necessary to adjust the length of the waist cable so that the center of gravity coincides with the straight direction. At this time, the left waist cable length state parameter is subtracted from X i G i In order to improve the right-leaning posture of the three-bar nine-cable prismatic tensegrity robot, the pressure of the two rear feet on the ground is equal, and when the rear feet are moved forward, the rear feet can move the same distance and the straight track will not deviate.

[0057] It should be noted that the initial range of each gait variable is determined according to the actual hardware conditions, where X i and H i The size setting should ensure that the length of the active cable does not exceed the range of the active cable in the actual hardware. The signal generation method is interpolation, that is, the designed gait determines the state at the end of each gait stage. The time series of the active cable length change can be obtained according to the gait table. The continuous active cable length change control signal is generated by interpolating the time series of the length change.

[0058] S3: Inputting the initial value of the gait variable of each active cable into the optimization algorithm for iteration to obtain the updated value of the gait variable of each active cable, and then inputting the updated value of the gait variable of the active cable into the optimization algorithm for repeated iteration until the specified number of iterations is reached. The optimization algorithm may be a genetic algorithm, a particle swarm algorithm, or an ant colony algorithm.

[0059] S4: Calculate each active cable update control signal and its corresponding evaluation parameter corresponding to the update value of the gait variable of each active cable in each iteration.

[0060] As an optional implementation, S4 includes: interpolating the time series of the change in the length of each active cable corresponding to the updated value of the gait variable of each active cable in each round of iteration to obtain an active cable update control signal corresponding to the change in the length of the active cable over time; inputting the active cable update control signal into the simulation program to obtain the coordinates after moving, and inputting the coordinates after moving and the coordinates before moving into the fitness function for calculation to obtain the fitness representing the evaluation parameter corresponding to the active cable update control signal.

[0061] Among them, the simulation needs to consider stopping the simulation when it is about to tip over and regenerating a new control signal through a genetic algorithm. The judgment condition for stopping can be set to that the vertical height difference between the upper end of each rod off the ground and the lower end of the rod touching the ground is greater than a fixed value. This fixed value can be set freely, depending on the tilt state that you want to abandon.

[0062] As an optional implementation, the fitness function is fit=(X(end)-X(0))-abs(Y(end)-Y(0)), where X(end) and Y(end) respectively represent the coordinates of the horizontal and vertical axes at the end of the robot's movement, and X(0) and Y(0) respectively represent the coordinates of the horizontal and vertical axes at the beginning of the robot's movement; abs represents the absolute value function.

[0063] S5: The active cable update control signal corresponding to the evaluation parameter satisfying the preset condition is used as the active cable target control signal; specifically, the active cable gait variable with the largest fitness in the population is selected as the final active cable gait variable, and then the time series signal of the active cable length change is obtained according to the robot movement gait table; the time series signal of the active cable length change is interpolated to obtain the control signal of each active cable length changing with time, so as to control the active cable and drive the passive cable to perform the gait action.

[0064] Furthermore, if fitness is used to represent the evaluation parameter, S5 includes: selecting the active search update control signal corresponding to the maximum fitness as the active search target control signal.

[0065] It should be noted that the active cable target control signal also needs to consider the maximum change speed of the active cable. If the control signal active cable change speed exceeds the maximum change speed that the active cable can achieve in the hardware, a new signal is regenerated through the optimization algorithm until the hardware can meet the actual length change speed.

[0066] The method for generating movement control signals of a tensegrity robot provided in this embodiment selects some cables as active cables from all cables through structural and force analysis, and the remaining cables are used as passive cables. The shape and posture of the robot are changed by changing the length of the active cables through motor drive to change the position of the center of gravity, thereby changing the pressure of the rod end on the ground to change the maximum friction between the rod end and the ground, and the robot moves on flat ground and uphill according to the set movement gait. The gait variables are iteratively optimized through an optimization algorithm, and the time series of the length change of each active cable is obtained according to the gait table, and then the length control signal of the active cable is generated by interpolation. Compared with the traditional unsupervised control algorithm, the method of the present invention generates feasible control signals faster and improves the movement speed and control accuracy of the three-rod and nine-cable prismatic tensegrity robot.

[0067] The following describes two movement modes to realize the generation process of active cable control signals corresponding to the straight movement of the tensegrity robot.

[0068] The first movement mode: When the movement behavior moves forward, the gait action includes: overall forward leaning and periodic first action combination. The first action combination includes: moving the rear two feet forward, moving the front foot forward and restoring the forward leaning. The movement effect simulation is as follows Figure 3 As shown, the following describes the generation process of the movement control signal generation of the tensegrity robot as follows.

[0069] First, determine the active cables and passive cables: According to the structural characteristics of the three-bar and nine-cable tensegrity robot, the active cables that need to be driven to change the tilt posture (forward, backward, etc.) are three waist cables, and the cables that need to move the feet are two bottom cables, a total of five active cables, and the remaining four are passive cables.

[0070] Then, the moving gait is designed according to the selected active cable: the complete gait cycle is the change from the initial neutral state to the overall forward leaning state, which does not belong to the cyclic gait, but is only a preparatory gait action for starting to move, followed by the forward moving rear two-foot gait action, followed by the forward moving front foot gait action, and finally the forward leaning is restored. When the three-bar nine-cable tensioned overall robot moves straight, these four gait actions, except for the first one which is only executed once, the remaining three gait cycles can make the robot move straight continuously. The designed gait table (Table 1) is as follows:

[0071] Table 1

[0072]

[0073] Furthermore, the initial parameters in the table are determined according to the designed gait and the initial length of the active cable. i The parameter range is set to [0,0.6], H i The parameter range of G is set to [-0.6, 0.6], iThe parameter range is set to [0,0.2]. The value of each parameter is randomly obtained within the parameter range as the initial value of the genetic algorithm iteration, and the parameters in the gait parameter table are used to interpolate to obtain a continuous active cable length change signal. The four time variables and eight length state variables in the gait are used as variables to be optimized in the genetic algorithm. The fitness function is fit=(X(end)-X(0))-abs(Y(end)-Y(0)), where X(0), Y(0), and Y(end) are the X and Y initial positions of the center of the triangle formed by the three rod ends of the three-rod nine-cable prismatic tensegrity robot in contact with the ground and the position after a complete gait. Among them, the physical structure model in the simulation program Simulink is exported through SolidWorks, and the motor adopts position voltage dual-loop PID control; finally, the priority algorithm is continuously iterated to obtain a better gait parameter result.

[0074] The second movement mode: When the movement behavior is forward, the gait action can also include: right forward leaning and periodic second action combination; the second action combination includes: moving the left foot forward, leaning forward to the left, moving the right foot forward, leaning forward as a whole, moving the front foot forward and restoring the right forward leaning. The movement effect simulation is as follows Figure 4 As shown, the following describes the generation process of the movement control signal generation of the tensegrity robot as follows:

[0075] First, determine the active cables and passive cables: According to the structural characteristics of the three-bar and nine-cable tensegrity robot, the active cables that need to be driven to change the tilt posture (forward, backward, etc.) are three waist cables, and the cables that need to move the feet are two bottom cables, a total of five active cables, and the remaining four are passive cables.

[0076] Then, the moving gait is designed according to the selected active cable: the designed gait complete cycle is the change from the initial neutral state to the right forward leaning state, which is not a cyclic gait, but only a preparatory gait action for starting to move, followed by the gait action of moving the left foot forward, and the next gait action is to lean left forward, move the right foot forward, lean forward as a whole, move the front foot forward, and finally restore the right forward leaning. When the three-bar nine-cable tensioning robot moves straight, these 7 gait actions, except for the first one which is only executed once, the remaining 6 gait cycles can make the robot move straight continuously. The designed gait table (Table 1) is as follows:

[0077] Table 2

[0078]

[0079] Then, the initial parameters in the table are determined according to the designed gait and the initial length of the active cable. i The parameter range is set to [0,0.6], Hi The parameter range is set to [-0.6, 0.6]. The parameters in the gait parameter table are interpolated to obtain a continuous active cable length change signal. The 7 time variables and 18 length state variables in the gait are used as variables to be optimized in the genetic algorithm, and the fitness function is: fit = (X(end)-X(0))-abs(Y(end)-Y(0)), where X(0), Y(0), X(end), and Y(end) are the X and Y initial positions of the center of the triangle formed by the three rod ends of the three-rod nine-cable prismatic tensegrity robot in contact with the ground and the position after the end of a complete gait. The physical structure model in Simulink is exported through SolidWorks, and the motor adopts position-voltage dual-loop PID control. Through continuous iteration of the genetic algorithm, a better gait parameter result is obtained.

[0080] Example 2

[0081] The present embodiment provides a mobile control signal generating device for a tensegrity robot, comprising the following modules. A selection module, for selecting a part of all cables of the tensegrity robot as active cables and the remaining parts as passive cables; a setting module, for determining each gait action in the mobile behavior of the tensegrity robot; setting the initial range of the gait variable corresponding to each gait action; generating the initial value of the gait variable of each active cable within the initial range; the gait variables include: a time variable t representing the time consumed to complete each gait action, an active cable length variable X representing the change in the current length of the active cable compared to the original length, an active cable extension variable H representing whether the active cable is extended or shortened, and a correction variable X for improving the deviation of the trajectory when the tensegrity robot moves straight. A positive variable G; an optimization module, used to input the initial value of the gait variable of each active cable into the optimization algorithm for iteration, to obtain the updated value of the gait variable of each active cable, and then input the updated value of the gait variable of the active cable into the optimization algorithm for repeated iteration until a specified number of iterations is reached; a calculation module, used to calculate the updated control signals of each active cable corresponding to the updated value of the gait variable of each active cable in each iteration and the corresponding evaluation parameters; a determination module, used to use the updated control signal of the active cable corresponding to the evaluation parameter satisfying the preset condition as the target control signal of the active cable, so as to control the active cable and drive the passive cable to perform the gait action.

[0082] Example 3

[0083] The present embodiment provides a movement control method of a tensegrity robot, comprising: obtaining an active cable target control signal based on a movement control signal generation method of the tensegrity robot; and controlling the tensegrity robot to perform corresponding gait movements using the active cable target control signal.

[0084] Example 4

[0085] This embodiment provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements a method for generating a movement control signal of a tensegrity robot or steps of a movement control method of a tensegrity robot when executing the computer program.

[0086] Example 5

[0087] The present embodiment provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, a method for generating a movement control signal of a tensegrity robot or steps of a movement control method of a tensegrity robot is implemented.

[0088] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for generating a motion control signal of a tensegrity robot, characterized in that: include: S1: Selecting some of the cables of the tensegrity robot as active cables and the remaining cables as passive cables; S2: determining each gait action in the movement behavior of the tensegrity robot; setting an initial range of a gait variable corresponding to each gait action; generating initial values ​​of the gait variables of each of the active cables within the initial range; The gait variables include: a time variable t representing the time taken to complete each gait action, an active cable length variable X representing the change of the current length of the active cable compared to the original length, an active cable extension variable H representing whether the active cable is extended or shortened, and a correction variable G for improving the trajectory deviation of the tensegrity robot when it moves straight; S3: inputting the initial value of the gait variable of each active cable into the optimization algorithm for iteration to obtain the updated value of the gait variable of each active cable, and then inputting the updated value of the gait variable of the active cable into the optimization algorithm for repeated iteration until a specified number of iterations is reached; S4: calculating each active cable update control signal and its corresponding evaluation parameter corresponding to the update value of the gait variable of each active cable in each iteration; S5: Using the active cable update control signal corresponding to the evaluation parameter satisfying the preset condition as the active cable target control signal to control the active cable and drive the passive cable to perform the gait action.

2. The method for generating a movement control signal of a tensegrity robot according to claim 1, characterized in that: In S1: the selection principle of the active cable is: select a waist cable that can change the tilt posture of the tensegrity robot and a bottom cable that is necessary to move the tensegrity robot to contact the ground as the active cable.

3. The method for generating a movement control signal of a tensegrity robot according to claim 1, characterized in that: The active cable length variable X, active cable extension variable H and correction variable G are all percentages; The product of the active cable length variable X and the corresponding original length of the active cable represents the change in the length of the active cable in the current round of iteration compared to the original length of the active cable; The amplitude of the active cable expansion and contraction variable H is the same as that of the active cable length variable X. The sign of the active cable expansion and contraction variable H is positive or negative. When it is positive, it indicates that the length is elongated relative to the initial moment, and when it is negative, it indicates that the length of the active cable is shortened relative to the initial moment.

4. The method for generating a movement control signal of a tensegrity robot according to claim 1, characterized in that: The S4 includes: Interpolating the time series of the length changes of each active cable corresponding to the updated value of the gait variable of each active cable in each iteration, to obtain an active cable update control signal corresponding to the length change of the active cable over time; The active cable update control signal is input into a simulation program to obtain the coordinates after the movement, and the coordinates after the movement and the coordinates before the movement are input into a fitness function for calculation to obtain the fitness representing the evaluation parameter corresponding to the active cable update control signal.

5. The method for generating a movement control signal of a tensegrity robot according to claim 4, characterized in that: The fitness function is fit=(X(end)-X(0))-abs(Y(end)-Y(0)), where X(end) and Y(end) respectively represent the coordinates of the horizontal and vertical axes at the end of the robot's movement, and X(0) and Y(0) respectively represent the coordinates of the horizontal and vertical axes at the beginning of the robot's movement; abs represents the absolute value function.

6. The method for generating a movement control signal of a tensegrity robot according to claim 5, characterized in that: The step S5 includes: selecting the active search update control signal corresponding to the maximum fitness as the active search target control signal.

7. The method for generating a movement control signal of a tensegrity robot according to claim 1, characterized in that: The optimization algorithm is a genetic algorithm, a particle swarm algorithm or an ant colony algorithm.

8. A device for generating a movement control signal of a tensegrity robot, characterized in that: include: A selection module, for selecting a portion of all cables of the tensegrity robot as active cables and the remaining portion as passive cables; A setting module, used to determine each gait action in the movement behavior of the tensegrity robot; and to set an initial range of a gait variable corresponding to each gait action; generating initial values ​​of the gait variables of each of the active cables within the initial range; The gait variables include: a time variable t representing the time taken to complete each gait action, an active cable length variable X representing the change of the current length of the active cable compared to the original length, an active cable extension variable H representing whether the active cable is extended or shortened, and a correction variable G for improving the trajectory deviation of the tensegrity robot when it moves straight; An optimization module, used for inputting the initial value of the gait variable of each active cable into the optimization algorithm for iteration, obtaining the updated value of the gait variable of each active cable, and then inputting the updated value of the gait variable of the active cable into the optimization algorithm for repeated iteration until a specified number of iterations is reached; A calculation module, used for calculating each active cable update control signal and its corresponding evaluation parameter corresponding to the update value of each gait variable of the active cable in each iteration; The determination module is used to use the active cable update control signal corresponding to the evaluation parameter satisfying the preset condition as the active cable target control signal to control the active cable and drive the passive cable to perform the gait action.

9. A method for controlling the movement of a tensegrity robot, characterized in that: include: Based on the method for generating a movement control signal of a tensegrity robot according to any one of claims 1 to 7, an active cable target control signal is obtained; The active cable target control signal is used to control the active cable of the tensegrity robot to extend and retract and drive the passive cable to move, so as to perform corresponding gait movements.

10. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the processor implements the steps of the method for generating a movement control signal of a tensegrity robot according to any one of claims 1 to 7 or the movement control method of a tensegrity robot according to claim 9.

Images