Robot control system, robot control method, and robot control apparatus
The robot control system addresses the challenge of controlling moving parts in diverse environments by setting allowable ranges and using model predictive control, ensuring stable and efficient robot operations despite numerous state combinations.
Patent Information
- Application Number
- JP2023080693
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-05-16
- Publication Date
- 2024-11-28
AI Technical Summary
Existing robot control methods struggle to stably control the moving section under diverse target motions in unprepared environments due to the vast number of arm and moving part state combinations, particularly failing to account for inertial forces during arm movements.
A robot control system that integrates a motion planning section, arm control unit, and movement control unit to generate appropriate control laws by setting allowable ranges for arm and moving part states, using model predictive control to adjust commands based on real-time sensor feedback and optimal control problems.
Enables efficient and stable control of robot movements in unprepared environments by reducing power consumption and joint wear, even with many arm and moving part state combinations, ensuring accurate target motions are achieved within predefined quality standards.
Smart Images

Figure 2024164946000001_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates to a robot control system, a robot control method, and a robot control device. [Background technology]
[0002] In recent years, labor shortages due to the declining birthrate and aging population, dangerous work in disaster zones, etc. have become social issues. To solve these issues, it is expected that robots will be used for various tasks in environments other than factories and other environments that are equipped so that robots can work safely and efficiently (hereinafter referred to as "unequipped environments").
[0003] In order for a robot with arms and moving parts to autonomously carry out tasks in unmaintained environments such as disaster zones, it is necessary to control the arms to skillfully manipulate the object being operated, as well as to control the movement appropriately to move close to the object being operated. Examples of robots that realize these include two-wheel drive carts equipped with manipulators, crawler robots, four-legged robots, humanoid robots, and unmanned aerial vehicles (UAVs).
[0004] As an example of this type of robot, the mobile manipulator described in Non-Patent Document 1 is known. Non-Patent Document 1 proposes a robot control method in which an arm controller and a base controller for a legged mobile manipulator are separately prepared, and a control rule (control command) for the moving part based on the behavior of the arms is acquired by deep reinforcement learning. [Prior art documents] [Non-patent literature]
[0005] [Non-Patent Document 1] Y. Ma, F. Farshidian, T. Miki, J. Lee and M. Hutter, "Combining Learning-based Locomotion Policy with Model-based Manipulation for Legged Mobile Manipulators," arXiv:2201.03871v1, 2022. Summary of the Invention [Problem to be solved by the invention]
[0006] The target actions that a robot can perform in an unmaintained environment are diverse, including actions that change the position or posture of an object, such as grasping, transporting, and placing an object, opening and closing a door, and operating a lever, as well as actions that process an object, such as welding, painting, cutting, and assembly.
[0007] In Non-Patent Document 1, since the arm controller and the movement controller are independent, when executing a new target motion for the arm, it is possible to respond by simply changing the control law of the arm controller. In addition, since the already established arm motion generation technology can be used to change the control law of the arm, it is possible to implement a new target motion relatively easily.
[0008] However, the control method of the moving unit in Non-Patent Document 1 has a problem that the moving unit can be stably controlled only under the conditions considered when learning the control law of the moving unit. Therefore, when there are many combinations of the arm state and the moving unit state when executing a target motion, it is difficult to obtain an appropriate control law for the moving unit. For example, when executing control to change the posture of the moving unit to expand the reach of the arm, the number of combinations of states that must be considered becomes enormous when the influence of the inertial force when the arm moves is included, so that it is difficult to obtain an appropriate control law for the moving unit.
[0009] Therefore, an object of the present invention is to provide a robot control system, a robot control method, and a robot control device that can easily obtain an appropriate moving part control law even when there are many combinations of arm states and moving part states to achieve a target movement. [Means for solving the problem]
[0010] In order to solve the above problems, the robot control system of the present invention is a robot control system having a robot having an arm and a moving unit, and a control device that controls the robot, wherein the control device is equipped with a motion planning unit that outputs a target motion for the movement of the arm and an acceptable range of the movement of the moving unit corresponding to the target motion, an arm control unit that outputs an arm control command and an arm state transition for the target motion, and a movement control unit that uses the arm state transition to generate a movement control command for the movement of the moving unit so that the movement falls within the acceptable range, and the robot is a robot control system in which the arm is controlled by the arm control command and the moving unit is controlled by the movement control command. Effect of the Invention
[0011] According to the robot control system, robot control method, and robot control device of the present invention, an appropriate moving part control law can be easily obtained even when there are many combinations of arm states and moving part states to achieve a target movement. [Brief description of the drawings]
[0012] [Figure 1] FIG. 1 is a diagram illustrating a configuration of a robot control system according to a first embodiment. [Figure 2A] FIG. 1 is a diagram illustrating a four-legged robot as an example of a robot. [Figure 2B] FIG. 1 is a diagram illustrating a humanoid robot as an example of a robot. [Figure 2C] FIG. 1 is a diagram illustrating an unmanned aerial vehicle as an example of a robot. [Diagram 3] FIG. 4 is a diagram for explaining a desired motion of an arm according to the first embodiment. [Figure 4]FIG. 2 is a diagram for explaining the state of the robot and the surrounding environment when the control according to the first embodiment is executed. [Diagram 5] 5A to 5C are diagrams illustrating changes in the position of the end effector of the arm over time when the control according to the first embodiment is performed. [Figure 6] FIG. 4 is a diagram illustrating the change over time in the amount of electric power output to a motor that drives a joint of a leg when the control according to the first embodiment is performed. [Figure 7] 10A to 10C are diagrams illustrating a desired motion of an arm according to the second embodiment. [Figure 8] FIG. 11 is a diagram illustrating the configuration of a robot control system according to a second embodiment. [Figure 9] FIG. 11 is a diagram for explaining the state of the robot and the surrounding environment when the control according to the second embodiment is executed. [Figure 10] FIG. 11 is a diagram illustrating the change over time in the moving speed of the base when the control according to the second embodiment is performed. [Figure 11] FIG. 11 is a diagram illustrating the change over time in the amount of electric power output to a motor that drives a joint of a leg when the control according to the second embodiment is performed. [Figure 12] FIG. 11 is a diagram illustrating the configuration of a robot control system according to a third embodiment. [Figure 13] 13A to 13C are diagrams for explaining a display example of a display unit according to a third embodiment. [Figure 14] FIG. 13 is a diagram for explaining the configuration of a robot control system according to a fourth embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0013] Hereinafter, examples of embodiments of the present invention will be described with reference to the accompanying drawings. The present invention is not limited to the following embodiments, and various numerical values in the embodiments are merely examples. In this specification and drawings, the same components or components having substantially the same functions are denoted by the same reference numerals, and duplicated descriptions will be omitted. EXAMPLES
[0014] FIG. 1 is a diagram for explaining the configuration of a robot control system 100 according to a first embodiment of the present invention. The robot control system 100 is a system having a control device 1, a robot 2, and an allowable range input unit 31. The control device 1 is composed of a motion planning unit 11, an arm control unit 12, and a movement control unit 13. Specifically, the control device 1 is a computer equipped with hardware such as a calculation device such as a CPU, a main storage device such as a semiconductor memory, an auxiliary storage device such as a hard disk, and a communication device. The calculation device executes a predetermined program to realize each functional unit such as the motion planning unit 11, but the following description will be given while appropriately omitting such well-known techniques. The robot 2 is composed of a manipulator (arm unit 21) that operates an object to be operated, and a movement unit 22 that is used when moving. The allowable range input unit 31 is a human-machine interface that is used by a user when inputting information required for robot control to the control device 1.
[0015] 2A is a diagram illustrating the configuration of a quadruped robot 2A, which is an example of the robot 2. This quadruped robot 2A has one six-joint arm 21 with an end effector 21a at its tip. In addition, the moving part 22 of the quadruped robot 2A has four three-joint legs 22a, and a base part 22b to which the arm 21 and the leg 22a are attached.
[0016] 2B is a diagram illustrating the configuration of a humanoid robot 2B, which is an example of the robot 2. This humanoid robot 2B has six-jointed arms 21 on the left and right, each arm having an end effector 21a at its tip. In addition, the moving part 22 of the humanoid robot 2B has two six-jointed legs 22a, and a base part 22b to which the arms 21 and the legs 22a are attached.
[0017] 2C is a diagram illustrating the configuration of an unmanned aerial vehicle 2C, which is an example of a robot 2. This unmanned aerial vehicle 2C has one six-joint arm 21 with an end effector 21a at its tip. In addition, the moving part 22 of the unmanned aerial vehicle 2C has four rotors 22c that generate lift and thrust, and a base part 22b to which the arm 21 and rotors 22c are attached.
[0018] The four-legged robot 2A and the humanoid robot 2B are provided with a drive mechanism for driving the arms 21 and the legs 22a at each joint. The unmanned aerial vehicle 2C is provided with a drive mechanism for driving the arms 21 at each joint, and a drive mechanism for controlling the attitude of the moving unit 22 is provided at each rotor 22c. In this case, the number of drive units of the arms 21 and the legs 22a may be less than the number of joints, and a mechanism may be used in which a single drive unit controls a plurality of joints. The number of joints of the arms 21 and the legs 22a, and the number of the arms 21, the legs 22a, and the rotors 22c are not limited to the examples of Figs. 2A to 2C and may be increased or decreased. In addition, there is no problem if other parts (for example, a tail, etc.) are provided. The mechanism for realizing the movement provided in the moving unit 22 is not limited to the legs 22a and the rotors 22c, and may be, for example, wheels or crawlers. In addition, the robot control system 100 of FIG. 1 can take into consideration both the case where the robot 2 exists in a real environment and the case where it exists in a virtual environment.
[0019] <Method of controlling arm portion 21> Here, a method for controlling the arm 21 of the robot 2 will be described. First, the motion planning unit 11 of the control device 1 commands the arm control unit 12 to execute a target motion by the arm 21. The target motion corresponds to any motion that the arm 21 can execute. For example, this includes moving an object to be operated (grasping, transporting, installing, operating, etc.), processing (painting, cutting, welding, assembling, wiping, etc.), and observation (instrument reading, surface inspection, appearance inspection, etc.).
[0020] The arm control unit 12 generates a series of movements to achieve a target movement based on the target movement commanded by the motion planning unit 11. Any method can be applied to generate the series of movements. For example, there are rule-based control methods based on a program created in advance, and learning-type control methods based on learned control rules.
[0021] The series of movements generated by the arm control unit 12 is aThe arm control command C1 is input to the arm 21, and the arm 21 operates based on the arm control command C1. The arm control command C1 is a command to rotate or translate each joint of the arm 21, and specifies, for example, time-series changes in the angle, angular velocity, and angular acceleration (torque) of each joint. When the arm control unit 12 generates an operation based on the difference between the arm control command C1 and the actual state of the arm 21 or the state of the surrounding environment, the arm control unit 12 can use information from various sensors equipped in the robot 2.
[0022] <Method of controlling the moving unit 22> Next, a method for controlling the moving unit 22 of the robot 2 will be described. In the case of a robot 2 equipped with an arm 21 and a moving unit 22, the movement control unit 13 of the control device 1 controls the moving unit 22 so as to follow a target state (for example, the position or posture of the base 22b) commanded by the motion planning unit 11 in order to achieve a target motion of the arm 21. At this time, the movement control unit 13 needs to control the moving unit 22 taking into consideration the effect of inertial force caused by the motion of the arm 21.
[0023] In order to control the moving part 22 so as to follow the target state, it is necessary to obtain a control input that reduces the difference between the actual state of the moving part 22 and the target state. If the problem is set so that the control input is as small as possible, then in a given evaluation period [t0, t f ], the following optimal control problem is obtained:
[0024]
number
[0025]
number
[0026]
number
[0027] Here, φ and L are evaluation functions, x is a state vector, u is an input vector, x ref is the target state vector, and P, Q, and R are weight matrices. x and x ref By including a variable related to the state of the moving part 22 in the target state vector, it is possible to realize control to track the target state. In addition, the terms defined in the evaluation function L are not limited to those in Equation 2. For example, a term to evaluate the amount of power consumption, or a term to evaluate the safety of the robot 2 to avoid collisions with itself or the surrounding environment can be added. ref The target trajectory of the state vector x can be defined by the trajectory of the time series change of
[0028] In order to solve the optimal control problem of Equations 1 to 3, it is necessary to estimate the time evolution of the state vector x. Therefore, the following constraints can be defined for the optimal control problem of Equations 1 to 3.
[0029]
number
[0030]
number
[0031]
number
[0032]
number
[0033] Here, equation 4 is the dynamics model of the robot 2, equation 5 is the initial condition, equation 6 is the equality constraint, and equation 7 is the inequality constraint.
[0034] If the optimal control problems of Equations 1 to 7 are solved offline before starting control of the robot 2, an optimized series of control inputs can be obtained in advance. On the other hand, when controlling the robot 2 in an unmaintained environment, the state transition may not necessarily occur as expected offline due to errors in the dynamics model and the effects of disturbances.
[0035] Therefore, in this embodiment, the movement control unit 13 performs the control period T l The robot state observed at that time is set as the initial condition of Equation 5, and the robot state is calculated as the time t0 from the current time t0 to the finite time ahead. f We decided to use a control method in which the optimal control problems of equations 1 to 7 are solved for up to t0, and only the result at t0 among the obtained control inputs is used as the control command C. This method is generally called model predictive control.
[0036] In the following embodiment, a description will be given assuming a case where a general model predictive control is used, but it is also possible to replace it with another control method related to the optimal control problem of formulas 1 to 7. For example, when the error of the dynamics model or the influence of disturbance is infinitesimally small, or when the robot 2 is controlled on a simulator, a series of control inputs can be determined from the results of solving the optimal control problem of formulas 1 to 7 offline. Alternatively, a hierarchical calculation method for calculating the force and torque generated at the center of gravity of the robot 2 and then calculating the control inputs to be output to each joint or rotor 22c, or a method for acquiring some of the parameters (weight matrix, evaluation period, target state, etc.) in the optimal control problem of formulas 1 to 7 by machine learning can also be adopted. In addition, the model predictive control includes a gradient-based method using differential dynamic programming, multiple shooting, etc., and a sampling-based method using the Monte Carlo method, path integration, etc., and any of these methods may be used. In addition, the time t f The value of is arbitrary, and by setting a large value, it is possible to calculate the control command C taking into consideration a more future state. On the other hand, the control period T l About f =t0+T land calculate the control command C by taking into consideration only the next control input to the robot 2 and the resulting state change.
[0037] In the following, as a specific example, a case will be described in which the optimal control problems of Equations 1 to 7 are applied to a four-legged robot 2A. When the four-legged robot 2A is the control target, the dynamics model in Equation 4 can be expressed as follows.
[0038]
number
[0039]
number
[0040] Here, p com , l com are the translational momentum and angular momentum of the center of gravity of the quadruped robot 2A, and the position x b , rotation θ b , translational velocity v b , rotational speed ω b , the joint angle θ of the leg 22a and the arm 21 l and θ a , joint angular velocity ω l and ω a It is a function that depends on n l and n a is the number of contact points between the legs 22a and the arms 21, r l,i and r a,i is the length from the center of gravity to the contact point i of the leg 22a and the arm 21, f l,i and f a,i is the contact force generated at the contact point i between the leg 22a and the arm 21, and τ l,i and τ a,i is the contact torque generated at the contact point i between the leg 22a and the arm 21, and mg is the force generated at the center of gravity by gravity.
[0041] The method of describing the dynamics model is not limited to formulas 8 and 9. For example, in order to describe the dynamics model in more detail, terms related to centrifugal force and Coriolis force may be added to formula 8, or terms related to the inertial force of leg 22a may be omitted in order to reduce the amount of calculation.
[0042] Furthermore, when the quadruped robot 2A is the object to be controlled, the equality constraints and inequality constraints in Equations 6 and 7 can be given as follows:
[0043]
number
[0044]
number
[0045]
number
[0046] where v l,i is the velocity of the contact point i of the leg 22a, f l,i x , f l,i y , f l,i z are the forces acting at contact point i in the tangential directions x and y and normal direction z with respect to the contact surface. Equations 10 and 11 define the conditions for contact point i not to slip on the contact surface. Furthermore, the equality constraint of equation 6 and the inequality constraint of equation 7 can be added or deleted depending on the performance and target motion of the four-legged robot 2A. For example, upper and lower limits of the joint angles and joint angular velocities of leg 22a can be added as inequality constraints.
[0047] In the above-described case where the quadruped robot 2A is the control target, the dynamics model of the equations 8 and 9 and the equality constraints and inequality constraints of the equations 10 to 12 are used to solve the optimal control problem of the equations 1 to 7, whereby the state x b , v b , θb , ω b Either of the following is the goal state x b,ref , v b,ref , θ b,ref , ω b,ref The joint angle θ of the leg 22a follows the l (or joint angular velocity ω l , joint angular acceleration α l ) and output it to the movement unit 22 as a movement control command C2.
[0048] Here, when model predictive control based on the optimal control problem of Equations 1 to 7 is performed in a real environment, differences occur between predicted values and actual measured values of the state of the quadruped robot 2A and the surrounding environment due to the influence of the dynamics model and disturbances. Therefore, sensor information provided in the quadruped robot 2A (for example, joint states acquired from encoders provided in the joints of the legs 22a, and posture information acquired from an inertial measurement unit provided in the base 22b) can be fed back to the movement control unit 13 and used as a correction of the control command C and an initial condition of Equation 5.
[0049] Furthermore, the joint angle θ of the arm 21 appearing in the dynamic model of Equations 8 and 9 a (or joint angular velocity ω a , joint angular acceleration α a ) and the arm state transition θ generated by the arm control unit 12 a * (or joint angular velocity transition ω a * , joint angular acceleration transition α a * ) or by adding equations 13, 14, and 15 as equality constraints of equation 6 in the optimal control problem of equations 1 to 7, it is possible to generate a movement control command C2 that takes into account the effect of the inertial force of the arm 21.
[0050]
number
[0051]
number
[0052]
number
[0053] Here, any one of Equation 13, Equation 14, and Equation 15 may be omitted. If a load is applied to the end effector 21a of the arm section 21, an equality constraint regarding the force fa acting on the end effector 21a may be added. In this case, in the process in which the arm control section 12 calculates the arm control command C1, for example, when the arm control section 12 uses model predictive control or when it uses learning control based on prediction of the environmental state, the arm control section 12 may calculate the arm control command C1 in the evaluation period [t0,t f ] arm state transition θ a * If it is possible to obtain the arm state transition θ a * can be output directly to the movement control unit 13. On the other hand, in the process in which the arm control unit 12 calculates the arm control command C1, f ] arm state transition θ a * If the past control command C is not acquired, P Refer to the future Control Directive C F By extrapolating, the arm state transition θ a * can be estimated.
[0054] In the optimal control problem of equations 1 to 7, the state variable x that appears in equation 2 i Weight Q i If you set it large, the state variable x i The goal state x i,refIt is possible to improve the tracking ability to the target state. On the other hand, in an optimal control problem based on set parameters (weight matrix, evaluation period, target state, etc.), a solution for making the state of the moving unit 22 follow the target state may not be calculated. In that case, the state of the moving unit 22 may deviate from the target state, causing the target motion of the arm 21 to fail, or the moving unit 22 to fall over because it cannot maintain a stable posture. Alternatively, even if the state of the moving unit 22 can follow the target state, a large burden may be placed on the joints of the leg 22a in order for the moving unit 22 to maintain the target state, which may consume excessive power or accelerate deterioration of the drive unit. A method of adjusting parameters in advance so that the state of the moving unit 22 can follow the target state and the load on the drive unit is reduced may also be considered.
[0055] However, in situations where complete information on the state of the object to be operated and the surrounding environment cannot be obtained in advance, or where these states change dynamically, it is difficult to prepare parameters that are effective for all situations in advance. Moreover, it is not realistic to calculate such parameters online in real time while controlling the robot 2.
[0056] In order to solve this problem, in this embodiment, the following method is proposed for setting in advance allowable ranges for state variables in the optimal control problem of Equations 1 to 7 based on the characteristics of the target movement of the arm 21.
[0057] First of all, it is not essential that the state of the moving unit 22 strictly follows the target state in order to execute a target operation of the arm 21. For example, when the target operation is processing an object to be operated and the required quality is satisfied if the processing result falls within a certain range, it is sufficient to control the moving unit 22 so that the processing can be performed within that range. Also, when the arm control unit 12 controls the arm 21 based on feedback of information from a sensor equipped in the robot 2, as long as the state of the moving unit 22 falls within a predetermined range, it is sufficient to correct the arm control command C1 based on the difference between the actual state and the target state of the moving unit 22 to achieve the target operation.
[0058] Therefore, by considering the characteristics of the target movement of the arm 21, it is possible to set in advance the allowable range for the state variables to achieve the target movement of the arm 21. This makes it possible to select the state of the moving unit 22 within the range that satisfies the allowable range, increasing the probability of calculating a solution to achieve the target movement of the arm 21. Furthermore, it is expected that the amount of power consumption and the load on the drive unit when controlling the moving unit 22 can be reduced.
[0059] There are two methods for setting the allowable range of the state variable x in order to achieve the target motion of the arm 21, for example.
[0060] The first method of setting the allowable range is to add the following equation as an inequality constraint of equation 7 in the optimal control problem of equations 1 to 7.
[0061]
number
[0062] where x i is the state variable that sets the tolerance, and x min and x max is the x required to achieve the movement goal of the arm 21 i In this case, the state variable x i may be a variable other than the state of the moving unit 22. For example, the state variable x i Alternatively, the position or posture of the contact point of the arm 21 may be set as follows. Also, the allowable range of the formula 13 may be set for multiple state variables. In this case, the lower limit value x min and upper limit x max may be input by the user via the tolerance input unit 31, or may be automatically generated by the motion planning unit 11 using machine learning or the like.
[0063] The second method for setting the allowable range is to adjust the weight matrix in the evaluation function of Equation 2 in the optimal control problem of Equations 1 to 7. For example, i Weight Q i By reducing the target state xi,ref Therefore, we can reduce the tracking ability of Q so that Equation 13 is satisfied when solving the optimal control problem of Equations 1 to 7. i By setting the weight Q, it is possible to calculate the movement control command C2 that can achieve the target movement of the arm 21. i may be input by the user via the tolerance input unit 31, or may be automatically generated by the operation planning unit 11 using machine learning or the like.
[0064] It is also possible to combine these two methods of setting the tolerance. For example, if we set the tolerance of Equation 13 to x i After explicitly setting i The ease of change of Q i It is possible to adjust it using
[0065] In this embodiment, a specific example has been described for a quadruped robot 2A. However, by replacing the dynamics model of equation 4 and the equality and inequality constraints of equations 6 and 7 according to the form of the robot 2, the present invention can also be extended to cases where the robot 2 is a humanoid robot 2B or an unmanned aerial vehicle 2C.
[0066] For example, when controlling the humanoid robot 2B, the number of contact points of the leg 22a and the arm 21 is set to n l and n a and treating the base 22b and the head collectively as the trunk, the dynamics models of Equations 8 and 9 and the constraints of Equations 10, 11, and 12 can be used as they are.
[0067] In addition, when controlling the unmanned aerial vehicle 2C, the dynamic model of equation 4 can be expressed as a translational motion equation in the inertial coordinate system (equation 17) and a rotational motion equation in the aircraft coordinate system with the center of gravity of the unmanned aerial vehicle 2C as the origin (equation 18).
[0068]
number
[0069]
number
[0070] Here, p com , l com are the translational momentum and angular momentum of the center of gravity of the unmanned aerial vehicle 2C, and the position x b , rotation angle θ b , translational velocity v b , rotational speed ω b , the joint angle θ of the arm 21 a , joint angular velocity ω a It is a function that depends on n a is the number of contact points of the base 22b, r a,i is the length from the center of gravity to the contact point i of the arm 21, f a,i is the contact force generated at the contact point i of the arm 21, and τ a,i is the contact torque generated at the contact point i of the arm 21, mg is the force generated at the center of gravity by gravity, R is the rotation matrix for converting from the aircraft coordinate system to the inertial coordinate system, F w and τ w are the thrust and torque generated in the unmanned aerial vehicle 2C by the rotation of the rotor 22c. w and torque τ w is the rotation speed ω of the rotor 22c w It is proportional to the square of .
[0071] The method of describing the dynamics model of the unmanned aerial vehicle 2C is not limited to Equation 17 and Equation 18. For example, a term relating to air resistance can be added to the translational equation of Equation 17, or a term relating to the viscous resistance of air can be added to the rotational equation of Equation 18. Also, the constraints in Equation 6 and Equation 7 do not necessarily need to be given, but for example, the rotation speed ω w The upper bound on can be given as an inequality constraint in Eq.
[0072] As described above, by solving the optimal control problems of Equations 1 to 7 using the dynamic models of Equations 17 and 18, the state x b , v b , θ b , ω b Either of the following is the goal state x b,ref, v b,ref , θ b,ref , ω b,ref The rotation speed ω of the rotor 22c of the unmanned aerial vehicle 2C is set to follow the w (or the thrust of each rotor 22c, the angular velocity of each rotor 22c, etc.) can be calculated and output to the moving part 22 as a moving part control command.
[0073] <Specific operation of the robot control system according to this embodiment> Next, a specific operation of the robot control system 100 when the robot 2 is a four-legged robot 2A will be described with reference to FIGS.
[0074] 3 is a diagram illustrating a target motion in which the robot 2 (four-legged robot 2A) paints a painting area Wa on a wall surface W with a paint sprayer S held by an arm 21 while moving horizontally (in the x direction) relative to the wall surface W. Below, using the target motion in FIG. 3 as an example, a method for determining an allowable range for achieving the target motion and a method for setting the allowable range as an inequality in Equation 13 in the optimal control problem of Equations 1 to 7 will be described.
[0075] In order to achieve the target motion shown in Fig. 3, the arm control section 12 of the control device 1 first issues a control command to the arm 21 of the robot 2 so that the joints Ja to Jd and Jf of the arm 21 maintain the initial state shown in Fig. 3. For the remaining joint Je, a control command is issued based on feedback of sensor information (posture information of the base 22b, visual sensor information, etc.) so that the spray direction of the paint spray S is always oriented vertically (y-axis direction) relative to the wall W. This target motion can be achieved by moving the moving section 22 in the horizontal direction (x direction) relative to the wall W under arm control command C1, which is an integration of these control commands. Here, if the distance from the nozzle of the paint spray S to the wall W is within a predetermined range, the sprayed paint will not spread and the required quality of this target motion will be satisfied.
[0076] In order to satisfy the required quality of this target motion, the posture (x bx ,x by ,x bz ,θ bx ,θby ,θ bz ), (x bx,ref ,x by,ref ,x bz,ref ,θ bx,ref ,θ by,ref ,θ bz,ref )=(v x,0 t, x by,0 ,x bz,0 ,θ bx,0 ,θ by,0 ,θ bz,0 ) can be set as the target posture. In this case, x bx,ref The base 22b moves at a constant speed v x0 represents the target posture to move in the x direction, and the other postures to maintain the initial values.
[0077] In addition, in order to satisfy the required quality of the target operation of painting, the position x of the end effector 21a of the arm 21 in the y direction is ay For the inequality constraint of Eq. 13, ay,min ≦x ay ≦x ay,max Here, if the required quality is satisfied, the inequality constraint of Equation 13 may be set for other state variables such as the position or orientation of the base 22b. In addition, if the deviation of the paint in the vertical direction (z direction) of the wall surface W is acceptable, the position x az For the inequality constraint x az,min ≦x az ≦x az,max Alternatively, you can set x ay The paint diffusion due to the change in also affects the deviation of the paint from the vertical direction of the wall W, so we set the inequality constraint x ay-z,min ≦Ax ay +Bx az ≦x ay-z,max It is also possible to set, where A and B are proportional constants.
[0078] In the configuration of the robot control system 100 in FIG. 1, the procedure for achieving this target motion is as follows.
[0079] First, the motion planning unit 11 commands the arm control unit 12 to execute the desired motion. At the same time, the motion planning unit 11 outputs the desired posture with respect to the base unit 22b and the allowable range with respect to the position of the end effector 21a of the arm unit 21 to the movement control unit 13.
[0080] Next, the arm control unit 12 issues an arm control command C1 for achieving the target motion at a control period T a At the same time, the arm control unit 12 outputs to the movement control unit 13 the arm state transition θ a * In the case of this target motion, information on the state of the joint J of the arm 21, which does not change with time, is output as the arm state transition θ a * At this time, the angle of the joint Je of the arm 21 is actually controlled based on the feedback of the sensor information, but it is assumed that the change in the inertial force due to the change in the angle of the joint Je is small and the effect on the dynamics of the moving part 22 is small.
[0081] Then, the movement control unit 13 calculates the arm state transition θ using model predictive control based on the optimal control problem of Equations 1 to 7. a * Taking the information from the above into consideration, a movement control command C2 is calculated so as to follow the target posture of the base portion 22b, and the control period T b In this way, the moving unit 22 outputs the arm state transition θ a * The system will be appropriately controlled based on the information provided by the operator.
[0082] 4 to 6 are diagrams for explaining the control results when the robot 2 is controlled to execute the target motion of FIG. 3. FIG. 4 is a schematic diagram of the robot 2 and the surrounding environment, and FIG. 5 is a diagram for explaining the control results when the robot 2 is controlled to execute the target motion of FIG. 3. ay FIG. 6 shows the change over time in the amount of electric power output to the motor that drives the joint of the leg 22a.
[0083] <<When this invention is not applied (when the tolerance range is not considered)>> First, the movement control unit 13 of the control device 1 detects the position x of the end effector 21a. ay The robot control result when the allowable range for is not taken into consideration will be described. In this case, the posture of the base 22b of the robot 2 follows the target posture, so that the base 22b continues to maintain the initial posture as shown in the left diagram of FIG. 4. As a result, as shown in the result 5a when the allowable range is not taken into consideration in FIG. 5, the position x ay can be kept within the tolerance of 5c.
[0084] However, in the state shown in the left diagram of Fig. 4, the center of gravity of the arm 21 of the robot 2 is located in the positive direction of the y-axis from the center of gravity of the moving part 22, and a moment that rotates the moving part 22 around the x-axis is generated due to the arm 21 in this state. Therefore, if the posture of the base part 22b is maintained as the target posture while resisting this moment, the amount of power consumed by the motor driving the joint of the leg part 22a increases, as shown in the result 6a in Fig. 6 when the allowable range is not considered. In this way, if the robot 2 is controlled without considering the allowable range of the end effector position, there is a risk that the robot 2 will consume excessive power and that the joint of the leg part 22a will deteriorate more quickly.
[0085] <<When this invention is applied (when the tolerance range is taken into consideration)>> Next, the movement control unit 13 of the control device 1 detects the position x of the end effector 21a. ay The control result when the allowable range for is taken into consideration will be described. In this case, as shown in the right diagram of FIG. 4, when the posture of the base part 22b of the robot 2 is θ bx 5. However, as shown in the result 5b when the tolerance range in FIG. 5 is taken into account, the position x ay has been able to stay within the acceptable range of 5c.
[0086] At this time, as shown in the right diagram of FIG. 4, the posture of the base part 22b of the robot 2 is set to θ bxBy rotating in the x-axis direction, the center of gravity of the arm 21 on the y-axis can be made to substantially coincide with the center of gravity of the moving part 22. This makes it possible to suppress the moment about the x-axis caused by the arm 21, and therefore to reduce the amount of power consumed by the motor driving the joint of the leg 22a, as shown in result 6b in FIG. 6 when the allowable range is taken into consideration. As a result, the amount of power consumed to achieve the target motion and the load on the joint of the leg 22a can be reduced.
[0087] 3 to 6, spray painting is given as the target motion, but the allowable range can be set in a similar manner even if the target motion is changed to painting using a roller or brush, polishing, welding, wiping, etc. In this case, in the case of polishing or wiping, the end effector 21a of the arm 21 needs to come into contact with the wall W, so the position x in the normal direction (y direction) of the wall W is ay In addition, when an external force is applied to the end effector 21a of the arm 21 due to contact with the wall surface W, the external force is acquired by a force sensor or the like provided on the arm 21, and the dynamics model f a , it is possible to control the moving part taking into consideration the external force acting on the arm part 21.
[0088] As described above, according to the robot control system of this embodiment, by setting the allowable range of the state variables in advance, it is possible to easily obtain an appropriate moving part control law even when there are many combinations of arm states and moving part states to achieve a target movement. EXAMPLES
[0089] Second Embodiment Next, a robot control system 100 according to a second embodiment of the present invention will be described with reference to Fig. 7 to Fig. 11. Note that a duplicated description of points common to the first embodiment will be omitted.
[0090] FIG. 7 shows a diagram of a four-legged robot 2A moving from a position x s When the object moves along the x-axis from the target object Ob, the object moves to the position x m1 is a diagram for explaining a target motion in which the arm 21 grasps an operation target object Ob. In this embodiment, a method for determining the allowable range for achieving this target motion and a method for setting the allowable range as the inequality of Equation 13 in the optimal control problem of Equations 1 to 7 will be explained for a case in which the allowable range changes based on the state of the arm 21, using this target motion as a specific example.
[0091] In order to achieve this target motion, the arm control unit 12 first recognizes the state of the operation object Ob from image information I1 of the surrounding environment captured by the camera 23 provided in the robot 2. Next, based on time-series information on the operation object Ob, the arm control unit 12 predicts the future trajectory of the operation object Ob and generates a target trajectory of the arm 21 for gripping the operation object Ob. This is equivalent to a control method of the arm 21 for gripping the operation object Ob moving on a belt conveyer or the like. Therefore, by moving the moving unit 22 connected to the arm 21 instead of the operation object Ob moving, the existing control method for the arm 21 can be used as it is as a control method in the arm control unit 12.
[0092] Here, the arm control unit 12 is assumed to be able to evaluate whether the arm 21 can grasp the operation object Ob when the moving unit 22 moves from the current position at a certain speed, based on the time-series image information I1 of the camera 23 capturing the operation object Ob. In this case, the upper limit v of the speed of the moving unit 22 for the arm 21 to grasp the operation object Ob is bx,max At this time, the distance r between the end effector 21a of the arm 21 and the operation target object Ob can be determined. a-o As the object Ob approaches, the end effector 21a of the arm 21 cannot anticipate the predicted trajectory of the object Ob. bx,max Thus, v bx,max is a-o Since the tolerance range for achieving this target operation varies depending on r a-o The value of the saturation voltage will vary based on the following:
[0093] As a result, the attitude (x bx ,x by ,x bz ,θbx ,θ by ,θ bz The target attitude at time t+Δt for (x bx,ref ,x by,ref ,x bz,ref ,θ bx,ref ,θ by,ref ,θ bz,ref )=(x bx,t +v bx,max (t)Δt, x by,0 ,x bz,0 ,θ bx,0 ,θ by,0 ,θ bz,0 ), where x bx,t is the position of the base 22b on the x-axis at time t, v bx,max (t) is the upper limit of the speed of the moving part 22 at time t. In addition, the allowable range for the arm 21 to grasp the operation object Ob is the translational speed v bx For the inequality constraint in Eq. 13, v bx,min ≦v bx ≦v bx,max Here, v bx,min is v bx,max We can set an arbitrary speed smaller than v bx,min =v bx,max Set / 2.
[0094] Fig. 8 is a diagram for explaining the configuration of the robot control system 100 for achieving this desired motion. In the configuration of the robot control system 100 in Fig. 8, the procedure for achieving this desired motion is as follows.
[0095] First, the motion planning unit 11 commands the arm control unit 12 to execute the target motion. Next, the arm control unit 12 generates an arm control command C1 based on image information I1 captured by the camera 23, and calculates the control period T a At the same time, the arm control unit 12 outputs the arm state transition θ a * And the tolerance based on that v bx,max and generate arm state transition θ a *The arm state transition θ a * can be generated by predicting a target trajectory that the arm 21 should take in the future, for example, assuming that the operation target object Ob moves at a constant speed.
[0096] Next, the motion planning unit 11 determines a target posture for the posture of the base portion 22b and v bx The controller 13 then uses model predictive control based on the optimal control problem of Equations 1 to 7 to calculate the arm state transition θ a * Taking the information into consideration, a movement control command C2 is calculated so as to follow the target posture of the base portion 22b, and the control period T b At this time, if the movement control command C2 changes, the position of the operation target object Ob captured by the camera 23 also changes, so the arm control command C1 generated by the arm control unit 12 is corrected. Therefore, in the process in which the movement control unit 13 calculates the movement control command C2 based on the optimal control problem of the equations 1 to 7, the arm state transition θ a * will change.
[0097] Accurate arm state transition θ a * In order to take into account the above, the arm state transition θ a * For example, the arm state can be corrected by changing the distance r between the end effector 21a of the arm 21 and the operation target object Ob. a-o By approximating the arm state transition θ a * On the other hand, when using model predictive control, the updated arm state transition θ a * The calculation of the movement control command C2 based on the control period T b Since it is executed in , the arm state transition θ a *Even if the position of the moving control command C2 is not accurate to a certain extent, it is possible to calculate a moving control command C2 that does not deviate significantly from the optimal solution.
[0098] Here, the control period T a and the control period T of the movement control unit 13 b If different, arm state transition θ a * The update period of the tolerance range transition is T a、 T b The period during which the motion planning unit 11 outputs the target posture and the allowable range to the movement control unit 13 is T a Or T b The period may be the same as or different from that.
[0099] There may also be a case where the distance between the robot 2 and the operation object Ob is large and the arm control unit 12 cannot execute the target motion. In that case, a method can be adopted in which image information I1 from the camera 23 is fed back to the motion planning unit 11, and a command for the target motion is started to the arm control unit 12 at the timing when the distance between the robot 2 and the operation object Ob becomes smaller than a certain reference value. At this time, until the command for the target motion is started to the arm control unit 12, a command for a target position or target posture to move at a certain speed toward the operation object Ob can be given to the movement control unit 13.
[0100] 9, 10, and 11 are diagrams for explaining the control results when the robot 2 is controlled for the target motion. FIG. 9 is a schematic diagram of the robot 2 and the surrounding environment, and FIG. 10 shows the translation velocity v bx 11 shows the transition of the amount of electric power output to the motor that drives the joint of the leg 22a.
[0101] <<When this invention is not applied (when the tolerance range is not considered)>> First, the movement control unit 13 of the control device 1 detects the translation speed v of the base portion 22b. bx A control result 10a when the allowable range for the position x is not taken into consideration will be described. As shown in the left diagram of FIG. 9, the posture of the base part 22b of the robot 2 follows the target posture. mThe arm 21 can grasp the object to be operated. On the other hand, when the arm 21 of the robot 2 approaches the object to be operated Ob, the position of the center of gravity of the robot 2 moves in the negative direction with respect to the y-axis. As a result, in order to make the posture of the base 22b follow the target posture, the amount of power output to the motor that drives the joint of the leg 22a increases, as shown in the result 11a in FIG. 11 when the allowable range is not taken into consideration. Therefore, there is a risk that excessive power will be consumed and that deterioration of the joint of the leg 22a will be accelerated.
[0102] <<When this invention is applied (when the tolerance range is taken into consideration)>> Next, the movement control unit 13 of the control device 1 calculates the translation speed v bx A control result 10b when the allowable range for the moving speed v of the base portion 22b is taken into consideration will be described. bx In addition to the tolerance for the z-direction position x bz Tolerance for x bz,min ≦x bz ≦x bz,max is given. Compared to the case where the allowable range is not taken into consideration in the left diagram of FIG. 9, when the allowable range is taken into consideration in the right diagram of FIG. 9, the posture of base 22b moves in the negative direction with respect to the z axis. As a result, the position of the center of gravity of arm 21 moves in the positive direction with respect to the y axis, and the amount of power output to the motor that drives the joint of leg 22a decreases, as shown in result 11b when the allowable range is taken into consideration in FIG. 11. As a result, the amount of power consumption and the load on the joint of leg 22a can be reduced. At this time, when the position x bz In order to walk while keeping the moving speed v low, as shown in Figure 10b when considering the tolerance range. bx Although it is smaller, it is still within the allowable range of 10c.
[0103] In this embodiment, the gripping of a stationary object is taken as an example of the target motion, but the allowable range can be set in a similar manner even if the target motion is changed to gripping a dynamic object, placing an object to be gripped, pressing a switch, etc. In addition, when an external force is generated on the end effector 21a of the arm 21 due to gripping the operation target object Ob or contact with the surrounding environment, the external force is measured by a force sensor provided in the arm 21, and the force f appearing in the dynamics model of Equation 9 is calculated. a By setting it to, it is possible to control the moving part taking into consideration the external force acting on the arm part 21. Alternatively, by describing the state transition of the operation target object Ob and the surrounding environment in the dynamics model of Equation 9, their behavior can be taken into consideration more accurately.
[0104] Furthermore, in the system of this embodiment equipped with the camera 23, when formulas 8 to 12 are used, information from the camera 23 is fed back to the movement control unit 13, and obstacles and terrain in the surrounding environment are recognized, thereby making it possible to determine the target state of the base 22b and the target position of the contact point of the leg 22a. EXAMPLES
[0105] Next, a robot control system 100 according to a third embodiment of the present invention will be described with reference to Fig. 12 and Fig. 13. In this embodiment, a system having a support function for determining an allowable range for achieving a target motion of the arm 21 will be described. Note that a duplicated description of points common to the above embodiments will be omitted.
[0106] 12 is a diagram for explaining the configuration of a robot control system 100 of this embodiment. This robot control system 100 has a configuration in which a sensor 24, a calculation unit 32, and a display unit 33 are added to the configuration of the robot control system of FIG.
[0107] The allowable range input unit 31 receives information about the allowable range input by the user via a keyboard, a touch panel, a joystick, or the like, and outputs the information about the allowable range to the motion planning unit 11. Specifically, the information is input as an inequality constraint of Equation 13 for a certain state variable, or a weighting matrix in the optimal control problem of Equations 1 to 3. At this time, the value input as the upper and lower limits in the inequality of Equation 13 may be a constant, like the allowable range for the target motion in the first embodiment, or may be a function that depends on a certain variable.
[0108] The calculation unit 32 obtains the results of controlling the robot 2 based on the input information on the allowable range from information from the sensor 24, and calculates information to be presented to the user based on that information. Note that the calculation unit 32 is a functional unit realized by a computer separate from the computer constituting the control device 1, or a functional unit realized by the same computer as the control device 1.
[0109] The display unit 33 is a display device such as a liquid crystal display that displays information calculated by the calculation unit 32. Fig. 13 shows an example of information displayed on the display unit 33. Here, Fig. 13 assumes a case where the robot 2 (four-legged robot 2A) performs the target motion shown in Fig. 3. The user can check a simulation video 33a of the robot 2 operating within a set tolerance range 33b, a time-dependent change 33f of a state variable for which a tolerance range is set, and a time-dependent change 33g of power consumption.
[0110] If the state variable for which the tolerance is set does not fall within the tolerance 33b, if the robot 2 falls while executing the target motion, or if the robot 2 runs out of charge before completing the target motion, NG is displayed in the target motion success / failure 33d, and the user can correct the tolerance 33b based on the information. Also, based on the information on the torso motion speed 33c and the predicted completion time 33e of the target motion displayed on the display unit 33, it is also possible to correct the tolerance or target state regarding the movement speed 33c of the base 22b output from the motion planning unit 11 to the movement unit control unit 13. In this case, the configuration of the display unit 33 is not limited to that shown in FIG. 13, and it is also possible to additionally display control command values for the arm 21 and the movement unit 22, omit displaying any of the information shown in FIG. 13, or add an interface that allows the user to select the information to be displayed. Also, the calculation unit 32 can calculate an appropriate tolerance based on the information from the sensor 24 and present it on the display unit 33. Alternatively, the calculated tolerance can be directly output to the motion planning unit 11.
[0111] The configuration of the robot control system 100 in FIG. 12 can be applied to both cases where the robot 2 is controlled in a real environment and where it is controlled in a virtual environment. When the robot 2 is controlled in a real environment, a more realistic tolerance range can be verified for operation using a real machine. On the other hand, when the environmental conditions in which the robot 2 is actually operated are uncertain, it takes a lot of effort to comprehensively evaluate the uncertain environmental conditions. In addition, if the target operation fails, the robot 2 may interfere with the surrounding environment and be damaged. On the other hand, when the robot 2 is controlled in a virtual environment, it is possible to determine information regarding the tolerance range after evaluating robustness against various environmental conditions without using a real machine. EXAMPLES
[0112] Next, a robot control system 100 according to a fourth embodiment of the present invention will be described with reference to Fig. 14. In this embodiment, a case where the arm 21 is controlled by a user will be described. Note that a duplicated description of points common to the above embodiments will be omitted.
[0113] 14 is a diagram for explaining the configuration of the robot control system 100 of this embodiment. The arm operation command input unit 34 inputs arm operation commands (such as the position and posture of the end effector 21a of the arm 21, and the angles of each joint of the arm 21) through an input device (such as a joystick, slave robot, or keyboard). The arm control unit 12 generates arm control commands C1 (such as joint angles, joint angular velocities, and joint torques) that satisfy the arm operation commands, and calculates the arm control commands C1 in a control period T a At the same time, the arm state transition θ a * and outputs it to the movement control unit 13. At this time, the arm control unit 12 estimates the arm state transition θ a * The method of generating the arm state transition θ can be arbitrary. For example, the arm state transition θ can be generated by linear extrapolation or machine learning using information on the past arm control command C1. a * A method for predicting the above-mentioned
[0114] In the movement command input unit 35, the movement command (position, posture, translational speed, etc. of the base 22b) and the allowable range expressed as the inequality constraint of Equation 13 are inputted by the allowable range input unit 31. In the movement control unit 13, the arm state transition θ a * The movement control command C2 taking into account the above is calculated using the optimal control problem of Equations 1 to 7, and the control period T b Then, the signal is output to the transfer unit 22.
[0115] In this embodiment, since the arm movement command is input by the user, the allowable range for achieving the target movement depends on the user's operation. For example, a target movement is assumed in which the user operates the arm 21 while the moving unit 22 of the robot 2 moves at a constant speed, the camera 23 provided on the arm 21 captures an image of an instrument, and the user reads the information of the instrument from the image of the camera 23 displayed on the display unit 33. In this case, if the base 22b of the robot 2 vibrates significantly in the vertical direction while moving, the information of the instrument cannot be accurately read from the image of the camera 23. Therefore, the user can correct the allowable range of vertical vibration based on the quality of the image of the camera 23 displayed on the display unit 33 and input it to the movement command input unit 35, thereby supporting the achievement of the target movement that improves the image quality of the camera 23.
[0116] In this embodiment, instrument reading is used as an example of a target motion, but the tolerance range can be set in a similar manner in cases where the user operates the arm 21 based on a target motion such as grasping, placing, transferring, throwing, or pressing a switch. [Explanation of symbols]
[0117] 100 Robot Control System 1 Control device 11 Motion Planning Section 12 Arm control unit 13 Movement control section 2. Robot 21 Arm 21a End effector 22 Mobile section 22a Legs 22b base 22c Rotor 23 Camera 24 Sensors 31 Tolerance input section 32 Arithmetic section 33 Display section 34 Arm movement command input unit 35 Movement command input section
Claims
1. A robot having an arm and a moving part; A control device for controlling the robot; A robot control system having: The control device includes: a motion planning unit that outputs a target motion related to a motion of the arm unit and an allowable range of a motion of the moving unit corresponding to the target motion; an arm control unit that outputs an arm control command and an arm state transition related to the target motion; a movement control unit that uses the arm state transition to generate a movement control command regarding an operation of the movement unit so as to fall within the allowable range; Equipped with The robot control system is characterized in that the arm of the robot is controlled by the arm control command, and the moving part is controlled by the movement control command.
2. The motion planning unit generates a target trajectory of the motion of the moving unit within the allowable range, 2. The robot control system according to claim 1, wherein the movement control unit generates the movement control command so as to reduce an error from the target trajectory.
3. the moving part of the robot has legs or rotors, The robot control system according to claim 1 , wherein the movement control unit generates the movement control command related to a state of the joints of the leg or the rotor.
4. 2. The robot control system according to claim 1, wherein the arm control unit generates the arm state transitions by predicting future arm control commands using the state transitions of the arm control commands.
5. The robot control system according to claim 1 , wherein the permissible range is generated by setting at least one of a lower limit and an upper limit for a state of the moving section or an end effector of the arm section.
6. 6. The robot control system according to claim 5, wherein the state of the moving part or the end effector is any one of a position, a rotation, a translational velocity, a rotational velocity, a translational acceleration, a rotational acceleration, a translational jerk, and a rotational jerk.
7. The robot control system according to claim 5, wherein the movement control unit generates the movement control command based on a weighting matrix that adjusts the magnitude of fluctuation in the state of the moving unit or the end effector within the allowable range.
8. an allowable range input unit for inputting information regarding the allowable range; a calculation unit that calculates display information based on a result of controlling the robot using the allowable range; A display unit that displays the display information; The robot control system according to claim 1 , comprising:
9. The display information includes:
9. The robot control system according to claim 8, further comprising any one of the tolerance range, the state of the moving part or arm, the power consumed by the robot, and a prediction of the completion time of the target motion of the arm.
10. 9. The robot control system according to claim 8, wherein the information regarding the allowable range includes any one of a state of a moving unit or an end effector of the arm unit for which the allowable range is set, an upper limit of the allowable range, and a lower limit of the allowable range.
11. an arm movement command input unit for inputting an arm movement command related to the movement of the arm; a movement unit operation command input unit for inputting an allowable range of the movement of the movement unit; A sensor for acquiring information; A display unit that displays information acquired by the sensor, 2. The robot control system according to claim 1, wherein the arm control unit generates an arm control command and an arm state transition related to the arm operation command.
12. A robot control method for controlling a robot having an arm and a moving part, comprising: a motion planning step of outputting a target motion related to the motion of the arm unit and an allowable range of the motion of the moving unit corresponding to the target motion; an arm control step of outputting an arm control command and an arm state transition related to the target motion; a movement control step of generating a movement control command for an operation of the moving unit using the arm state transition so as to fall within the allowable range; a robot control step of controlling the arm unit according to the arm control command and controlling the moving unit according to the movement control command; A robot control method comprising:
13. A control device for controlling a robot having an arm and a moving unit, a motion planning unit that outputs a target motion related to a motion of the arm unit and an allowable range of a motion of the moving unit corresponding to the target motion; an arm control unit that outputs an arm control command and an arm state transition related to the target motion; a movement control unit that uses the arm state transition to generate a movement control command regarding an operation of the movement unit so as to fall within the allowable range; A robot control device comprising:
Citation Information
Patent Citations
Walking robot
JP2007007797A
Ultraviolet light sanitizing cart
JP2022028618A
Arm and Body Coordination
US20220193900A1
