JP2012157965A - Device and method for determining moving method of robot, and program for the same - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a device for determining a moving method of a robot, which can determine an optimal trajectory even when a dynamics model of an objective robot is unknown.SOLUTION: The device 100 for determining a moving method of a robot, includes: an actual machine control part 102 for determining a control value candidate that is a set of control value candidates for controlling a plurality of driving parts provided with the robot so that the robot is moved to a second position from a first position; an actual measurement value obtaining part 104 for obtaining an actual measurement value that is a physical quantity indicating driving states of the plurality of driving parts controlled by the control value candidate; an evaluation part 106 for determining an evaluation value of the control value candidate corresponding to the actual measurement value, based on the actual measurement value; and a control value determination part 108 for determining whether to adapt the control value candidate as a set of control values for determining a moving method of the robot, based on the evaluation value.

Description

  The present invention relates to a robot motion method determination device and the like, and more particularly, to a robot motion method determination device that determines a motion method for moving a robot from a predetermined first position to a second position.

  In recent years, there has been a demand for a reduction in the amount of fossil fuel used in order to reduce carbon dioxide emissions. Therefore, not only when using automobiles and household electrical appliances, but also at the time of manufacturing them, it is required to reduce the amount of potato fossil fuel and electric power used. Therefore, there is an increasing need to reduce the power consumption of industrial robots widely used in the manufacturing process of industrial products.

  In general, the static and dynamic characteristics of a robot having a plurality of drive joints exhibit strong non-linearity, and the drive force of each joint varies in a complicated manner depending on the position, posture, speed, and acceleration of the robot. Here, when moving the output part (for example, the hand part of the robot arm) of the robot from the first position to the second position, the displacement of the output part (for example, the hand part) of the robot with respect to time is “trajectory”. Or, it is called “exercise method”. That is, hereinafter, the robot trajectory (or the robot movement method) represents a time change of the position of the robot movable part when the robot movable part is moved from the first position to the second position.

  The trajectory of an industrial robot that performs routine work can be determined in advance. Therefore, the power consumption of the robot can be reduced by optimizing the trajectory in advance so that the driving force of each joint or the energy required for driving is reduced.

  Conventionally, as a method for optimizing the trajectory of such a robot, simulation based on a dynamic model of a target robot has been attempted (for example, see Non-Patent Document 1).

  By searching for the optimal trajectory by simulation, it is possible to repeatedly search a large number of trajectories by taking a calculation time, and it is possible to determine an optimal trajectory that further reduces the power consumption of the robot.

Michihiro Hayashi, Hiroshi Tachiya, Naoki Asakawa, “Determining the trajectory of a multi-degree-of-freedom manipulator using heuristic methods”, Transactions of the Japan Society of Mechanical Engineers (C), February 2009, Vol. 75, No. 750, p. 14-21 S., Lin, B., Kernighan, “An Effective Heuristic Algorithm for the Traveling-Salesman Problem”, Operations Research, 1973, Vol.21, No.2, p.498-516

  However, in the conventional method using simulation, it is necessary to know the dynamic model of the target robot, but it is generally very difficult to know this.

  FIG. 14 is a diagram illustrating an outline of a trajectory generation method using simulation based on a dynamic model of a robot according to related technology.

  As shown in FIG. 14, in this method, the computer 900 first identifies parameter set candidates for determining the trajectory 906 of the robot by the heuristic method 904. The computer 900 obtains an analysis value of the driving force required when the robot is operated on the trajectory 906 to which this parameter set is applied. The computer 900 can determine the optimal trajectory 906 by repeatedly searching for candidates for the trajectory 906 by the heuristic method 904 so that the analytical value becomes smaller, for example, using the analytical value as an evaluation value.

  In this method, in order to obtain an evaluation value, it is indispensable that the computer 900 has a dynamic model 902 of the target robot as a mathematical formula therein.

  FIG. 15 is a diagram illustrating a specific example of a robot dynamic model 902 according to related technology.

  As shown in FIG. 15, the robot dynamic model 902 includes, for example, parameters that determine the motion state of each of the joints A to C, the link lengths of the links B and C, the mass and the center of gravity, and It includes various parameters of the target robot, such as the efficiency of bearings, reducers and actuators. In order to perform simulation, it is necessary to determine these parameters in advance with high accuracy.

  However, many of industrial robots on the market are unclear about the values of these parameters or the parameters themselves of the dynamic model, and the method cannot be applied. Therefore, there is a problem that it is extremely difficult for a robot user to optimize the trajectory of the robot by simulation using a dynamic model.

  Therefore, an object of the present invention is to provide a robot motion method determination device that can determine an optimum trajectory even when the dynamic model of the target robot is unknown.

  A robot motion method determination device according to an aspect of the present invention is a robot motion method determination device that determines a motion method for moving a robot from a predetermined first position to a second position. The control value candidate that is a set of control value candidates for controlling a plurality of drive units included in the robot is determined so that the robot moves from the first position to the second position, and the determined control value candidate Corresponding to the actual measurement value from the actual value control unit, the actual value control unit that acquires the actual value that is the physical quantity indicating the driving state of the plurality of drive units controlled by the control value candidates, and the actual value An evaluation unit that determines an evaluation value of a control value candidate, and a control value determination unit that determines whether or not to adopt a control value candidate as a set of control values for determining a robot movement method based on the evaluation value With.

  According to this configuration, the robot motion method determination device can obtain the evaluation value of the control value candidate by acquiring the actual measurement value corresponding to the control value candidate from the actual machine without the dynamic model of the target robot. it can. Therefore, the robot motion method determination device determines whether or not the control value candidate should be adopted as the final set of control values when determining the robot motion method based on the evaluation value obtained from the actual measurement value. And can be determined. As a result, it is possible to provide a robot motion method determination device capable of determining an optimal trajectory even when the dynamic model of the target robot is unknown.

  Furthermore, an evaluation storage unit is provided, the actual machine control unit determines a plurality of different control value candidates, and the evaluation unit associates each of the plurality of control value candidates with the evaluation value of the control value candidate. The control value determining unit stores the control value candidate corresponding to the evaluation value with the best evaluation among the plurality of evaluation values stored in the evaluation storage unit as a set of control values. You may decide.

  According to this configuration, the robot motion method determination device causes the actual machine to execute each of the plurality of control value candidates, and stores the evaluation value obtained from the actually measured value in the storage unit in association with each of the control value candidates. Can be made. Therefore, the robot motion method determination apparatus 100 can search for control value candidates with the goal of optimizing the evaluation value by repeating the evaluation with the actual machine until a predetermined end condition is satisfied. As a result, a more optimal trajectory can be determined.

  The robot motion method determination device according to another aspect of the present invention further includes a similarity calculation unit that calculates a similarity between two control value candidates, and the actual machine control unit is included in the plurality of control value candidates. A plurality of control value candidates may be determined such that the similarity between any two control value candidates is not similar to the similarity indicated by a predetermined similarity threshold.

  In general, for control value candidates that are similar to control value candidates for which evaluation values have already been acquired, it is expected that evaluation values that are not significantly different from evaluation values that have already been acquired are often obtained even when evaluation is performed using actual devices. The Therefore, the actual machine control unit can reduce the number of useless evaluations by the actual machine by limiting the control value candidates to be evaluated by the actual machine to control value candidates that are not similar to the already evaluated control value candidate to a certain degree or more. . As a result, the time required for the robot motion method determination device to determine the final set of control values can be reduced.

  Specifically, the control value is a value corresponding to at least one of the driving time, driving distance, driving speed, driving acceleration, and position of the driving unit, and the actual measurement value is a current value, power, and The evaluation unit is at least one of the electric energy, and the evaluation unit determines the evaluation value so that the smaller the sum of the actual measurement values or the peak value of the actual measurement values in the plurality of driving units is smaller, the similarity calculation unit Calculates the similarity as the reciprocal of the sum of absolute values of the differences between the corresponding control value candidates between the two control value candidates, and the actual machine control unit controls any two controls included in the plurality of control value candidates. A plurality of control value candidates are determined such that the degree of similarity between the value candidates is less than a predetermined threshold, and the control value determination unit is the most of the plurality of evaluation values stored in the evaluation storage unit. Control value candidates corresponding to large evaluation values are set as a set of control values. It may be determined to use.

  In addition, the actual machine control unit, when the evaluation value corresponding to one control value candidate among any two control value candidates included in the plurality of control value candidates is less than a predetermined evaluation reference value May determine a plurality of control value candidates such that one control value candidate and the other control value candidate are not similar to the similarity indicated by a predetermined threshold.

  In general, as a result of the evaluation that has already been made, a control value candidate that is similar to a control value candidate that is poorly evaluated is expected to have a poor evaluation. On the other hand, if the control value candidate is similar to the control value candidate with good evaluation, it can be expected that the optimal trajectory is included in the trajectory similar to the trajectory. Therefore, the actual machine control unit determines that it is wasteful to evaluate the control value candidate similar to the control value candidate that has already been evaluated as a result of evaluation, and evaluates with the actual machine again. On the other hand, a control value candidate similar to a control value candidate that has been found to have a high evaluation is transmitted to the actual machine and evaluated. Thus, the robot motion method determination device can determine a more optimal set of control values while suppressing the number of evaluations by the actual machine.

  The present invention can be realized not only as such a robot motion method determination device, but also as a robot motion method determination method using characteristic means included in the robot motion method determination device as a step. Such characteristic steps can also be realized as a program for causing a computer to execute. Needless to say, such a program can be distributed via a recording medium such as a CD-ROM (Compact Disc Read Only Memory) and a transmission medium such as the Internet.

  Furthermore, the present invention can be realized as a semiconductor integrated circuit (LSI) that realizes part or all of the functions of such a robot motion method determination device, or a robot motion including such a robot motion method determination device. Or as a method decision system.

  As described above, the present invention can provide a robot motion method determination device capable of determining an optimal trajectory even when the dynamic model of the target robot is unknown.

It is a conceptual diagram which shows the use condition of the movement method determination apparatus of the robot which concerns on Embodiment 1 and 2 of this invention. It is a figure which shows the functional block of the movement method determination apparatus of the robot which concerns on Embodiment 1 of this invention. It is a 1st figure which shows an example of the design variable matrix used for SHA in Embodiment 1 and 2 of this invention. It is a 2nd figure which shows an example of the design variable matrix used for SHA in Embodiment 1 and 2 of this invention. It is a figure which shows an example of the narrowing-down search used for SHA in Embodiment 1 and 2 of this invention. It is a figure which shows the procedure of the process in which the movement method determination apparatus of the robot which concerns on Embodiment 1 determines a track | orbit. It is a figure which shows an example of the change of the moving speed of the robot hand tip corresponding to a robot command. It is a figure which shows the functional block of the movement method determination apparatus of the robot which concerns on Embodiment 2. FIG. It is a figure which shows the procedure of the process in which the movement method determination apparatus of the robot which concerns on Embodiment 2 determines a track | orbit. It is a figure which shows an example of the range determined by the real machine control part to be similar based on similarity. It is an external view of a 6-DOF vertical articulated robot used in verification as an example of an actual machine. It is a figure which shows an example of the result of having compared the electric current value of the motor with the track | orbit optimized by the movement method determination apparatus of the robot which concerns on Embodiment 2, and a standard track. It is a block diagram which shows the hardware constitutions of the computer system which implement | achieves the movement method determination apparatus of the robot concerning Embodiment 1 and 2 of this invention. It is a figure which shows the outline of the trajectory generation method using the simulation based on the dynamic model of the robot which concerns on related technology. It is a figure which shows the specific example of the dynamic model of the robot which concerns on related technology.

  DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of a robot motion method determining apparatus according to the present invention will be described below in detail with reference to the drawings.

(Embodiment 1)
First, the configuration of the robot motion method determination device according to the first embodiment of the present invention will be described.

  FIG. 1 is a conceptual diagram showing a usage state of a robot motion method determination device 100 according to the present embodiment.

  As shown in FIG. 1, the robot movement method determination device 100 according to the present embodiment does not have a target robot dynamic model. Instead, an input / output interface with the actual robot 200 is provided.

  Specifically, the robot motion method determination apparatus 100 determines a set of control value candidates (hereinafter referred to as control value candidates) of the real machine 200 using the heuristic method 550. The control value of the real machine 200 is, for example, a parameter value for controlling a plurality of drive units included in the real machine 200.

  The robot motion method determination apparatus 100 transmits the determined control value candidates to the actual machine 200.

  Next, the actual machine 200 actually operates according to the received control value candidate (that is, on the trajectory 552). For example, when the robot arm included in the actual machine 200 is moved from the first position to the second position, the robot arm can take various trajectories, and the trajectory 552 is one of them.

  The actual machine 200 includes a measuring device 202 that measures a physical quantity indicating the driving state of the drive unit, such as current values or electric power amounts flowing through a plurality of motors included in the actual machine 200. The robot motion method determination device 100 acquires the actual measurement value of the physical quantity measured by the measurement device 202.

  The robot motion method determination device 100 evaluates the acquired actual measurement value according to a predetermined standard (for example, an evaluation function). Thereafter, the robot motion method determination device 100 repeatedly generates a new trajectory 552 by the heuristic method 550 until a predetermined condition is satisfied, gives this trajectory 522 to the actual machine 200, acquires an actual measurement value, and evaluates it. I do. The robot motion method determination apparatus 100 determines the trajectory 522 having the best evaluation among the plurality of trajectories 552 generated as described above as the optimal trajectory 552.

  As described above, the robot motion method determination device 100 evaluates the control value candidates using the real machine 200, so that even if the dynamic model of the real machine 200 that is the target robot is unknown, the robot movement method is determined. The method determining apparatus 100 can determine an optimal trajectory.

  FIG. 2 is a diagram showing functional blocks of the robot motion method determining apparatus 100 according to the embodiment of the present invention.

  As illustrated in FIG. 2, the robot motion method determination device 100 includes an actual machine control unit 102, an actual measurement value acquisition unit 104, an evaluation unit 106, a control value determination unit 108, and an evaluation storage unit 112.

  As described above, the robot movement method determining apparatus 100 determines a movement method for moving the robot from a first position to a second position.

  At that time, the real machine control unit 102 controls the plurality of drive units included in the real machine 200 so that the robot shown as the real machine 200 in FIG. 2 moves from the first position to the second position. A control value candidate that is a set of candidates is determined, and the determined control value candidate is transmitted to the real machine 200.

  Here, the control value is, for example, a value corresponding to at least one of the driving time, driving distance, driving speed, driving acceleration, and position of each driving unit included in the actual machine 200.

  Specifically, the actual machine control unit 102 determines control value candidates by a heuristic technique called SHA. Details thereof will be described later.

  Note that since it is difficult for the actual machine control unit 102 to determine an optimal control value candidate at a time, it is preferable to determine a plurality of different control value candidates from each other and perform repeated evaluation using the actual machine 200.

  The actual measurement value acquisition unit 104 is an interface that acquires actual measurement values that are physical quantities indicating the drive states of a plurality of drive units included in the real machine 200 controlled by the control value candidates. As the actual measurement value, for example, at least one of current, electric power, and electric energy flowing through the driving unit can be considered. Further, any other physical quantity that can be measured for the actual machine 200 (for example, the position of the tip of the robot arm included in the actual machine 200, the acceleration, the surface temperature of the drive unit, the illuminance, the amplitude of the ambient sound, etc.) is used as the actual measurement value. May be.

  The evaluation unit 106 determines an evaluation value of a control value candidate corresponding to the actual measurement value from the actual measurement value acquired by the actual measurement value acquisition unit 104. For example, if the actual measurement value is the amount of power of each of the plurality of drive units included in the actual machine 200, the evaluation unit 106 may determine the total value as the evaluation value. In this case, the smaller the evaluation value, the lower the power consumption and the better the evaluation.

  In addition, for the purpose of energy saving and cost saving by controlling the peak value of power consumption to be a certain value or less as in so-called demand control, for example, the power of each of a plurality of drive units is used as an actual measured value, The maximum value of power consumption in the entire drive unit over 200 operating hours may be used as the evaluation value. In this case, the smaller the evaluation value, the smaller the peak power is, indicating that the evaluation is good.

  That is, in the case where the power consumption of the actual machine 200 is intended to be suppressed, if the measured value is at least one of the current, power, and power amount flowing through the drive unit, the evaluation unit 106 The evaluation value may be determined to be larger as the total of the actual measurement values in the driving unit is smaller.

  Further, the evaluation unit 106 stores each of the plurality of control value candidates determined by the actual machine control unit 102 in an evaluation storage unit, which will be described later, in association with the evaluation value of the control value candidate.

  Based on the evaluation value, the control value determination unit 108 determines whether or not to adopt the control value candidate as a set of control values for determining the exercise method of the actual machine 200.

  Specifically, the control value determination unit 108 determines to adopt the control value candidate corresponding to the evaluation value with the highest evaluation among the plurality of evaluation values stored in the evaluation storage unit as a set of control values. To do. More specifically, the control value determination unit 108 determines to adopt a control value candidate corresponding to the largest evaluation value among a plurality of evaluation values stored in the evaluation storage unit 112 as a set of control values. To do.

  As described above, the evaluation storage unit 112 is a storage unit that stores control value candidates and their evaluation values in association with each other. Specifically, the evaluation storage unit 112 includes RAM (Random Access Memory), ROM (Read Only Memory), SRAM ( Realized by Static Random Access Memory).

  Next, referring to FIG. 3 and FIG. 4, a search algorithm based on a heuristic method used by actual machine control unit 102 included in robot motion method determination apparatus 100 according to the present embodiment to determine evaluation value candidates. SHA which is is demonstrated.

  A GA (Genetic Algorithm) is a typical search algorithm based on the heuristic method. Lin et al., Proposed in Non-Patent Document 2, use the SHA modified by the inventors so as to be compatible with the robot motion method determining apparatus 100.

  In SHA, a design variable matrix called a design variable matrix, each column indicates a design parameter to be determined, and each row constructs a matrix indicating a discrete value (that is, a control value) that can be taken by each design parameter, and the matrix is operated. Thus, the combination of variables that is evaluated best is searched.

FIG. 3 is a diagram showing an example of a design variable matrix used for SHA in the present embodiment. The element S _{ij in} the i row and j column in the figure represents the i th control value of the j th design parameter. In the present embodiment, the design parameter is, for example, any one of a driving time, a driving distance, a driving speed, a driving acceleration, and a position of a driving unit included in the real machine 200 at a certain time. Further, the control value is a value corresponding to one of predetermined drive times (1 second, 3 seconds, 5 seconds, 7 seconds, 9 seconds, etc.) if the design parameter is the drive time, for example. If the design parameter is the driving distance, it is a value corresponding to one of the predetermined driving distances (1 m, 3 m, 5 m, 7 m, 9 m, etc.).

  Hereinafter, a specific procedure for operating the design variable matrix (hereinafter also referred to as SHA processing step) will be described with reference to FIGS.

  Step 1: When the number of design parameters to be determined is n, the actual machine control unit 102 corresponds to each of the n design parameters and each row corresponds to a discrete value that can be taken as a control value of the design parameter. Create a design variable matrix. It is assumed that the number of discrete values that can be taken as control values for each design parameter is the same. FIG. 3 shows a design variable matrix in which the design variable is 6 and the number of possible discrete values is 9, respectively.

Step 2: The actual machine control unit 102 randomly selects one discrete value in each column of the design variable matrix and determines an initial path. The determined initial path and reference path T _0. In FIG. 3, elements S _{61 to} S ₅₆ connected by a solid line are selected as T ₀ . In the following, control value candidates in SHA are also referred to as “routes”, assuming a route from p ₁ to p ₆ .

Step 3: The actual machine control unit 102 performs element selection operations in order from the first column of the design variable matrix. First, in the first column, randomly selects an element of the fixed number in addition to the elements that are selected as the reference path T _0. In FIG. 4A, white circle elements (S ₄₁ , S ₆₁ , S ₈₁ ) are selected. These elements are called “Look-ahead-Base Point (LBP)” and are candidates for new design variables.

Step 4: The actual machine control unit 102 randomly selects an element in the next row following each element set as the LBP, and generates a plurality of paths as shown in FIG. _{Specifically,} the path comprising path including the _{S 41} and _{S 22,} a path including the _{S 61} and _{S 32,} the _{S 61} and _{S 52,} as _well, the path including the _{S 81} and _{S 82} are generated.

Step 5: The actual machine control unit 102 evaluates each of the reference path determined in Step 2 and the plurality of newly generated paths in Step 4 with the actual machine 200, and the obtained evaluation value is the control value with the best control value. The combination is set as a new reference route. For example, the path including the _{S 41} and _{S 22} shown in FIG. 4 (b) and is selected as a new reference path _{T 1.}

Thereafter, the real machine control unit 102 sequentially advances the target column to the next column and performs an operation (steps 3 to 5 above) for updating a route with a good evaluation value as a reference route (steps 3 to 5 above) (shown in FIG. 4). In the example, repeat until the fifth column). For example, FIG. 4 (c), carried out subject to the element selection operations a second column (Step 3), shows a case where the reference path T ₂ is selected.

  Step 6: The actual machine control unit 102 does not select an element in the final column (sixth column in the example shown in FIG. 4), and takes the same number of paths as the discrete values for all elements included in the final column. A set of discrete values having the best evaluation value is generated and updated as a reference path.

Step 7: As shown in FIG. 4D, the real machine control unit 102 moves the last column (sixth column) of the design variable matrix to the first column, and moves the other columns to the next column. Reconstruct the design variable matrix. The above operation is repeated for the obtained matrix. The actual machine control unit 102 repeats the operations in steps 3 to 7, and when the design variable (p _{1 in the} example shown in FIG. 4) that is the first column first becomes the first column again. The SHA process is terminated.

  The control value determination unit 108 included in the robot movement method determination device 100 performs the final operation for determining the movement method of the actual machine 200 by selecting the control value candidate corresponding to the route having the best evaluation value through the repetitive operation by the SHA process. It is adopted as a set of typical control values.

  In the heuristic method, since a solution is searched using random numbers, even if a result in the vicinity of the optimal solution is obtained at the time of iterative calculation, a better solution in the vicinity of the result is not always searched. In addition, it is expected that a better solution of the evaluation function value can be obtained as the width of the discrete values that each design parameter can take is subdivided, but if the number of discrete values is increased, the number of rows becomes enormous and the search is difficult. Become.

  Therefore, the actual machine control unit 102 according to the present embodiment first has a control value candidate having a good evaluation value by performing the operations of Step 1 to Step 7 for the design variable matrix in which the width of the discrete values is set relatively coarse. Then, the control variable included in the obtained control value candidate is set to the median value, and the design variable matrix is further subdivided into step sizes. The search process of step 1 to step 7 may be executed again. The real machine control unit 102 can optimize the control value more accurately by using the reference path finally obtained by this search as a solution. The initial search is referred to as “preliminary search”, and the search using the design variable matrix obtained by subdividing the step size in the vicinity of the search result is referred to as “narrowed search”.

  The narrowing search will be described more specifically with reference to FIG.

  FIG. 5 is a diagram illustrating an example of a refinement search used for SHA in the present embodiment.

  In FIG. 5A, the driving speeds of the robot hand tip at three times (t1, t2, t3) are set as design parameters p1, p2, and p3, respectively, and the respective control values (that is, the robot hand tip speed) are set. The result of the preliminary search when determining is shown.

  At the time of starting the preliminary search, the search range 702 for p1, the search range 704 for p2, and the search range 706 for p3 all have a common width (upper and lower limit values).

  Here, the preliminary search is completed, and as shown in FIG. 5A, the actual control unit 102 determines v1 as the control value of p1, v2 as the control value of p2, and v3 as the control value of p3. Suppose that

  In this case, it is considered that the optimal solutions of p1 to p3 are likely to exist in the vicinity of v1 to v3, respectively. Therefore, the real machine control unit 102 sets a new search range 708 in which the search range 702 of p1 is narrowed to the vicinity of v1 as compared with the time of the preliminary search, for example, as shown in FIG. . Similarly, the actual machine control unit 102 sets new search ranges narrowed down to the vicinity of v2 and v3 for the search range 704 and the search range 706, respectively. Here, if the number of discretizations in the search range after narrowing down and the search range before narrowing down are the same, the granularity of discretization becomes finer in the search range after narrowing down. Therefore, the actual machine control unit 102 can determine a more optimal control candidate value in the vicinity of the control candidate value determined by the preliminary search by executing SHA again for the search range after narrowing down.

  The actual machine control unit 102 does not necessarily need to perform a narrowing search. For example, if the evaluation value of the control value determined in the preliminary search is lower than a predetermined reference value, a new search range is set by expanding the search range of each design parameter. SHA may be performed again for the target. A search range 710 shown in FIG. 5C is an example of this.

  Further, for example, when the control value determined by the preliminary search is biased either above or below the median value of the search range, the actual machine control unit 102 looks like the search range 712 shown in FIG. In addition, after the search range is shifted so as to eliminate the bias (that is, the control value determined by the preliminary search is brought closer to the median of the search range), the SHA is performed again in the new search range. May be.

  The SHA performed by the actual machine control unit 102 in the present embodiment described above has the following advantages compared with the GA.

  (1) In GA, it is necessary to convert a design variable into a gene model expressed by a binary sequence, but in SHA, a real value can be directly handled as a design variable.

  (2) In GA, it is necessary to adjust a plurality of parameters that influence the execution of solution search, such as the number of generations, crossover rate, and mutation rate. Although these values have a significant effect on the solution search results, the decision must be largely based on experience. On the other hand, there are no parameters that are difficult to adjust with SHA.

  (3) At the time of iterative search, the GA needs to acquire evaluation values for a large number of solution candidates every time. When the processing of the evaluation function is complicated, the entire processing amount becomes enormous. On the other hand, since the number of solution candidates evaluated by SHA is small, the processing time for each time is short, and the entire processing time can be shortened.

  In particular, the advantage of shortening the processing time described in (3) is an important advantage in the present invention. When using a robot simulated in a computer as a dynamic model, the processing time does not matter so much, but when using the actual robot, a person always stands by beside the robot in consideration of safety issues. Need to do. In addition, in the case of a robot that has already been introduced in the field, the time that can be occupied for the trajectory determination work is usually limited to a short time. Therefore, compared with GA and SHA related to the related art, the inventors' original SHA using the refinement search is more suitable for realizing the robot motion method determination device 100 according to the present embodiment. Yes.

  Next, with reference to FIG. 6, the procedure of the process in which the robot movement method determination device 100 according to the present embodiment determines the trajectory will be described.

  First, the actual machine control unit 102 acquires a condition for trajectory search determined by the user, for example, by a GUI (Graphical User Interface; not shown) or by reading an external file. More specifically, (1) design parameters necessary for determining the trajectory of the robot, and (2) upper and lower limit values that can be taken by the control values that are the values of the design parameters and the number of divisions thereof are acquired ( S102, S104).

  Next, when the actual machine control unit 102 determines that the SHA process end condition as shown in the SHA process step 7 is not satisfied (No in S106), the actual machine control unit 102 executes the SHA process.

  Here, when there is a control value candidate whose evaluation value is not yet determined in the above-described SHA processing step 5, the actual machine control unit 102 selects one of the control value candidates whose evaluation value is not determined ( S108), and transmits to the real machine 200 (S112). More specifically, the control value candidate is transmitted to the actual machine 200 as a parameter in the robot operation program of the actual machine 200.

  Next, the actual device 200 operates according to the transmitted control value candidate (S114). In the meantime, the physical quantity indicating the drive state of the drive unit of the real machine 200 is measured (S116). For example, a power meter is connected to each of a plurality of motors included in the actual machine 200, and the power consumption of each motor is measured. The wattmeter may be included in the actual device 200 or may be connected to the outside of the actual device 200.

  The actually measured values of the power consumption of the motors measured in this way are transmitted to the actual value acquisition unit 104 provided in the robot motion method determination device 100 by wireless or wired communication. The actual measurement value is evaluated by the evaluation unit 106 using a predetermined evaluation function, and an evaluation value for the control candidate value is determined (S118). Thereafter, the evaluation unit 106 causes the evaluation storage unit 112 to store the evaluation value and the control candidate value corresponding to the evaluation value.

  As described above, as the evaluation function, for example, a function that determines a higher evaluation value as the total power consumption of the real machine 200 is smaller, a function that determines a higher evaluation value as the peak power is smaller, and the like are considered. You may determine an evaluation function on the basis.

  The above steps S108 to S118 are repeated until evaluation values are determined for all control value candidates in the above-described SHA processing step 5. Furthermore, since the control value determination unit 108 updates the optimal value of the control value candidate, whether or not a control value candidate having a better evaluation than the control value candidate corresponding to the current reference route has been found in the SHA processing step 5 described above. Is determined (S120). As a result of the determination, if a control value candidate with better evaluation is found (Yes in S120), the control value determining unit 108 sets a control value candidate with better evaluation as a reference route (S122). On the other hand, when it is determined that a control value candidate with better evaluation has not been found (No in S120), the control value determination unit 108 does not update the reference route.

  Next, the control value determination unit 108 determines whether to reset the SHA search condition or abort the search process by SHA based on the transition of the best evaluation value corresponding to the reference route (S124).

  Here, the resetting of the search condition means, for example, that the control value determining unit 108 changes to that shown in FIG. 5C when the rate at which the reference route is changed in step S122 is lower than a predetermined value. As shown, it means to decide to reset the search conditions so that the upper and lower limits of each design parameter are widened. Further, when the evaluation value is significantly improved in step S122, the control value determination unit 108 resets the search condition so that the upper and lower limits of each design parameter are narrowed as shown in FIG. You may decide to set.

  In addition, for example, when the search process is terminated, for example, when the update of the reference route is delayed in step S122 (more specifically, when the reference route is not updated continuously for a predetermined number of times, When the update rate of the reference route falls below a predetermined value), the control value determination unit 108 aborts the search by SHA (Yes in S126), and the control value candidate that has shown the best evaluation value so far is Deciding to adopt as a set of control values for deciding the exercise method of the actual machine 200.

  On the other hand, if the control value determining unit 108 determines that the search for the set of control values by SHA is to be continued (No in S126), the control value determining unit 108 returns to the process of step S106 performed by the actual machine control unit 102 again.

  In addition, when the actual machine control unit 102 determines that the SHA process end condition as shown in the SHA process step 7 is satisfied (Yes in step S106), the actual machine control unit 102 ends the SHA process. . After that, the control value determination unit 108 determines to adopt the control value candidate that showed the best evaluation value at that time as a set of control values for determining the exercise method of the actual machine 200.

  Next, specific examples of control value candidates in the present embodiment will be described in detail.

  When the trajectory is determined by the dynamic model 902 according to the related technology, the temporal change of the joint angle displacement and the hand tip displacement is expressed by a polynomial, and the robot trajectory is expressed by appropriately determining the coefficient. . However, in the present embodiment, when operating the actual machine 200, it is difficult to control the displacement of each joint as expressed by a polynomial or the like. Therefore, in the robot motion method determination device 100 according to the present embodiment, the robot command parameters that are standard in industrial robots are used as design parameters, and combinations of the values are used as control value candidates, so that the actual machine Represents 200 trajectories. This will be described in more detail below.

  In general, most industrial robots are controlled by robot commands. For example, the moving speed and coordinates of the tip of the robot hand are controlled by transmitting a plurality of robot command sets as described below to the robot.

SPEED 500MM / S
LMOVE (500, 1000, 250)

  Here, the SPEED command specifies the moving speed, and the LMOVE command specifies the hand tip coordinates. The hand tip is controlled so as to pass the hand tip coordinates specified by the LMOVE command at the moving speed specified by the SPEED command.

  The robot operates by designating the speed and position one after another using such a set of commands. These “speed commands” and “position commands” are different to the extent that they can be replaced with SPEED → velocity, LMOVE → move linear, etc. even if the robots are of different manufacturers. Therefore, the trajectory of the robot is determined by designating a set of a speed command value (speed command value) and a position command value (position command value) for each drive unit. Therefore, the robot motion method determining apparatus 100 controls the actual machine 200 by transmitting specific values of the respective command values to the actual machine 200 as control value candidates, and thereby the evaluation values corresponding to the control value candidates. Can be obtained.

  For example, the robot motion method determination apparatus 100 uses a position command value x, y, z and a speed command value v1, v2, v3 in the following robot command as design parameters, and controls a set of these values. By transmitting to the real machine 200 as a value candidate, the corresponding evaluation value is acquired.

LMOVE ([x], [y], [z])
SPEED [v1] MM / S
LMOVE ([x] +100, [y] +100, [z] +100)
SPEED [v2] MM / S
LMOVE ([x] +200, [y] +200, [z] +200)
SPEED [v3] MM / S
LMOVE ([x] +300, [y] +300, [z] +300)

  FIG. 7 shows changes in the moving speed of the hand tip of the actual machine 200 corresponding to the robot command.

  FIG. 7 is a diagram illustrating an example of a change in the moving speed of the hand tip corresponding to the robot command. The hand tip of the actual machine 200 moves from the initial value (x, y, z) to a point 651 specified by (x + 100, y + 100, z + 100), passes through the point 651 at a speed v1, and then (x + 200, y + 200). , Z + 200), move to the point 652 designated. The hand tip of the real machine 200 passes through the point 652 at the speed v2, then goes to the point 653 designated by (x + 300, y + 300, z + 300), passes through the point 653 at the speed v3, and then stops.

  Actually, since the robot controller smoothly controls between designated coordinates, the speed change is smooth. However, since the speed change is the same for the same robot command, this method can be used as a method for expressing the trajectory.

  As the speed of the robot changes, the joint drive torque and power consumption of the robot at that time also change. Therefore, if a control value candidate (here, a set of speed command value and movement command value) can be appropriately determined, an optimal trajectory that minimizes the power consumption of the actual machine 200 can be determined.

  That is, the robot motion method determining apparatus 100 according to the present embodiment determines the motion method for moving the actual machine 200, which is a robot, from a predetermined first position to a second position. A method determination device that determines a control value candidate that is a set of control value candidates for controlling a plurality of drive units included in a robot such that the robot moves from a first position to a second position. An actual machine control unit 102 that transmits the determined control value candidates to the robot, an actual value acquisition unit 104 that acquires actual values that are physical quantities indicating the drive states of a plurality of drive units controlled by the control value candidates, The evaluation unit 106 that determines the evaluation value of the control value candidate corresponding to the actual measurement value from the value, and the control value candidate as a set of control values for determining the robot movement method based on the evaluation value And a control value determining unit 108 that determines whether Luke.

  The robot motion method determining apparatus 100 according to the present embodiment further includes an evaluation storage unit 112, the actual machine control unit 102 determines a plurality of different control value candidates, and the evaluation unit 106 includes a plurality of control value candidates. Each of the control value candidates is stored in the evaluation storage unit 112 in association with the evaluation value of the control value candidate, and the control value determination unit 108 is the most of the plurality of evaluation values stored in the evaluation storage unit 112. Control value candidates corresponding to evaluation values with good evaluation may be determined to be adopted as a set of control values.

  According to the above configuration, since the trajectory is determined by the robot motion method determination device 100, it is not necessary to construct a dynamic model of the actual machine 200. Therefore, according to the robot motion method determination device 100, the optimal trajectory can be determined even when the dynamic model of the real machine 200 that is the target robot is unknown.

  In particular, the cooperation of the manufacturer of the real machine 200 has been indispensable for the construction of the dynamic model in the past, but according to the robot movement method determination device 100 according to the present embodiment, it is easy only on the user side of the real machine 200. In addition, the trajectory can be optimized. Therefore, it can be easily applied to the actual machine 200 that is already operating in the factory.

  Furthermore, although it is extremely difficult to create a dynamic model including the effects of individual differences, environment, and secular changes that the actual machine 200 has, according to the robot motion method determination device 100 according to the present embodiment, The trajectory can be optimized in consideration of variations in characteristics of the actual machine 200.

  The robot motion method determining apparatus 100 according to the present embodiment has been described above.

  Next, the robot motion method determining apparatus 100 according to the second embodiment that can further reduce the search time for a set of control values until the trajectory is determined will be described.

(Embodiment 2)
FIG. 8 is a diagram illustrating functional blocks of the robot motion method determination device 100 according to the second embodiment. Components similar to those of the robot motion method determining apparatus 100 according to the first embodiment shown in FIG. 2 are denoted by the same reference numerals, and detailed description thereof is omitted.

  The difference from the robot motion method determination device 100 according to the first embodiment is that the robot motion method determination device 100 according to the present embodiment further calculates a similarity between two control value candidates. The point is that the unit 110 is provided.

  When the evaluation value of the trajectory is acquired using the actual machine 200 instead of the dynamic model 902, the number of repetitions of the evaluation by the actual machine 200 is a large practical limitation.

  Therefore, the actual machine control unit 102 according to the present embodiment uses the similarity between two control value candidates calculated by the similarity calculation unit 110 to reduce the number of repetitions of evaluation by the actual machine 200.

  Specifically, the real machine control unit 102 indicates that the similarity between any two control value candidates included in the plurality of control value candidates transmitted to the real machine 200 is similar to the similarity indicated by a predetermined similarity threshold. A plurality of control value candidates are determined so as not to be similar to each other. Note that the real machine control unit 102 determines similarity only between control value candidates at which the difference between the start positions is less than a predetermined threshold among the plurality of control value candidates. It is preferable. Here, the start position is the position of the actual machine 200 at the start of control, and is, for example, the initial value of the hand tip coordinates. Details will be described below.

  FIG. 9 is a flowchart showing a processing procedure of the robot motion method determination device 100 according to the present embodiment. In addition, the same code | symbol is attached | subjected to the process similar to FIG. 6, and detailed description is abbreviate | omitted.

  The main difference between the processing procedure of the robot movement method determination device 100 according to the first embodiment and the processing procedure of the robot movement method determination device 100 according to the present embodiment is as shown in FIG. This is the point to which step S110 shown in FIG.

  Similar to the real machine control unit 102 according to the first embodiment, the real machine control unit 102 according to the present embodiment returns to the real machine 200 when there is a control value candidate for which an evaluation value is not yet determined in the SHA processing step 5. In order to send, one control value candidate is selected from them (S108).

  Here, the actual machine control unit 102 according to the present embodiment calculates the similarity between the control value candidate selected in step S108 and each of the plurality of control value candidates stored in the evaluation storage unit 112. It is determined whether or not a control value candidate similar to the control value candidate calculated by the unit 110 and selected in step S108 is already stored in the evaluation storage unit 112 (S110).

  For example, if the similarity calculation unit 110 calculates the similarity so that the more similar the two control value candidates are, the larger the value is, the similarity calculated by the similarity calculation unit 110 may be any of the similarities. If it is equal to or greater than the predetermined threshold (Yes in S110), the actual machine control unit 102 discards the control value candidate selected in Step S108 without sending it to the actual machine 200. Thereafter, the actual machine control unit 102 returns to the process of step S108. If any one of the similarities calculated by the similarity calculating unit 110 is less than a predetermined threshold (No in S110), the real machine control unit 102 determines the control value selected in Step S108. The candidate is transmitted (transferred) to the real machine 200 (S112).

  Here, various methods can be considered for the similarity calculation unit 110 to calculate the similarity.

  For example, the similarity calculation unit can calculate the similarity as the reciprocal of the sum of the absolute values of the differences between the control value candidates corresponding to the two control value candidates. In this case, the similarity is greater as the control value candidates are more similar. Further, the similarity calculation unit 110 may calculate the similarity as the sum of the absolute values of the differences between the corresponding control value candidates for the two control value candidates. In this case, the similarity is smaller as the two control value candidates are more similar.

  FIG. 10 is a diagram showing an example of a range determined by the actual machine control unit 102 as being similar based on the similarity thus obtained.

  Control value candidates whose hand tip speed command values at times corresponding to the points 661, 662, and 663 are v1, v2, and v3 are set as first control value candidates. The thick line shown in FIG. 10 is obtained by connecting each control value included in the first control value candidate with a straight line. At this time, (1) the difference of the hand tip position at the start of the operation is within a predetermined threshold distance, and (2) the speed command value of the hand tip speed at the time corresponding to the points 661, 662 and 663. However, it is determined that the second control value candidates that are all included in the region between the two thin lines drawn above and below the thick line shown in FIG. 10 are similar to the first control value candidates.

  As described above, the robot motion method determination device 100 according to Embodiment 2 of the present invention further includes the similarity calculation unit 110 that calculates the similarity between two control value candidates, and the actual machine control unit 102 includes a plurality of A plurality of control value candidates may be determined so that the similarity between any two control value candidates included in the control value candidate is not similar to the similarity indicated by a predetermined similarity threshold.

  More specifically, the control value is a value corresponding to at least one of the drive time, drive distance, drive speed, drive acceleration, and position of the drive unit, and the actually measured value is the current value, power, and , And the evaluation unit 108 determines the evaluation value so that the smaller the sum of the actual measurement values or the peak value of the actual measurement values in the plurality of drive units, the greater the similarity. The calculation unit 110 calculates the similarity as the reciprocal of the sum of the absolute values of the differences between the corresponding control value candidates between the two control value candidates, and the real machine control unit 102 is an arbitrary number included in the plurality of control value candidates. A plurality of control value candidates are determined such that the degree of similarity between the two control value candidates is less than a predetermined threshold, and the control value determination unit 108 includes a plurality of control value candidates stored in the evaluation storage unit 112. Corresponds to the largest of the evaluation values That the control value candidate may be determined to adopt as a set of control values.

  In general, for control value candidates that are similar to control value candidates that have already acquired evaluation values, evaluation values that are not significantly different from evaluation values that have already been acquired are often obtained even when evaluation is performed using the actual device 200. It is done. Therefore, the real machine control unit 102 reduces the number of useless evaluations by the real machine 200 by limiting the control value candidates evaluated by the real machine 200 to control value candidates that differ from the already evaluated control value candidates by a predetermined threshold or more. Can do. As a result, the time required for the robot motion method determination apparatus 100 to determine the final set of control values can be reduced.

  Next, a verification result when the robot trajectory is actually determined using the robot motion method determination device 100 according to the present embodiment will be described.

  FIG. 11 is an external view of a six-degree-of-freedom vertical articulated robot used in verification as an example of the actual machine 200.

  In the verification, the trajectory was optimized when the hand portion 204 of the robot shown in FIG. 11 was moved from the first position to the second position. The evaluation function at that time was determined such that the smaller the peak value of the current flowing through the motor that drives the joint of the actual machine 200, the higher the evaluation.

  FIG. 12 compares the current values of the motor that drives the robot when the robot is driven on the trajectory 752 optimized by the robot motion method determining apparatus 100 according to the present embodiment and the standard trajectory 750. It is a figure which shows an example of the result obtained.

  Here, the standard trajectory 750 is the trajectory considered to be the most frequent as the trajectory of the robot before the trajectory is optimized. Specifically, the standard trajectory 750 is determined as follows. The starting position of the standard track 750 is the center of the movable area of the actual machine 200. In addition, the actual machine 200 is pre-installed with a speed change pattern as a standard, and the standard-mounted change pattern is used as a set of control values used for movement from the first position to the second position.

  As shown in FIG. 12, the trajectory 752 determined using the robot motion method determination device 100 according to the present embodiment has a current peak value reduced by about 20% compared to the trajectory 750.

  In addition, the robot motion method determination device 100 according to the present embodiment can determine a trajectory that can obtain the effect of reducing the current peak value by repeating several tens of times. This number of times can be said to be a number of practically no problem when the evaluation is performed using the actual machine 200.

  In the first and second embodiments described above, the evaluation unit 106 included in the robot motion method determination device 100 acquires an actual measurement value from one actual machine 200 and performs the evaluation. However, when there are a plurality of actual machines 200, actual measurement values may be acquired from the plurality of actual machines 200 and evaluated. In that case, it is conceivable that the evaluation unit 106 uses an average value, median value, maximum value, minimum value, etc. of a plurality of actually measured values for determining an evaluation value of a corresponding control value candidate.

  In addition, a control value determined using one real machine 200 can be applied to a plurality of other real machines 200 used for the same purpose. As a result, power and the like are suppressed without performing a search for other robots of the same model, so that it is possible to determine a number of robot movement methods very efficiently.

  Note that the robot motion method determination device 100 according to Embodiment 1 does not necessarily include the evaluation storage unit 112. For example, when an evaluation value corresponding to a control value candidate transmitted to the actual device 200 exceeds a predetermined threshold value for the first time, the control value determination unit 108 adopts the control value candidate as a set of control values. You may decide. In this case, the robot motion method determination device 100 can achieve the same effects of the invention without the evaluation storage unit 112.

  Note that the similarity calculation unit 110 according to Embodiment 2 may calculate the similarity using a square error instead of the absolute value of the difference. Further, the similarity may be calculated depending on whether or not each control value is included in a predetermined range. For example, if the control value of each design parameter is included in a predetermined range, 1 is added to the similarity, and if it is not included, 1 is subtracted. The similarity may be calculated by performing the design parameters.

  Note that the similarity calculation unit 110 according to the embodiment may calculate the similarity for each design parameter included in the control value candidate instead of calculating one similarity for each control value candidate. In this case, the actual machine control unit 102 according to Embodiment 2 may determine that two control value candidates are similar if the similarity is equal to or greater than a predetermined threshold for at least one design parameter.

  In addition, in the case of Yes in step S110 illustrated in FIG. 9, the real machine control unit 102 according to the second embodiment uses an evaluation value in advance instead of discarding all control value candidates similar to the already evaluated control value candidates. Only the control value candidates that are less than the evaluation reference value defined in (1) may be discarded.

  In general, as a result of the evaluation that has already been made, a control value candidate that is similar to a control value candidate that is poorly evaluated is expected to have a poor evaluation. On the other hand, if the control value candidate is similar to the control value candidate with good evaluation, it can be expected that the optimal trajectory is included in the trajectory similar to the trajectory. Therefore, the actual machine control unit 102 according to the second embodiment has an evaluation value in which an evaluation value corresponding to one control value candidate is determined in advance among any two control value candidates included in the plurality of control value candidates. Only when it is less than the reference value, a plurality of control value candidates are determined so that one control value candidate and the other control value candidate are not similar to the similarity indicated by a predetermined threshold. Also good.

  For example, when the similarity calculation unit 110 calculates a larger similarity as the two control value candidates are similar, the actual machine control unit 102 according to the second embodiment is included in the plurality of control value candidates. Of the two arbitrary control value candidates, only when the evaluation value corresponding to one control value candidate is less than a predetermined evaluation reference value, one control value candidate and the other control value candidate A plurality of control value candidates may be determined such that the similarity is less than a predetermined threshold.

  In addition, although the actual machine control unit 102 according to the present embodiment uses SHA as a search algorithm for an optimal control value candidate, the search algorithm used by the actual machine control unit 102 is not limited to this. For example, other well-known heuristic search algorithms such as annealing and tabu search may be used.

  The actual machine control unit 102 may generate control value candidates by randomly determining each control value. Even in this case, the actual machine control unit 102 can achieve the effects of the invention by discarding similar control value candidates without executing an excessive number of repetitions. However, it is considered that the number of iterations until convergence to the optimum value is increased as compared with the case of using the search algorithm such as the above SHA.

  In the comparison process with the threshold value in the present embodiment, the determination “above” may be determined as “beyond”, and the determination “below” may be determined as “less than”. Conversely, the determination “over” may be determined as “over”, and the determination “under” may be determined as “under”.

  Note that the robot motion method determination device 100 described in this embodiment can also be realized by a computer. FIG. 13 is a block diagram illustrating a hardware configuration of a computer system that implements the robot motion method determining apparatus 100.

  Referring to FIG. 13, robot movement method determining apparatus 100 includes a computer 34, a keyboard 36 and mouse 38 for giving instructions to computer 34, and a display 32 for presenting information such as the calculation results of computer 34. And a CD-ROM (Compact Disc-Read Only Memory) device 40 for reading a program executed by the computer 34 and a communication modem (not shown).

  A program, which is a process performed by the robot movement method determination device 100, is stored in the CD-ROM 42, which is a computer-readable medium, and is read by the CD-ROM device 40. Alternatively, the data is read by the communication modem 52 through a computer network.

  The computer 34 includes a CPU (Central Processing Unit) 44, a ROM (Read Only Memory) 46, a RAM (Random Access Memory) 48, a hard disk 50, a communication modem 52, and a bus 54.

  The CPU 44 executes a program read via the CD-ROM device 40 or the communication modem 52. The ROM 46 stores programs and data necessary for the operation of the computer 34. The RAM 48 stores data such as parameters at the time of program execution. The hard disk 50 stores programs and data. The communication modem 52 communicates with other computers via a computer network. The bus 54 connects the CPU 44, the ROM 46, the RAM 48, the hard disk 50, the communication modem 52, the display 32, the keyboard 36, the mouse 38, and the CD-ROM device 40 to each other.

  Furthermore, some or all of the constituent elements constituting each of the above-described devices may be configured by a single system LSI (Large Scale Integrated Circuit). The system LSI is an ultra-multifunctional LSI manufactured by integrating a plurality of components on a single chip, and specifically, a computer system including a microprocessor, ROM, RAM, and the like. . A computer program is stored in the RAM. The system LSI achieves its functions by the microprocessor operating according to the computer program.

  Furthermore, some or all of the constituent elements constituting each of the above-described devices may be configured from an IC card that can be attached to and detached from each device or a single module. The IC card or module is a computer system that includes a microprocessor, ROM, RAM, and the like. The IC card or the module may include the super multifunctional LSI described above. The IC card or the module achieves its function by the microprocessor operating according to the computer program. This IC card or this module may have tamper resistance.

  Further, the present invention may be the method described above. Further, the present invention may be a computer program that realizes these methods by a computer, or may be a digital signal composed of the computer program.

  Furthermore, the present invention provides a computer-readable recording medium such as a flexible disk, hard disk, CD-ROM, MO, DVD, DVD-ROM, DVD-RAM, BD (Blu-ray Disc). (Registered trademark)), a memory card such as a USB memory or an SD card, or a semiconductor memory. Further, the digital signal may be recorded on these recording media.

  Further, the present invention may transmit the computer program or the digital signal via an electric communication line, a wireless or wired communication line, a network represented by the Internet, a data broadcast, or the like.

  The present invention may be a computer system including a microprocessor and a memory, wherein the memory stores the computer program, and the microprocessor operates according to the computer program.

  In addition, the program or the digital signal is recorded on the recording medium and transferred, or the program or the digital signal is transferred via the network or the like and executed by another independent computer system. You may do that.

  Furthermore, the above embodiment and its modifications may be combined.

  The embodiment disclosed this time should be considered as illustrative in all points and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.

  The present invention can be applied to a robot motion method determination device, and in particular, to a robot motion method determination device that determines a motion method for moving a robot from a predetermined first position to a second position. it can.

32 Display 34 Computer 36 Keyboard 38 Mouse 40 CD-ROM device 42 CD-ROM
44 CPU
46 ROM
48 RAM
50 Hard Disk 52 Communication Modem 54 Bus 100 Robot Motion Method Determination Device 102 Actual Machine Control Unit 104 Actual Value Acquisition Unit 106 Evaluation Unit 108 Control Value Determination Unit 110 Similarity Calculation Unit 112 Evaluation Storage Unit 200 Actual Machine 202 Measurement Device 550 Heuristic Method 552 , 750, 752 Trajectory 702, 704, 706, 708, 710, 712 Search range 900 Computer 902 Dynamic model 904 Heuristic method 906 Trajectory

Claims (9)

An apparatus for determining a movement method of a robot for determining a movement method for moving a robot from a predetermined first position to a second position,
A control value candidate that is a set of control value candidates for controlling a plurality of drive units included in the robot is determined and determined so that the robot moves from the first position to the second position. An actual machine control unit that transmits the control value candidates to the robot;
An actual measurement value acquisition unit that acquires an actual measurement value that is a physical quantity indicating a drive state of the plurality of drive units controlled by the control value candidate;
An evaluation unit that determines an evaluation value of the control value candidate corresponding to the actual measurement value from the actual measurement value;
And a control value determining unit that determines whether or not to adopt the control value candidate as the set of control values for determining the robot motion method based on the evaluation value. . Furthermore, an evaluation storage unit is provided,
The actual machine control unit determines a plurality of control value candidates different from each other,
The evaluation unit stores each of the plurality of control value candidates in the evaluation storage unit in association with the evaluation value of the control value candidate,
The control value determination unit determines to adopt the control value candidate corresponding to the evaluation value with the highest evaluation among the plurality of evaluation values stored in the evaluation storage unit as the set of control values. The robot motion method determination device according to claim 1. Furthermore, a similarity calculation unit that calculates the similarity between the two control value candidates is provided,
The actual machine control unit is configured to prevent the similarity between any two control value candidates included in the plurality of control value candidates from being similar to the similarity indicated by a predetermined similarity threshold. The robot motion method determining apparatus according to claim 2, wherein control value candidates are determined. The control value is a value corresponding to at least one of a driving time, a driving distance, a driving speed, a driving acceleration, and a position of the driving unit,
The actual measurement value is at least one of a current value, electric power, and electric energy,
The evaluation unit determines the evaluation value to be larger as the sum of the actual measurement values or the peak value of the actual measurement values in the plurality of driving units is smaller,
The similarity calculation unit calculates the similarity as a reciprocal of the sum of absolute values of differences between the control value candidates corresponding to the two control value candidates;
The actual machine control unit determines the plurality of control value candidates such that the similarity between any two control value candidates included in the plurality of control value candidates is less than a predetermined threshold;
The control value determination unit determines to adopt, as the set of control values, the control value candidate corresponding to the largest evaluation value among the plurality of evaluation values stored in the evaluation storage unit. 4. The robot motion method determination device according to 3. The actual machine control unit, when an evaluation value corresponding to one control value candidate among any two control value candidates included in the plurality of control value candidates is less than a predetermined evaluation reference value The plurality of control value candidates are determined such that the one control value candidate and the other control value candidate are not similar to the similarity indicated by a predetermined threshold value. Robot movement method determination device. A method for determining a movement method of a robot for determining a movement method for moving the robot from a predetermined first position to a second position,
A control value candidate that is a set of control value candidates for controlling a plurality of drive units included in the robot is determined and determined so that the robot moves from the first position to the second position. An actual machine control step of transmitting the control value candidates to the robot;
An actual value acquisition step of acquiring an actual value that is a physical quantity indicating a driving state of the plurality of driving units controlled by the control value candidate;
An evaluation step for determining an evaluation value of the control value candidate corresponding to the actual measurement value from the actual measurement value;
A control value determining step for determining whether to adopt the control value candidate as the set of control values for determining the robot motion method based on the evaluation value. . A program for causing a computer to execute the method for determining the movement method of the robot according to claim 6. A computer-readable recording medium on which the program according to claim 7 is recorded. An integrated circuit for determining a motion method for moving a robot from a predetermined first position to a second position, comprising:
A control value candidate that is a set of control value candidates for controlling a plurality of drive units included in the robot is determined and determined so that the robot moves from the first position to the second position. An actual machine control unit that transmits the control value candidates to the robot;
An actual measurement value acquisition unit that acquires an actual measurement value that is a physical quantity indicating a drive state of the plurality of drive units controlled by the control value candidate;
An evaluation unit that determines an evaluation value of the control value candidate corresponding to the actual measurement value from the actual measurement value;
An integrated circuit comprising: a control value determining unit that determines whether or not to adopt the control value candidate as a set of control values for determining a motion method of the robot based on the evaluation value;

Images