JP6164457B2 - Test system using network - Google Patents

Description

  The present invention relates to a test system using a network.

2. Description of the Related Art Conventionally, with the development of communication networks such as the Internet and Intranet in recent years, attention has been paid to a technology that can link various robots or a robot and an interface via a network (for example, see Patent Document 1).
Using such communication control technology, there are cases in which a coupling test is performed with the machine elements (engine, transmission, steering system, etc .; hereinafter referred to as “actual machine”) arranged in different locations. , Connect a simulation machine that can simulate the operation of the other actual machine, connect a simulation machine that can simulate the operation of the one actual machine to the other actual machine, connect each actual machine and each simulation machine via a network, Test by exchanging data and sharing data.

During testing, each simulator must operate in real time based on data obtained from a real machine operating at a different location.
FIG. 1 shows that a real machine A1 and a simulation machine B1 are installed at a certain place (base 1), a simulation machine B2 and a real machine A2 are installed at another place (base 2), and the simulation machine B1 and the real machine A1 are mechanically coupled. It is the schematic which shows the conventional test system which mechanically combined real machine A2 and simulation machine B2.

  The simulated machine B1 simulates the operation of the actual machine A2, and the simulated machine B2 simulates the operation of the actual machine A1. Therefore, the simulated machine B1 and the actual machine A2 are connected by a network, and the simulated machine B2 and the actual machine A1 are connected by the network.

JP 2006-00977 A

  However, when operating the simulator according to data obtained through the network, if the communication cycle (determined by the packet transmission cycle) is long, it corresponds to the communication cycle from when the actual device transmits data until the next data is transmitted. During this time, the simulator may not be able to update data, and the simulator may not be able to correctly simulate the state of the actual device. In particular, if the data value to be communicated is a content that changes abruptly, the follow-up for reproducing the state may not be in time, and the test operation may become unstable.

Thus, the communication delay adversely affects the entire test system using the network.
In view of such circumstances, the present invention takes a time corresponding to one communication cycle from the time when the actual device transmits data to the next data, during which the simulated device cannot update the data. Even so, an object of the present invention is to provide a test system in which the simulator can simulate the state of the actual machine as accurately as possible.

  In the present invention, the operation of the first actual machine (A1) that is a machine element of an automobile at one site and the second actual machine (A2) that is another machine element coupled to the first actual machine (A1). A first simulation machine (B1) to be simulated is installed, and a second simulation machine that simulates the operations of the second actual machine (A2) and the first actual machine (A1) at a second site away from the site. Machine (B2) is installed, the second actual machine (A2) and the second simulated machine (B2) are mechanically coupled, and the first simulated machine (B1) and the first actual machine (A1) are machined. The first simulated machine (B1), the second actual machine (A2), the second simulated machine (B2), and the first actual machine (A1) are connected to the network so that data communication is possible. The present invention relates to a test system for performing an operation test.

According to the present invention, the test system is connected to the first simulator (B1), and the first data processor (1) for capturing data from the second actual machine (A2) through the network for each communication period ( 11) and a second data processing device (21) connected to the second simulator (B2) and for capturing data from the first actual device (A1) through the network for each communication cycle,
The first data processing device (11) calculates a predicted value for interpolating the data taken in every communication cycle, and the second data processing device (21) uses the data taken in every communication cycle. A predicted value for interpolation is calculated.
Here, the “predicted value” means that the data processing device captures data for each control cycle shorter than the communication cycle after the data is captured in a certain communication cycle until the data is captured in the next communication cycle. This is the predicted value for interpolation.

Reference numerals such as (A) and (B) are reference numerals used in the accompanying drawings, and the present invention is not limited thereby (the same applies to the claims).
According to the test system of the present invention, the first data processing device (11) uses the data of the second actual machine (A2) interpolated using the predicted value, and the first simulator (B1). Can be calculated, and the first simulator (B1) can be controlled using this control target value. The second data processing device (21) calculates a control target value for the second simulator (B2) using the data of the first actual machine (A1) interpolated using the predicted value, and this The second simulator (B2) can be controlled using the control target value .

However, when an abnormal state occurs in which the predicted value is more than the threshold value from the data of the second actual machine (A2), the first data processing device (11) calculates the control target value using the predicted value. Then, since the control error increases, it is preferable to calculate the control target value using only the data of the second actual machine (A2) at this time.
In this case, the calculated control target value is transmitted to the second data processing device (21), and the second data processing device (21) uses the received control target value to generate a second simulator. It is more preferable to control (B2).

By providing such a countermeasure against an abnormal state, when the data of the second actual machine (A2) is more than a threshold value away from the predicted value, it is possible to quickly cope with the data of the second actual machine (A2). Control based on can be continued.
The first data processing device (11) that has found the abnormal state transmits a flag indicating the abnormal state together with the calculated control target value to the second data processing device (21). The second data processing device (21) can start control of the second simulator (B2) with the reception of this flag as a trigger.

The above measures, the predicted value of the first data processing device (11) is a measure of when it is an abnormal state of leaves above the threshold from the data of the second actual (A2), the second data The same countermeasure can be taken when an abnormal state occurs in which the predicted value in the processing device (21) is more than the threshold value away from the data of the first actual machine (A1).

The actual machine A1 and the simulator B1 are installed at the base 1, the simulator B2 and the simulator A2 are installed at other locations (base 2), the machine A2 and the simulator B2 are mechanically coupled, and the simulator B1 and the machine A1 are machined. 1 is a schematic diagram illustrating a combined conventional test system. FIG. The actual machine A1 and the simulator B1 are installed at the base 1, the simulator B2 and the simulator A2 are installed at other locations (base 2), the machine A2 and the simulator B2 are mechanically coupled, and the simulator B1 and the machine A1 are machined. 1 is a schematic diagram showing a combined test system 1 of the present invention. It is a block diagram of the test system 1a performed conventionally. It is a flowchart which shows the test procedure which the data processing apparatus 11 connected to simulation machine B1 performed conventionally. It is a graph showing the time transition of the measurement data of the torque or rotation speed obtained from the actual machine A2, the control target value for the simulator B1, and the measurement data of the torque or rotation speed measured by the measuring device 13. FIG. 6 is an enlarged view showing one communication period of FIG. 5. 1 is a configuration diagram of a test system 1a according to an embodiment of the present invention. It is a flowchart for demonstrating the test procedure of this invention which the data processor 11 connected to simulator B1 performs. It is a graph showing the data of the torque or rotation speed measured with the actual machine, and the predicted value calculated with the simulation prediction apparatus. FIG. 10 is an enlarged view showing one communication cycle of FIG. 9 taken out. It is an enlarged view showing a state in which a predicted value is separated from measurement data of an actual machine when the measurement data of the actual machine moves peculiarly. If the measured data of the actual machine at the local site deviates from the predicted value predicted by the simulation, when the measured data of the actual machine is captured in the next communication cycle, the measured value of the actual machine is used instead of the predicted value used so far. It is a flowchart for demonstrating the test procedure of this invention which calculates a control target value based on data.

Embodiments of the present invention will be described below with reference to the accompanying drawings.
FIG. 2 shows that the actual machine A1 and the simulator B1 are installed at the base 1, the simulator B2 and the simulator A2 are installed at another place (base 2), the actual machine A2 and the simulator B2 are mechanically coupled, and the simulator B1 it is a schematic diagram showing a test system 1 coupling the actual A1.
In the embodiment of the present invention, for example, the actual machine A2 is a front-wheel drive-based 4WD vehicle, and the front and rear wheels are continuously driven by electronic control from front wheel 100: front wheel drive of rear wheel 0 to front wheel 50: rear wheel 50. The electronically controlled coupling mechanism can be varied, and the actual machine A1 is a prime mover coupled to the coupling mechanism.

The simulator B2 is a motor that simulates a prime mover, and the simulator B1 is a motor that simulates an electronically controlled coupling mechanism. The simulated machine B1 and the actual machine A1 are mechanically coupled by a shaft, and the simulated machine B2 and the actual machine A2 are also mechanically coupled by a shaft.
However, the machine element is not limited to such embodiments, it is possible to select any mechanical elements of a vehicle are linked together actual A2, as actual A1.

The simulated machine B1 and the actual machine A2 are connected via a network, and either or both of the rotational speed and rotational torque of the shaft are transmitted as measurement data. The simulated machine B2 and the actual machine A1 are also connected by a network, and either or both of the rotational speed and rotational torque of the shaft (hereinafter referred to as “torque or rotational speed”) are transmitted as measurement data through the network.
There are also provided models C1 and C2 for predicting the operations of the real machine A2 and the real machine A1, respectively, and providing data of their predicted values. The model C1 is connected to the simulator B1, and the simulator B1 receives data of the predicted value of the model C1. A model C2 is connected to the simulated machine B2, and the simulated machine B2 receives data of predicted values of the model C2.

Hereinafter, a specific configuration of the test system 1 will be described.
First, the configuration of a conventional test system 1a will be described with reference to FIG. FIG. 3 (a) receives the torque or rotation speed data from the actual machine A1, the simulation machine B1, and the actual machine A2 installed at the base 1, and gives the torque control target value or the rotation speed control target value to the simulation machine B1. FIG. 2 is a block diagram showing a data processing device 11, a propeller shaft 12 that couples an actual machine A1 and a simulated machine B1, and a measuring instrument 13 that measures the torque or the rotational speed of the propeller shaft 12. As described above, the actual machine A1 is a prime mover, and the simulated machine B1 is a motor.

FIG. 3 (b) shows the simulation machine B2 and the actual machine A2 installed at the base 2 and the torque or rotation speed data received from the actual machine A1 through the network, and the simulation machine B2 receives the torque control target value or the rotation speed control target value. FIG. 2 is a block diagram showing a data processing device 21 for providing the power, a propeller shaft 22 that couples the simulator B2 and the actual machine A2, and a measuring instrument 23 that measures the torque or the rotational speed of the propeller shaft 22. As described above, the actual machine A2 is an electronically controlled coupling mechanism, and the simulated machine B2 is a motor. A flywheel 25 simulating a rear wheel tire is coupled to the actual machine A2.

  The test procedure performed by the data processing apparatus 11 connected to the simulator B1 in the above configuration will be described with reference to FIG. 4 showing a flowchart. When the data processing device 11 receives torque or rotation speed data from the real machine A2 through the network every communication cycle (communication cycle is not limited, for example, 10 msec) (step S1), the torque or rotation speed is changed accordingly. A control target value is calculated (step S2). The data processing device 11 updates the control target value calculated last time and stored in the memory in the data processing device 11 using the calculated control target value (step S3). Then, using this updated control target value, motor drive control of the simulator B1 is performed (step S4). The implemented control target value is recorded in the memory.

The test procedure performed by the data processing device 21 connected to the simulator B2 is performed in the same manner.
FIG. 5 is a graph showing the time transition of the torque or rotation speed measurement data obtained from the actual machine A2, the control target value for the simulator B1, and the torque or rotation speed measurement data measured by the measuring instrument 13. is there. The measurement data of the actual machine A2 is represented by a thin solid line, the control target value is represented by a thick staircase broken line, and the measurement data of the measuring instrument 13 is represented by a curved solid line. The communication cycle is indicated by a vertical broken line.

  As can be seen from FIG. 5, the measurement data of the actual machine A2 continuously changes, but the data processing device 11 calculates the control target value for the simulator B1 because of the network communication cycle as described above. The reception timing of the measurement data of the actual machine A2, which is the basis, is constrained by the communication cycle. Even if the measurement data of the machine A2 fluctuates finely, it cannot sufficiently follow it, and the control target value for the machine A2 changes stepwise. End up. That is, until the next communication cycle, the measurement data of the torque or the rotational speed of the actual machine A2 is fixed, and based on this, the control target value calculated at the beginning of the communication cycle is held at that value. Furthermore, depending on the content of the control performed by the data processing device 11 on the simulator B1, an overshoot of the measurement data measured by the measuring device 13 can be seen at the moment when the staircase rises.

  FIG. 6 is an enlarged view showing one communication period of FIG. Both ends (x marks) of one communication cycle indicate the time when the measurement data of the actual machine A2 is taken, and among them, each control cycle of the data processing device 11 is indicated by a circle. The control target value for the simulation machine B1 does not change during the communication cycle, corresponds to a flat part of the staircase, and becomes a constant value even though the measurement data of the actual machine A2 is moving. Therefore, it cannot be said that the actual machine reproduction accuracy of the simulation machine B1 is very good. Further, if the overshoot becomes large, the control cannot be performed at the end, and there is a risk of divergence. The communication cycle may be shortened, but the communication speed allowed for the network is limited.

Therefore, in the present invention, as shown in FIG. 7, the configuration of the test system 1 in which the simulation prediction device 14 is added to the data processing device 11 and the simulation prediction device 24 is added to the data processing device 21 is adopted. FIG. 7A shows the configuration at the site 1, and FIG. 7B shows the configuration at the site 2.
The configuration other than the simulation prediction devices 14 and 24 is the same as that shown in FIG. The simulation predicting device 14 predicts the movement of the measurement data for each control cycle after the data processing device 11 takes in the measurement data of the real machine A2 for each communication cycle and then takes in the measurement data. Then, the data processing device 11 calculates the control target value based on the measurement data of the actual machine A2 captured every communication cycle and the model state quantity data predicted every control cycle (hereinafter referred to as “predicted value”). . Similarly, the simulation prediction device 24 predicts the movement of the measurement data after the data processing device 21 captures the measurement data of the real machine A1 and then captures the measurement data. The control target value is calculated based on the predicted value.

  FIG. 8 is a flowchart for explaining the test procedure of the present invention performed by the data processing apparatus 11 connected to the simulator B1. When the data processing device 11 receives torque or rotation speed data from the real machine A2 through the network for each communication cycle (step T1), the simulation prediction device 14 assumes that the control cycle (the control cycle is shorter than the communication cycle). For example, every time the communication period is 10 msec, the control period is 2 msec.) Every time, the predicted value of the torque or the number of revolutions is calculated (step T2) and passed to the data processing device 11. The data processing device 21 calculates a control target value for the simulator B2 using the calculated predicted value (step T3). Furthermore, using this control target value, the control target value calculated last time and stored in the memory in the data processing device 11 is updated every control cycle (step T4). Then, using the updated control target value, the simulator B2 is controlled (step T5).

The test procedures performed by the data processing device 21 and the simulation prediction device 24 connected to the simulator B1 are performed in the same manner.
As described above, instead of using the torque or rotation speed data from the actual machine through the network for each communication cycle as it is during the communication cycle, the simulation prediction device is used to predict the change in the actual torque or rotation speed. Using the predicted data, the control target value for the simulator is calculated for each control cycle.

FIG. 9 is a graph showing torque or rotation speed data measured with an actual machine and a predicted value of the simulation prediction apparatus. The actual measurement data is represented by a thin solid line, and the predicted value is represented by a thick broken line. The communication cycle is indicated by a vertical broken line.
FIG. 10 is an enlarged view showing one communication period of FIG. Both ends (x marks) of one communication cycle indicate the time when the measurement data of the actual machine is taken, and among them, the control cycle of the data processing device is indicated by a circle. In this way, the control target value is calculated based on the predicted value, and thus exhibits good followability with respect to actual measurement data. As a result, the rising amount of the staircase as shown in FIG. 5 is reduced, and the overshoot is also reduced accordingly. The effect of improving the reproduction accuracy of the simulated machine is obtained.

  The content of the simulation prediction process performed by the simulation prediction apparatus is not limited, but, for example, the measurement data of the actual machine captured in the most recent communication cycle, and each measurement data of the actual machine captured N (N is an integer of 1 or more) times before Based on the above, approximately obtain a straight line close to each measurement data, find its slope, connect the straight line with that slope to the measured data of the actual machine captured in the most recent communication cycle, find the value in each control cycle, They can be used as predicted values in each control cycle.

  Next, the abnormal operation correction control according to the modified embodiment of the present invention will be described. In the test procedure of the present invention described above, the simulation prediction apparatus predicts the measurement data of the actual machine by performing a simulation prediction process. For example, when the measurement data of the actual machine moves peculiarly, the predicted value is the actual machine. It is also possible that the measurement data will be separated from the measured data. For example, FIG. 11 is an enlarged view showing this state, and the measurement data Xa of the actual machine changes greatly in the adjacent communication cycle (x mark), which is far from the predicted value Xm calculated for each control cycle (◯ mark). It has been. In such a case, if the control target value is obtained based on the predicted value, the control target value based on the predicted value used up to now and the actual The control target value calculated based on the measured data is greatly separated.

  Therefore, as shown in the flowchart of FIG. 12, for example, when the data processing apparatus 11 of the simulator B1 receives the data Xa of the measurement data of the real machine A2 through the network every communication cycle (step T1), the received data Xa is predicted. The value Xm is compared, and it is checked whether or not the absolute value of the difference exceeds the abnormal state correction threshold value Xc (step T11). If the absolute value of the difference does not exceed the abnormal state correction threshold value Xc, the abnormal state correction flag is turned off (step T12). Next, it is checked whether or not the abnormal state correction flag of the counterpart site transmitted from the counterpart site is turned on (step T13). If it is off, it is the same as shown in steps T2 to T5 of FIG. Then, the predicted value is calculated (step T2), the control target value for the simulator B1 is calculated using the calculated predicted value (step T3), and the data processing device is calculated the previous time using this control target value. The control target value stored in the memory in 21 is updated every control cycle (step T4), and the simulator B1 is controlled using the updated control target value (step T5).

  If the abnormal state correction flag of the partner site is on in step T13, the steady state control target value transmitted from the partner site is used as the control target value for the simulator B1 instead of using the predicted value (step S13). T14) Using this control target value, the simulator B1 is controlled (steps T3 to T5). Thus, the steady state control target value of the real machine A2 processed at the counterpart site and transmitted from the counterpart site is adopted. The reason for transmitting the steady state control target value without transmitting the measurement data such as the torque and the rotational speed of the actual machine A2 measured at the partner site is to reduce the communication load.

  If the absolute value of the difference between the data Xa and the predicted value Xm exceeds the abnormal state correction threshold value Xc in step T11, the process proceeds to step T15, and the abnormal state correction flag is turned on (step T15). Then, a control target value is obtained based on the measurement data Xa of the actual machine A2 instead of the predicted value Xm (step T16). This control target value is referred to as a “steady state control target value” in the sense of a control target value that is corrected to correct the abnormal state and return to the steady state.

  Further, the data processing device 21 transmits this steady state control target value to the data processing device 11 at the partner site (step T17). This is to set a flag so that both simulators are in an abnormal state when either one is shifted. The data processing device 21 does not use the predicted value, but uses the steady state control target value as the control target value for the simulator B1, and uses the control target value to control the simulator B1 (step T3). ~ T5).

  By adopting the above procedure, if the measured data of the actual machine deviates from the predicted value predicted by the simulation at the local site, the measured data of the actual machine will be used until the next communication cycle. Because the control target value is calculated based on the actual machine measurement data instead of the predicted value, the correction that the control target value for the simulated machine and the actual machine measurement data are largely separated at the local site Can be prevented. Therefore, it is possible to prevent a situation in which overshoot increases and control is diverged.

In addition, the fact that the measurement data of the actual machine at the local site has deviated from the predicted value predicted by the simulation is also notified to the data processing apparatus at the remote site, and the control target value for the simulated machine B2 and the real machine A1 are also transmitted at the remote site. It is possible to prevent the measurement data from being greatly separated by correction.
The present invention is not limited to the embodiments of the invention described above, and various modifications can be made without departing from the scope of the present invention. For example, the actual machine A2 may be a steering system, and the actual machine A1 may be a tire system. In this case, the simulator B1 is a motor that drives the tire to turn, and the simulator B2 is a motor that applies a reaction force of the tire to the rotation of the steering shaft. The actual machine A2 may be a wheel-in motor of an automobile, and the actual machine A1 may be a tire system.

DESCRIPTION OF SYMBOLS 1, 1a ... Test system, 11, 21 ... Data processing apparatus, 14, 24 ... Simulation prediction apparatus

Claims (4)

The first actual machine (A1) that is a machine element of an automobile at one base and the second actual machine (A2) that is another machine element coupled to the first actual machine (A1) And a second simulator (B2) for simulating the operations of the second actual machine (A2) and the first actual machine (A1) at a second base remote from the base. The second actual machine (A2) and the second simulator (B2) are mechanically coupled, the first simulator (B1) and the first actual machine (A1) are mechanically coupled, The operation test is performed so that data communication is possible by connecting the first simulator (B1), the second actual machine (A2), the second simulator (B2), and the first actual machine (A1) to the network. In the test system,
A first data processor (11) connected to the first simulator (B1) and for capturing data from the second actual machine (A2) through the network every communication period; and a second simulator (B2) And a second data processing device (21) for capturing data from the first actual machine (A1) through the network for each communication cycle,
The first data processing device (11) is for interpolating data for each control cycle shorter than the communication cycle from the time of taking in the data in the communication cycle to the time of taking in the data in the next communication cycle . The predicted value is calculated, and the second data processing device (21) takes every control cycle shorter than the communication cycle after taking the data in the communication cycle until it takes in the data in the next communication cycle. Calculate the predicted value to interpolate the data ,
The first data processing device (11) calculates a control target value for the first simulator (B1) using the data of the second actual machine (A2) interpolated using the predicted value, The first simulator (B1) is controlled using the control target value, and the second data processor (21) uses the data of the first actual machine (A1) interpolated using the predicted value. The control system calculates a control target value for the second simulator (B2) and controls the second simulator (B2) using the control target value . The first data processing device (11) uses only the data of the second real machine (A2) when the data of the second real machine (A2) is separated from the predicted value by a threshold or more. The test system according to claim 1 , wherein the control target value is calculated. The first data processing device (11) transmits the calculated control target value to the second data processing device (21), and the second data processing device (21) receives the received control target value. The test system according to claim 2 , wherein the second simulator (B2) is controlled by using. The test system according to claim 3 , wherein the first data processing device (11) transmits a flag indicating an abnormal state together with the calculated control target value to the second data processing device (21). .

Images