JP2021174426A - Source code automatic generation system - Google Patents

Abstract

To provide a source code automatic generation system and a method which can generate a source code of a mounting model including a code in a part depending on a device at a high execution rate while suppressing consumption of a storage capacity.SOLUTION: A source code automatic generation system 11 includes: an automatic code generation part 18 for generating a control application source code 19 on the basis of a block describing control processing operation; a device analysis part 20 for analyzing target information 21b related to a CPU to be mounted and device information 21a related to an IO device, which are included in a controller model; a device driver generation part 22 for generating only a device driver source code 24 required for mounting the IO device corresponding to the CPU from a device driver library 23 on the basis of the information analyzed by the device analysis part; and a code coupling part 25 for outputting the control application source code and the device source code as an integral source code 26.SELECTED DRAWING: Figure 1

Description

The present invention relates to an automatic source code generation system and method, and more specifically, includes a device-dependent part code built in a controller (microcomputer) by using a program for automatically generating source code in model-based development. The system and method for automatically generating the source code of the implementation model.

In the field of electronic device development that requires advanced control of automobiles, aircraft, robots, etc., the control contents and control targets of electronic devices are modeled by mathematical formulas, and control devices and control programs are simulated using these models. There is a method called model-based development that develops and verifies.

For model description and simulation in model-based development, for example, MATLAB (registered trademark) provided by MathWorks (registered trademark) that can easily calculate the matrix required for simulation, and the controller model graphically in a block diagram. It is done by software such as descriptive Simulation®.

By using MATLAB (registered trademark) / Simulink (registered trademark), the developer can simulate the designed controller model on a computer and verify the validity of the processing operation to be realized, and also verified it. For the controller model, for example, by applying an automatic code generation tool such as TargetLink (registered trademark) provided by dSPACE, it is possible to automatically generate source code that can be controlled by an actual controller.

However, in order to actually execute control according to the controller model, the existence of various devices such as peripheral devices such as sensors and actuators and IO devices for exchanging those signals is essential, so control is executed. The source code to be executed also requires the code of the device driver, which is a dedicated processing routine for adapting the application (control software) to the device.

However, the automatic code generation tool does not support the automatic generation of device drivers with different specifications for each device, and it was necessary to manually code the device driver implementation. Therefore, a technique capable of automatically generating source code for such a device-dependent part has been proposed (see, for example, Patent Document 1).

For example, in the development of an ECU that controls each function of an automobile, various IO devices (ADC, DAC, DIO, PWM, etc.) are used in the controller (microcomputer) mounted on the ECU in order to realize a function of cooperating with the outside. ) Is built-in.

In this case, according to the method described in Patent Document 1, each device is modeled as an auxiliary block, and a library in which each driver is registered is prepared. Furthermore, the auxiliary block is provided with a driver interface section that extracts the driver from the library, and by targeting the driver interface section for automatic code generation, each auxiliary block corresponds to the controller model to which the auxiliary block is connected. The source code combined with the driver of the device to be used can be automatically generated.

Japanese Patent No. 5137367

However, in the method described in Patent Document 1, when the controller is changed or the device built in the controller is changed, it is necessary to change the corresponding driver as well, and the auxiliary block is provided accordingly. Since it is also necessary to change the driver interface part, it took time and effort to correct these changes.

For this, for example, it is conceivable to provide a wrapper that makes the driver interface part that accesses the library independent from the auxiliary block. In that case, the auxiliary block is provided with a wrapper interface section for accessing the wrapper instead of the driver interface section, and a driver interface section is provided on the wrapper side.

Further, the wrapper can be fetched from the library via the driver interface section, and the driver used when the auxiliary block performs input / output processing is configured so that the wrapper interface section reads from the wrapper. Then, the wrapper interface part is targeted for automatic code generation.

By reconstructing the relationship between the auxiliary block and the library in this way, the device dependence of the auxiliary block itself can be eliminated, and the time and effort required for modification when changing the device can be reduced. However, since the wrapper takes out the driver, there is a problem that the execution speed is deteriorated.

Furthermore, even if such measures are taken, in order to support various devices that may be used, a library in which drivers for all of these devices are registered is required.

For this reason, electronic computing devices such as microcomputers with a built-in CPU do not actually use device drivers corresponding to various devices in case the controller configuration or device type is changed. There was a problem that wasted resources (storage capacity) were required when implementing the model.

Therefore, the present invention provides an automatic source code generation system and method capable of automatically generating the source code of an implementation model including the code of a device-dependent part by suppressing the consumption of storage capacity and at a high execution speed. I am aiming.

In order to achieve the above object, the source code automatic generation system of the present invention describes the control processing operation related to the CPU by mounting the source code to be mounted on the CPU built in the microcomputer inside the ECU (Electric Control Unit). A source code automatic generation system that is automatically generated based on a controller model having a block and an IO block corresponding to an IO device that enables the CPU to be connected to other devices, and describes the processing operation of the control. An automatic code generator that generates control application source code based on blocks, a device that analyzes target information that is included in the controller model and is information about the CPU to be implemented, and device information that is information about the IO device. From the template of the device driver library, which is a database in which the template of the driver code is stored, based on the analysis unit, the target information analyzed by the device analysis unit, and the device information, the IO device corresponding to the CPU can be used. It includes a device driver generator that generates only the device driver source code required for implementation, and a code coupling unit that combines the control application source code and the device drive source code and outputs it as an integrated source code. It is characterized by that.

Further, in order to achieve the above object, in the source code automatic generation method of the present invention, the source code for mounting the source code to be mounted on the CPU built in the microcomputer inside the ECU (Electric Control Unit) is subjected to the control processing operation related to the CPU. A source code generation method that automatically generates a source code based on a controller model having a block to be described and an IO block corresponding to an IO device that enables the CPU to be connected to other devices, and describes the processing operation of the control. The control application source code based on the block is generated by the automatic code generation program, and the target information which is the information about the CPU to be implemented and the device information which is the information about the IO device included in the controller model are analyzed. , Only the device driver source code required to implement the IO device corresponding to the CPU from the device driver library template, which is a database in which the driver code template is stored, based on the analyzed target information and the device information. Is generated, and the control application source code and the device drive source code are combined to output as an integrated source code.

According to the source code automatic generation system and method of the present invention, the device analysis unit configures only the necessary device driver source code based on the device information and the target information, and integrates it by combining with the source code of the control application. Since it can be output as source code, the storage capacity of the library is small and the execution speed is fast.

The block diagram which shows one form example of the source code automatic generation system. The block diagram which shows typically the control system including an ECU. The block diagram which shows the application model which modeled the CPU and IO device in the ECU of FIG. FIG. 6 is a block diagram schematically showing a flow in which each device driver is generated in an implementation model of a control system including the application model of FIG. A flowchart showing the integrated source code generation process.

FIG. 1 is a block diagram showing an example of one form of the source code automatic generation system 11 of the present invention. As shown in FIG. 1, the source code automatic generation system 11 includes a control design unit 12 which is software that operates on a computer and can be operated by a user to design a control and verify the control contents. There is.

The control design unit 12 reads the model 13 for simulation including the controller model 13a in which the control logic of the controller is described by the block diagram and the plant model 13b in which the state of the controlled object is described by the block diagram. The various IO blocks 14 that model the various IO devices necessary for connecting the controller and the controlled object to each other at the time of mounting the model 13 are registered.

A device information tag 14a, which is metadata for search / analysis, is set in each of the various IO blocks 14, and the type of IO device corresponding to the IO block 14 can be determined from the outside by reading the device information tag 14a. It is possible.

Further, the control design unit 12 is provided with a model editing unit 15, a simulation processing unit 17, and an automatic code generation unit 18.

The model editing unit 15 is a GUI (graphical user interface) capable of performing editing operations such as connecting the read model 13 and the registered IO blocks 14 and changing the settings of the IO blocks 14 by the user's operation.

Further, the model editing unit 15 can create an implementation model 16 capable of executing feedback control simulation by forming a closed loop in which the controller model 13a, the plant model 13b, and the IO block 14 are interconnected. It is configured in.

Further, in the model editing unit 15, the target CPU (microcomputer), that is, the type of CPU to be mounted can be specifically specified by information such as a model number, and the information to that effect is the mounting model 16. The block that models the CPU included in the above is set as a CPU information tag 16a that can be discriminated from the outside.

The simulation processing unit 17 can verify the operation of the mounting model 16 created by the model editing unit 15 by simulation, and is configured to be able to display the simulation result to the user.

The automatic code generation unit 18 automatically generates the control application source code 19 that enables the controller model 13a or the plant model 13b of the mounting model 16 whose operation has been verified by the simulation processing unit 17 to be executed as a control application, respectively. It is configured to be possible.

Further, the source code automatic generation system 11 includes a device analysis unit 20 capable of acquiring information from the control design unit 12. The device analysis unit 20 reads the mounting model 16 including the device information tag 14a and the CPU information tag 16a constructed by the model editing unit 15 from the control design unit 12, analyzes the configuration of the mounting model 16, and analyzes the mounting model. It outputs device information 21a including information such as types, numbers, and settings of 16 IO devices, and target information 21b which is information of the CPU to be mounted.

Further, the source code automatic generation system 11 is recorded on a predetermined recording medium capable of communicating between the device driver generation unit 22 that generates the driver code and the device driver generation unit 22, and is a driver corresponding to an arbitrary IO device. It includes a device driver library 23, which is a database in which code templates are stored.

The device driver generation unit 22 reads the device information 21a and the target information 21b output from the device analysis unit 20, and acquires the driver code related to the IO device corresponding to the specific CPU as the target from the device driver library 23. , Only the device driver source code 24 necessary for mounting the IO device corresponding to the specific CPU incorporated in the mounting model 16 can be generated.

The source code automatic generation system 11 includes a control design unit 12 and a code coupling unit 25 capable of receiving information from the device driver generation unit 22.

The code coupling unit 25 is configured to be capable of generating integrated source code 26 in which the control application source code 19 generated by the automatic code generation unit 18 and the device driver source code 24 generated by the device driver generation unit 22 are combined. Has been done.

Hereinafter, the source code automatic generation system 11 will be described by taking as an example the development of an ECU (Electronic Control Unit) that operates a power window of an automobile.

FIG. 2 is a block diagram schematically showing a control system including an ECU that moves a power window. As shown in FIG. 2, the power window 31 includes a motor 32 that opens and closes the window, and an encoder 33 that outputs the internal state of the motor 32 as a signal.

The ECU 34 includes a motor drive circuit 35 that drives the motor 32, and a target CPU 36 that can mount a control application for outputting a signal to the motor drive circuit 35.

The switch 37 opens and closes the door window of the automobile. A DI signal 38, which is an on / off digital input signal, is transmitted from the switch 37, and an ENC signal 39 including a pulse signal or the like according to the rotation angle of the motor 32 is transmitted from the encoder 33 of the power window 31.

The target CPU 36 transmits a PWM signal 40, which is an output signal for driving the motor 32, and a DO signal 41, which is a digital output signal for designating forward / reverse rotation of the motor 32, to the motor drive circuit 35. It is configured to be possible.

Further, the target CPU 36 includes DI 42, which is an IO device that enables reception of the DI signal 38, ENC 43, which is an IO device that enables reception of the ENC signal 39, and PWM Out 44, which is an IO device that enables transmission of the PWM signal 40. And DO45, which is an IO device that enables transmission of the DO signal 41, are built-in.

FIG. 3 is a block diagram showing an application model (controller model) 46 in which the target CPU 36 and each IO device in the ECU 34 of FIG. 2 are modeled in the control design unit 12 of FIG.

As shown in FIG. 3, the application model 46 describes a control logic description block 47 (present) in which an algorithm for determining how to drive the motor 32 in response to an input signal sent from the switch unit 37 to the target CPU 36 is described. A block that describes the processing operation of the control of the invention) is provided.

Further, in the control logic description block 47, DI blocks 48a and 48b, which are a type of IO block 14, and ENC block 49 are connected to the input side, and PWMOut block 50 and DO block 51 are connected to the output side, respectively. ing.

Here, the DI blocks 48a and 48b are models of the DI 42, the DI block 48a corresponds to the UP of the switch 37, and the DI block 48b corresponds to the DOWN of the switch 37. The ENC block 49 is a model of the ENC 43. Further, the PWM Out block 50 is a model of the PWM 44, and the DO block 51 is a model of the DO 45.

FIG. 4 is a block diagram schematically showing a flow in which each device driver is generated in the ECU mounting model 52 including the application model 46 of FIG. 3 designed by the control design unit 12 of FIG.

As shown in FIG. 4, the ECU mounting model 52 is a form of the mounting model 16 formed by modeling the power window 31 of FIG. 2, the ECU 34, and the switch 37 and combining them with the model editing unit 15 of FIG. Is.

The device analysis unit 20 analyzes the configuration of the ECU mounting model 52, and based on the CPU information tag 16a attached to the application model 46 and the device information tag 14a attached to the IO block 14, the type of the target CPU 36 And the number and type of IO devices connected to the target CPU 36.

Then, the type of the specified target CPU 36 is output as the target information 21b, and the setting contents of the specified IO device are output as the device information 21a.

That is, the device analysis unit 20 acquires the information of the DI 42 from the DI block 48, the information of the ENC 43 from the ENC block 49, the information of the PWM 44 from the PWM Out block 50, and the DO block 51 in the application model 46. The DO45 information can be obtained from the device information 21a and included in the device information 21a.

Further, the device driver generation unit 22 obtains the source code of each device driver (DI driver, ENC driver, PWMOut driver, DO driver) corresponding to the target CPU 36 from the device driver library 23 based on the device information 21a and the target information 21b. (Template) can be acquired, modified based on the device information 21a and the target information 21b, and then included in the generated device driver source code 24.

For example, in the case of a DO driver, after the DO driver template is acquired, the source code of the template is modified according to the output state at the end specified by the user in advance. In the case of an ADC driver, after the template of the ADC driver is acquired, the source code of the template is modified according to the A / D (analog / digital) conversion method specified in advance by the user. Is done.

Hereinafter, the process of generating the device-dependent integrated source code 26 by the source code automatic generation system 11 will be described with respect to the ECU mounting model 52.

FIG. 5 is a flowchart showing the generation process of the integrated source code 26. The device analysis unit 20 monitors the control design unit 12, and as shown in FIG. 5, the control design unit 12 controls the mounting model 16 (ECU mounting model 52) by the automatic code generation unit 18. When the generation of the application source code 19 (code of the control logic description block 47) is started (step S1), the device analysis of the mounting model 16 (ECU mounting model 52) is performed (step S2), and the target information and the device information 21a are performed. (Step S3).

When the device analysis unit 20 acquires the device information 21a and the target information 21b, the device analysis unit 20 sends the acquired device information 21a and the target information 21b to the device driver generation unit 22. When the device driver generation unit 22 receives the device information 21a and the target information 21b from the device analysis unit 20, the device driver generation unit 22 acquires a device driver template corresponding to the CPU based on the target information 21b from the device driver library 23 (step S4), and also obtains the device driver template. Based on the setting contents of each device included in the device information 21a, the acquired device driver is modified (step S5), and an appropriate device driver source code 24 is constructed (step S6).

When the device driver source code 24 is constructed, the device driver generation unit 22 sends the constructed device driver source code 24 to the code coupling unit 25. When the code coupling unit 25 receives the device driver source code 24 from the device driver generation unit 22, the automatic code generation unit 18 confirms whether the control application source code 19 has already been generated, and if it has not been generated, the code coupling unit 25 is generated. Wait until (step S7).

If the control application source code 19 is generated, the code coupling unit 25 receives the control application source code 19 from the automatic code generation unit 18 and combines the received control application source code 19 with the device driver source code 24. , Build integrated source code 26.

In this embodiment, the controller model 13a, which describes the control processing operation shown in FIG. 1, is mounted on the CPU in the ECU. However, in the present invention, the dynamic characteristics of the controlled object shown in FIG. The plant model 13b that describes the above can also be applied to a CPU in hardware in HILS (Hardware in the Loop Simulation).

As described above, according to the source code automatic generation system 11 of the present invention, the control design unit 12 has the controller model 13a or the plant model among the mounting models 16 in which the simulation model 13 and the IO block 14 are combined. When the control application source code 19 according to 13b is generated by the automatic code generation unit 18, the device analysis unit 20 acquires the device information 21a related to the IO block 14 and the target information 21b of the CPU to be mounted from the mounting model 16, and the device driver. The generation unit 22 generates the device driver source code 24 based on the device information 21a and the target information 21b, and further, the generated control application source code 19 and the device driver source code 24 are combined by the code coupling unit 25. Since it is output as the integrated source code 26, it is possible to operate a controller to which an arbitrary device is connected and automatically generate a device-dependent source code that suppresses the consumption of storage capacity from the implementation model.

The present invention is not limited to the above embodiment, and various modifications can be made within the scope of the invention. For example, in this embodiment, device analysis of the implementation model by the device analysis unit is performed according to the start of control application source code generation by the automatic code generation unit, but device analysis and automatic code generation for the same implementation model. If the resulting device driver source code and control application source code are combined, the execution of device analysis does not necessarily have to coincide with the start of control application source code generation. The user may manually determine the execution timing of.

Further, in this embodiment, the device driver source code is generated by the device driver generation unit, but it is not always automatically generated if an appropriate device driver source code is created based on the device information output from the device analysis unit. It does not have to be done, and it may be created by the user instead of the device driver generator.

11 ... Source code automatic generation system, 12 ... Control design department, 13 ... Model, 13a ... Controller model, 13b ... Plant model, 14 ... IO block, 14a ... Device information tag, 15 ... Model editorial department, 16 ... Implementation Model, 17 ... Simulation processing unit, 18 ... Automatic code generation unit, 19 ... Control application source code, 20 ... Device analysis unit, 21a ... Device information, 21b ... Target information, 22 ... Device driver generation unit, 23 ... Device driver library , 24 ... device driver source code, 25 ... code coupling, 26 ... integrated source code, 31 ... power window, 32 ... motor, 33 ... encoder, 34 ... ECU, 35 ... motor drive circuit, 36 ... target CPU, 37 ... Switch, 38 ... DI signal, 39 ... ENC signal, 40 ... PWM signal, 41 ... DO signal, 42 ... DI, 43 ... ENC, 44 ... PWM, 45 ... DO, 46 ... Application model, 47 ... Control logic description block, 48a, 48b ... DI block, 49 ... ENC block, 50 ... PWMOut block, 51 ... DO block, 52 ... ECU mounting model

Claims (2)

The source code for mounting on the CPU built in the microcomputer inside the ECU (Electronic Controller Unit) is attached to the block that describes the control processing operation related to the CPU and the IO device that enables the CPU to be connected to other devices. A source code automatic generation system that automatically generates based on a controller model that has a corresponding IO block.
An automatic code generator that generates control application source code based on a block that describes the processing operation of the control,
A device analysis unit that analyzes target information, which is information about the CPU to be mounted, and device information, which is information about the IO device, included in the controller model.
Based on the target information analyzed by the device analysis unit and the device information, it is necessary to implement an IO device corresponding to the CPU from the template of the device driver library which is a database in which the template of the driver code is stored. A device driver generator that generates only the device driver source code,
A code coupling unit that combines the control application source code and the device drive source code and outputs it as an integrated source code.
Source code automatic generation system with. The source code for mounting on the CPU built in the microcomputer inside the ECU (Electronic Controller Unit) is attached to the block that describes the control processing operation related to the CPU and the IO device that enables the CPU to be connected to other devices. A source code generation method that automatically generates based on a controller model that has a corresponding IO block.
A control application source code based on a block that describes the processing operation of the control is generated by an automatic code generation program.
The target information, which is information about the CPU to be mounted, and the device information, which is information about the IO device, included in the controller model are analyzed.
Based on the analyzed target information and the device information, only the device driver source code necessary for implementing the IO device corresponding to the CPU is extracted from the device driver library template, which is a database in which the driver code template is stored. Generate and
A source code generation method that outputs as integrated source code by combining the control application source code and the device drive source code.

Images