WO2017052059A1

WO2017052059A1 - Real-time control system, real-time control device and system control method

Info

Publication number: WO2017052059A1
Application number: PCT/KR2016/008031
Authority: WO
Inventors: 이정호; 오준호; 임정수
Original assignee: 주식회사 레인보우
Priority date: 2015-09-21
Filing date: 2016-07-22
Publication date: 2017-03-30

Abstract

A system control method, according to an embodiment of the present invention, comprises the steps of: operating by a plurality of agents using shared memory and having processes that are independent from each other; obtaining, from each reference generated from the plurality of agents and stored in the shared memory, hardware control data for controlling one or more devices; and transmitting a control signal, on the basis of the reference, to the one or more devices selected from the hardware control data.

Description

Real time control system, real time control device and system control method

The present invention relates to a real time control system, a system control apparatus and a system control method. More specifically, the present invention relates to a real-time control system, a system control apparatus and a system capable of accurate real-time processing, easy to develop and debug, and robust in hardware.

Currently, research on robots is being actively conducted in many institutions at home and abroad. Robots can be largely divided into hardware and software, and they are integrated to form a system.

Components of the robot hardware include a driver and a controller for moving the robot joint, a battery and a power controller, a communication module, a sensor, an exoskeleton of the robot, an electronic circuit, and a battery. These different types of elements are combined according to the characteristics of each desired robot to form a robot hardware platform.

Each of these elements will vary in size, appearance, performance, and manufacturer, depending on the design purpose, resulting in an endless variety of robotic hardware platforms. Indeed, there are many different robots around the world. Therefore, research is underway around the world to investigate the performance and function of a common target robot platform and to develop one representative standard robot platform.

In addition, there is a study on the standard robot software that can be used in common, as there is a study on the development of a standard robot platform as described above in hardware. Software to control robotic devices such as drive controllers and sensors belonging to the robot hardware, software to help develop robot motion easily, software to help robots operate by determining the relationship between tasks, software to provide navigation or vision solutions There are various types of software, and the development of standard software is being conducted for the purpose similar to the development of standard hardware.

However, in the development of the standard software, it is difficult to provide a standard solution for solving the common requirements of various developers.

In particular, as a lot of changes are made in hardware, it is not only difficult to collaborate among developers, but also a problem may occur that processes are overlapped by functions or processes developed by several people at the same time. Accordingly, the real-time synchronization is difficult, the stability or robustness of the hardware may be degraded, and the development time also takes a long time.

In addition, in hardware development, there is a problem in that driving performance may be degraded by adding an additional function change in order to maintain compatibility with the standard software.

The present invention is to solve the above problems, in a system that requires real-time, while several independent processes for the same hardware control and processing can coexist, the operation of the robot can be stably controlled accordingly, An object of the present invention is to provide a real-time system, a system control device, and a system control method that can provide robustness and scalability.

According to an aspect of the present disclosure, there is provided a method of controlling a system, the method including: operating a plurality of agents using a shared memory and having mutually independent processes; Obtaining hardware control data for controlling one or more devices from each reference generated from the plurality of agents and stored in the shared memory; And transmitting a control signal according to the reference to the one or more devices selected from the hardware control data.

In addition, an apparatus according to an embodiment of the present invention for solving the above problems, the system control apparatus, a plurality of agents having mutually independent processes; A shared memory storing a reference generated according to the operations of the plurality of agents; And obtaining hardware control data for at least one device control from each reference generated from the plurality of agents and stored in the shared memory, and controlling the control signal according to the reference to the at least one device selected from the hardware control data. It includes a device control module for delivering.

And, a real-time system for solving the above problems, one or more hardware devices including a joint or sensor device of the system; A plurality of agents having processes independent of each other in association with the system; A shared memory storing a reference generated according to the operations of the plurality of agents; From each reference generated from the plurality of agents and stored in the shared memory, hardware control data for controlling one or more devices is obtained, and the control signal according to the reference is sent to the one or more devices selected from the hardware control data. A device control module for transmitting; And a user system for managing real-time operating cycles and execution operations of the device control module and the agents.

According to an aspect of the present invention, there is provided a method of controlling a real-time device system, the method including: obtaining state information of a current step from a device to be controlled; Transferring the status information to one or more agents processing a reference operation using the status information of the current step; Obtaining, from the agent, reference information calculated beforehand based on state information of a previous step; And transmitting the reference information to the device as a control signal corresponding to the current step.

In addition, the apparatus according to an embodiment of the present invention for solving the above problems, in the device system control apparatus, obtains the state information of the current step from the device to be controlled, and reference by using the state information of the current step A device communication unit for transmitting the status information to one or more agents processing an operation; And a reference acquisition unit for obtaining reference information pre-calculated based on state information of a previous step from the agent, wherein the device communication unit transfers the reference information as a control signal corresponding to the current staff to the device.

In addition, a system for solving the above problems includes a device control system comprising: a plurality of agents having mutually independent processes; A shared memory storing a reference generated according to the operations of the plurality of agents; And obtaining the state information of the current step from the controlling device, and transferring the state information to the plurality of agents processing a reference operation using the state information of the current step, and from the agent to the state information of the previous step. And a device control module configured to obtain reference information calculated in advance based on the reference information, and to transmit the reference information to the device as a control signal corresponding to the current step.

On the other hand, the method for solving the above problems can be implemented with a program for executing the method on a computer and a recording medium on which the program is recorded.

According to an embodiment of the present invention, a plurality of agents having mutually independent processes and a shared memory in which references generated according to operations of the plurality of agents are stored are provided, and the reference to the hardware device is controlled using the reference. By providing a separate device control module, in a robot system requiring real-time, while several independent processes for the same hardware control can coexist, the robot operation can be stably controlled accordingly.

Accordingly, according to an embodiment of the present invention, even if each agent is independently developed, it is possible to synthesize and select a reference through a shared memory, thereby reducing the possibility of mutual collision and securing robust real-time. In addition, since agent replacement and real-time debugging are facilitated when an error occurs, collaboration convenience and scalability can be brought.

In addition, according to an embodiment of the present invention, in the thread synchronization between processes, by providing a method for processing one step delay (One Step Delay) using a shared memory, to minimize the processing delay between the device control module and the agent and the device, By optimizing the real-time synchronization process accordingly, it is possible to provide a control system that enhances the quality and real-time of the output signal and further refines and minimizes errors and jitter.

1 is a conceptual diagram schematically showing an entire system according to an embodiment of the present invention.

2 is a flowchart illustrating a control method of a robot system according to an exemplary embodiment of the present invention.

3 to 4 are diagrams for describing a relationship between a shared memory and a system according to an exemplary embodiment of the present invention.

5 is a diagram for explaining data exchange between a device control module and an agent according to an embodiment of the present invention.

6 is a block diagram illustrating a device control module according to an embodiment of the present invention.

7 is a flowchart illustrating a control operation of the robot system according to another exemplary embodiment.

8 illustrates a configuration of a device control module according to an embodiment of the present invention.

9 is a flowchart illustrating an operation of a device control module and agents according to an embodiment of the present invention.

10 is a timing diagram illustrating an operation timing between a device, a device control module, and an agent according to an embodiment of the present invention.

11 to 14 illustrate time elements and test results for performance verification of a control system according to an exemplary embodiment of the present invention.

The following merely illustrates the principles of the invention. Therefore, those skilled in the art, although not explicitly described or illustrated herein, can embody the principles of the present invention and invent various devices that fall within the spirit and scope of the present invention. Furthermore, all conditional terms and embodiments listed herein are in principle clearly intended for the purpose of understanding the concept of the invention and are not to be limited to the specifically listed embodiments and states. Should be.

In addition, it is to be understood that all detailed descriptions, including the principles, aspects, and embodiments of the present invention, as well as listing specific embodiments, are intended to include structural and functional equivalents of these matters. In addition, these equivalents should be understood to include not only equivalents now known, but also equivalents to be developed in the future, that is, all devices invented to perform the same function regardless of structure.

Thus, for example, it should be understood that the block diagrams herein represent a conceptual view of example circuitry embodying the principles of the invention. Similarly, all flowcharts, state transitions, pseudocodes, and the like are understood to represent various processes performed by a computer or processor, whether or not the computer or processor is substantially illustrated on a computer readable medium and whether the computer or processor is clearly shown. Should be.

The functionality of the various elements shown in the figures, including functional blocks represented by a processor or similar concept, can be provided by the use of dedicated hardware as well as hardware capable of executing software in association with appropriate software. When provided by a processor, the functionality may be provided by a single dedicated processor, by a single shared processor or by a plurality of individual processors, some of which may be shared.

In addition, the explicit use of terms presented in terms of processor, control, or similar concept should not be interpreted exclusively as a citation to hardware capable of running software, and without limitation, ROM for storing digital signal processor (DSP) hardware, software. (ROM), RAM, and non-volatile memory are to be understood to implicitly include. Other hardware for the governor may also be included.

In the claims of this specification, components expressed as means for performing the functions described in the detailed description include all types of software including, for example, a combination of circuit elements or firmware / microcode, etc. that perform the functions. It is intended to include all methods of performing a function which are combined with appropriate circuitry for executing the software to perform the function. The invention, as defined by these claims, is equivalent to what is understood from this specification, as any means capable of providing such functionality, as the functionality provided by the various enumerated means are combined, and in any manner required by the claims. It should be understood that.

The above objects, features, and advantages will become more apparent from the following detailed description taken in conjunction with the accompanying drawings, whereby those skilled in the art may easily implement the technical idea of the present invention. There will be. In addition, in describing the present invention, when it is determined that the detailed description of the known technology related to the present invention may unnecessarily obscure the gist of the present invention, the detailed description thereof will be omitted.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

Referring to FIG. 1, an entire system according to an embodiment of the present invention may include one or more devices 100, a device control module 200, a shared memory 300, one or more agents 400, and a user system 500. Include.

The device 100 may include one or more driving devices that finally perform the operation of the robotic system. The drive device may comprise a hardware device or a software device. The drive device may include, for example, at least one of a joint device, a sensor device including a sensor board, or a simulator device that controls drive for the articulated motor.

In addition, the device 100 may be controlled according to a control signal received from the device control module 200, and may output various data such as sensor data to the device control module 200. In addition, the term device 100 is not limited to hardware, but may be used as a concept including a software driver for driving an actual hardware device. Accordingly, each device 100 may be physically and softwarely connected to the device control module 200.

Each device 100 may form a communication network with the device control module 200. The communication network may form a system network using a controller area network (CAN) protocol for system stability.

For example, each device 100 may be connected to the device control module 200 through one or more CAN communication channels, and receive a message composed of CAN frame data according to a preset control period through the CAN communication channel. Alternatively, a message may be output to the device control module 200. Here, the message may include a motor control reference, an encoder value, a controller state value, a pulse width modulation (PWM) command, a sensor value, or various other setting or output values.

The device control module 200 obtains hardware control data for controlling one or more devices 100 from each reference generated from the plurality of agents 400 and stored in the shared memory, and the hardware control data. And transmits a control signal according to the reference to the one or more devices 100 selected from.

The device control module 200 may always reside on an operating system for controlling the robotic system and may be executed in the background.

The device control module 200 may uniquely directly communicate the device 100 with reference to the shared memory 300, and may transmit a control signal or receive a sensor signal through the communication channel.

For example, the device control module 200 may transfer a reference for controlling the joint device 100 to the joint device 100 or receive necessary sensor information from the sensor device 100.

In addition, the device control module 200 may include a real-time thread created on the operating system. The thread is synchronized with the motion generation operation cycle of the system to enable real time processing. In addition, the device control module 200 may further include a non-real time thread for processing data reading and conversion.

In addition, the plurality of agents 400 may be implemented as independent software modules having independent processes. For example, the agents 400 may each process different motions and perform a process for outputting a reference corresponding thereto. For example, each agent 400 may include a motion agent, a controller agent, a communication agent or a walking agent, a damping agent, and various other agents.

Since the agents 400 are functionally separated from each other, the agents 400 can create and operate respective threads without sharing heap, data, and static memory, and mutually share them. The necessary data for each can be provided to the shared memory 300, thereby allowing organic processing without mutual collision, and facilitates software development and processing.

In particular, according to an embodiment of the present invention, the plurality of agents 400 according to the operation of the mutually independent process, the sensor data from the shared memory 300 according to an embodiment of the present invention, each agent 400 Each may refer to hardware abstraction data and user-defined data of the shared memory 300 according to a defined process, and store the reference data generated based on the hardware abstraction data of the shared memory 300.

Here, the user defined data may include shared data for sharing information between the agents 400 and various data for driving other user-definable systems.

In addition, the hardware abstraction data may include abstracted reference, sensor data, motion owner variable, and command data to control the device 100. The device control module 200 may generate a control signal for each device 100 by using the hardware abstraction data and hardware information previously stored in the hardware database 250.

Accordingly, the device control module 200 identifies the control target device 100 using the hardware abstraction data extracted from the shared memory 300, and generates a control signal for the control target devices 100. The control signal according to the reference may be output to the control target device 100.

And, in order to ensure robust real-time, the processing cycle of each agent 400 needs to be shorter than the operation cycle of processing the motion information of the system. Accordingly, the agent 400 generates a reference from the sensor data, the device control module 200 generates and outputs a control signal from the reference through the shared memory 300, and the time for updating the sensor data is It may be included in the first operating period of the system. Thus, the series of operations can all be processed within the first operating period.

Meanwhile, the user system 500 may provide a user interface for controlling and monitoring the agent 400 and the device control module 200. In addition, the user system 500 may include middleware for controlling the agent 400, and may provide various interfaces that may be connected to other external systems.

Hereinafter, the robot system of the present invention and its control method will be described in more detail with reference to FIGS. 2 to 5.

Referring to FIG. 2, in the control method of the robot system according to an exemplary embodiment of the present invention, operations of a plurality of agents 400 having mutually independent processes are first performed (S101).

In operation S103, the device control module 200 obtains hardware abstraction data from a reference stored in the shared memory 300.

In addition, the device control module 200 generates a control signal for hardware control from the hardware abstraction data (S107), and transmits the generated control signal to one or more devices 100 (S109).

Meanwhile, the device control module 200 receives sensor data from the devices 100 corresponding to the sensor (S111), and updates the received sensor data in the shared memory 300 (S113).

In addition, the series of operation steps may be all processed within a first period corresponding to the real-time operation cycle of the robot system, thereby ensuring real-time.

For this process, as shown in FIG. 3, each of the agents 400 and the device control module 200 may perform data exchange and transfer processing using the shared memory 300.

According to the operation of each agent 400, the reference corresponding to each device 100 may be stored in the shared memory 300, and the device control module 200 may obtain a reference and use it to output a control signal. .

Such a plurality of agents 400 and the device control module 200 may configure a multi-agent system around the shared memory 300. As a result, each part performing independent work may be separately developed by several developers, or may have an advantageous structure in a robot system development environment in which they may collaborate.

By using the agent 400 according to an exemplary embodiment of the present invention, developers may use the shared memory 300 while interacting with the computational output of another agent 400 while ensuring a development space independent from the process concurrent development model. You will be able to send and receive.

4, hardware abstraction data and user-defined data may be included on the shared memory 300.

The hardware abstraction data may include sensor data, reference data, motion owner- and command data, and the device control module 200 may access only the hardware abstraction data area of the shared memory 300. .

Accordingly, the device control module 200 accesses the hardware abstraction data area of the shared memory 300 to update sensor data received from the device 100, or obtains updated reference data to control the device 100. You can generate a signal.

Here, the hardware abstraction data may have a data format converted by abstracting detailed data about the robot device control, and the device control module 200 may convert it into an actual hardware control signal and deliver it to the appropriate devices 100. have.

Accordingly, the agent 400 developer or user can facilitate the control without a deep understanding of the hardware. The developer or the user may transfer the abstracted hardware input information as a reference through the shared memory 300, and the device control module 200 may generate a low level control signal for controlling the device 100 from the hardware abstraction data.

In addition, since there may be scalability of hardware and changes and modifications of hardware, the device control module 200 may manage hardware information required to generate the control signal using the hardware database 250 described above. The hardware information may include, for example, a list of devices 100, joint motor information (deceleration ratio, encoder pulse, driver channel number, etc.), a communication protocol, and the like.

The device control module 200 may load the hardware database 250 to determine hardware information of the driving target device 100, thereby generating an optimal control signal for controlling the driving target device 100. can do. In addition, even if there is a change in hardware or using hardware of a new configuration, it is possible to apply only by modifying the hardware database 250, so that it is robust to hardware change and the hardware can provide scalable characteristics.

The hardware abstraction data may include reference data, sensor data, motion owner, and command data.

Here, the reference data may be updated according to the calculation result in each agent 400, and may include a target value in the current step for the device control module 200 to control each device 100. For example, the reference data may include a joint motion reference and a joint controller reference.

In addition, the sensor data may include measurement data that the device control module 200 receives from each device 100. The measurement data may include, for example, state information at a current step including an encoder value of the joint device and sensing data.

The command data may include command information for controlling the device control module 200 and the agent 400 at a higher system level, and may include command target process information and parameter information.

On the other hand, since the data of the other shared memory 300 is a value that is read by the agent 400, there is no confusion, but in the case of reference data, the values output for each agent 400 for the same joint device 100 can be different. In order to eliminate the possibility of collision, the shared memory 300 may include motion owner information.

5, in the data exchange between the device control module 200 and the agent 400 according to an embodiment of the present invention, a relationship between a motion owner and reference data may be described.

As illustrated in FIG. 5, the hardware abstraction data area of the shared memory 300 may include a memory area 350 for each agent 400 that can update reference data for each agent 400.

Accordingly, each agent 400 may update its calculated reference in its memory space area.

Here, each agent 400 may calculate and update reference data corresponding to each device 100. For example, when a total of 31 joint devices 100 exist from J1 to J31, a memory space area of each agent 400 may include a reference data area corresponding to each of the joint devices 100. .

In addition, the shared memory 300 may include a motion owner variable for each of the joint devices 100. Therefore, each motion owner variable space may include the same number of motion owner variables as the number of the joint devices 100.

Each motion owner variable may represent one agent having the authority to the joint device 100 among a plurality of preset agents 400. Accordingly, the device control module 200 may determine which agent 400 the control right for the joint device 100 depends on.

In addition, the control right for each joint device 100 may be transferred to another agent 400 or the device control module 200 according to the change of the motion owner variable.

To this end, the device control module 200 may first identify the agent 400 having the control right of the specific joint device 100 from the motion owner variable. The device control module 200 may collect reference data of the identified agent 400, and combine the collected reference data to generate overall reference data for the overall joint device 100.

In addition, the device control module 200 may generate a control signal for each device 100 by using the entire reference data, and may appropriately transmit the signal.

It will read the value of the corresponding joint reference, thereby constructing a reference for the entire joint of the robot, which is composed of only one set, and passing it to the robotic device for driving.

Through such a data exchange method, each joint of the robot can be controlled without collision in different agents 400. For example, if one agent 400 controls lower body joints through an algorithm for stabilizing lower body posture, and the other agent 400 generates a specific task motion of the upper body, the results of the two agents 400 are determined. In total, the whole body task of the robot can be performed. This enables efficient control according to the characteristics of the multi-agent system of the robot.

Referring to FIG. 6, the device control module 200 includes a motion selector 210, a controller signal accumulator 220, a signal combiner 230, and an information handler 240.

According to an embodiment of the present disclosure, the reference data for the joint may include two or more reference signals for joint motion control and detailed control. Accordingly, the agent 400 corresponding to each joint device 100 may generate the two or more reference signals as reference data and store the same in the shared memory 300.

In addition, as shown in FIG. 6, the reference signal may be referred to as a motion reference and a controller reference. The motion reference may include reference data that provides a dominant value for each joint, and the controller reference may include detailed reference data that is added to or subtracted from the motion reference. However, in the embodiment of the present invention, the reference is not limited to the name.

Accordingly, the motion reference output data M1 to Mm and the controller references M1 to Mm may be input to the device control module 200 from the shared memory 300.

In addition, one motion reference may be selected for each joint device 100, but all the controller references may be accumulated and added.

To this end, the motion selector 210 may select motion reference data corresponding to each joint device 100 from the motion reference data based on the motion owner variable information, and output the motion reference data to the signal combiner 230. have. Therefore, one motion reference data may be selected for one joint device 100.

In addition, the controller signal accumulator 220 may accumulate each controller reference data and output the result value to the signal combiner 230 regardless of the motion owner variable.

In addition, the signal combiner 230 may generate the reference data for each final joint device 100 by synthesizing the motion reference data and the controller reference data accumulated result value, and output them to the appropriate target joint devices 100. Can be.

Here, the signal combiner 230 may identify the type of the reference and classify the processing space according to the reference type. To this end, the signal combiner 230 may include a type identifier and a spatial processor.

For example, the reference data may have other types, such as task processing, as well as joint motion, such that a type identifier may identify whether the task type or the joint type is, and the spatial processor may determine each other according to the type. Can provide processing of other data spaces.

As such, by separating the motion reference from the controller reference, functional separation may be enabled in the process of generating the robot motion. For example, if a bipedal motion is generated, a basic walking pattern is generated in one agent 400 to generate a motion reference, a damping controller is designed in another agent 400, and another agent 400 is generated. By designing a controller to catch the vibration in the controller and outputting it to the controller reference, it is very easy to design and develop.

Meanwhile, the information handler 240 may perform a function of synthesizing sensor data collected from the sensor device 100 or other measurement target devices and outputting them to the shared memory 300.

In general, when a problem occurs in a real experiment using a robot, the robot must be driven again from the beginning. In the case of a mobile platform, the robot initialization process is simple, but when the initialization is difficult in the articulated system or the ground like a humanoid, and it is necessary to initialize it in the air by using a crane, etc., the entire initialization process is very cumbersome and time consuming. do.

Thus, as shown in Figure 7, the device control module 200 according to an embodiment of the present invention can debug and test the robot again without the process of initializing such a robot.

For this purpose, system initialization is first performed (S201), and a plurality of agents 400 having respective mutually independent processes operate (S202).

Thereafter, when a specific agent having an error exists (S203), the operation of the failed agent is stopped, and the device control module 200 or the other agent 400 changes the motion owner variable to another agent. (S205).

Accordingly, when the user tests the motion algorithm through the agent 400 and a problem is issued, the user simply passes the motion owner to another agent 400 or the device control module 200, and sends a code for the suspended agent 400. Can be modified.

When debugging is completed by compiling the created code (S207), the motion owner variable may be switched back to the original agent 400 (S209).

As such, the developer can continue the experiment after bringing the motion owner. As a result, it can accelerate development, and from the user's point of view, it can be further utilized to continuously observe the robot's joint reference on other special eggs to detect the collision and to switch the robot to the motion owner in case of a collision. It has the effect of allowing you to experiment safely.

8 illustrates a configuration of the device control module 200 according to an embodiment of the present invention.

In a real-time real-time control environment, in order to implement a strict hard real-time framework, real-time thread synchronization may be performed for the processes of the respective agents 400.

However, if synchronization does not occur properly between multiple processes during the time limited by each step (or period), unexpected jitter and jerk motion due to an update of one-step early joint reference or one-step late joint reference. This can happen. The more frequent these problems, the poorer the quality of motion or the more dangerous the overall task of the robotic system.

For example, when the device control module 200 and all the agents 400 are synchronized with a thread of a real time process having the same period as the control step, actual values of the output action may have some error. This causes a problem of not maintaining strong real time.

More specifically, the real-time threads included in each process are bound to be affected by jitter according to the operation time of each process, which causes a time axis shift between the processes. Although the device control module 200 synchronizes the most recent reference and sensor values, the device control module 200 may reflect the reference that is not updated according to the algorithm execution time, and degrade the quality of motion generation due to incorrect sensor data. As a result, they may not always be updated at offset time intervals according to a constant step period, or they may be updated too quickly, which may cause many problems.

Accordingly, the device control module 200 according to an exemplary embodiment of the present invention may acquire data using the shared memory 300 and one step delay and process synchronization between the agents 400 for thread synchronization between processes. By doing this, the above problems can be solved.

To this end, the device control module 200 performs step-based control, obtains the state information of the current step from the controlling device 100, and transmits the state information to one or more operation agents 400, and from the operation agent, In the previous step for the step, the reference information calculated in advance is obtained based on the state information of the previous step, and the reference information is transmitted to the device 100 as a control signal corresponding to the current step and calculated in the previous step. The reference motion is implemented in the device of the current step, and the sensor information of the current step is delivered to and processed by each operation agent 400, thereby enabling operation without time delay and jitter, based on one step period.

Accordingly, referring to FIG. 8, the device control module 200 according to an embodiment of the present invention may include a reference acquirer 250, a sensor device communicator 260, a synchronization signal processor 270, and a shared memory access unit 280. Include.

Here, each component of the device control module 200 may complete all control and operation processing in a control period corresponding to one step, and may perform the same processing again in the next step.

In addition, components of the device control module 200 may be processed using all or some functions of the above-described motion selector 210, controller signal accumulator 220, signal combiner 230, and information handler 240. May be implemented as a hardware processor or software module process.

First, the sensor device communication unit 260 is uniquely connected to each device 100 to transmit and receive control signals and sensor data between the device control module 200. Here, the sensor data may be a concept including state information of each device 100, and the state information may include not only information sensed by a separate sensor module but also operating environment information of the device 100.

Then, the sensor device communication unit 260 receives the sensor values updated in the current step from each device 100, and updates the sensor values in the sensor data of the shared memory 300 through the shared memory access unit 280. do.

In addition, according to the synchronization signal processed by the synchronization signal processor 270, each of the agents 400 may access the shared memory 300 to obtain the updated sensor values of the current step. In one embodiment, the sensor value of the current step may be used in the computation process for the reference of the next step.

In addition, the reference obtaining unit 250 shares the reference data previously calculated based on the state information (sensor data, etc.) of the previous step by each agent 400 in the previous step with respect to the current step as a reference of the current step. It may be obtained from the memory 300.

The sensor device communication unit 260 also generates a control signal of the current step by using the obtained reference data according to the synchronization signal received from each agent 400 by the synchronization signal processing unit 270, The generated control signal is transferred to each device 100.

Accordingly, the device control module 200 may process the motion control of the device 100 by converting the updated reference into the control signal of the current step in the shared memory 300 according to the state information of the previous step. This allows each of the agents 400 to operate separately according to a given sufficient control time, while the motion control signals for the devices 100 can be output directly in synchronization with the current step without any additional complex computations, thereby providing motion timebase error, jitter Alternatively, errors and the like can be solved, and motion quality can be improved according to stable operation.

9 is a flowchart illustrating the operation of the device control module 200 and the agent 400 according to an embodiment of the present invention.

Referring to FIG. 9, first, the device control module 200 receives a changed sensor value from each device 100 through the sensor device communication unit 260 to update the shared memory 300 (S301).

When the update is completed, the device control module 200 transmits a synchronization start signal SYNC SIGNAL to each agent 400 through the synchronization signal processor 270 (S303).

When the synchronization start signal is received, each of the agents 400 may obtain sensor data of the current step from the shared memory 300 and calculate a reference of the next step based on the obtained sensor data.

On the other hand, according to the synchronization start signal, each agent 400 updates the reference value calculated in the previous step on the shared memory 300 (S305).

The reference of the previous step may include a value calculated in advance from the sensor data of the previous step. Accordingly, the output time can be minimized because only the result value is updated in the current step.

When the update is completed, the device control module 200 receives a response signal (ACK SIGNAL) from each agent 400 through the synchronization signal processing unit 270 (S307).

Accordingly, the synchronization signal processor 270 of the device control module 200 determines whether a response signal is received from all agents 400 (S309), and when all are received, the reference obtainer of the device control module 200 is received. 250 obtains the updated previous step reference value on the shared memory 300 through the shared memory access unit 280, and transmits a control signal according to the reference value of the previous step to each device 100. Transfer (S311).

In operation S313, the device control module 200 requests a next sensor value through the sensor device communication unit 260.

Here, steps S301 to S311 may be all performed within one step corresponding to the operation cycle time of the robot system. In addition, the one step may be synchronized to the operation cycle of the device control module 200. Since the current sensor data based arithmetic process of each of the agents 400 is used to obtain a reference of the next step, the processing time can be minimized for the current step, and the synchronization can be maintained without error. The system according to the embodiment of the present invention can provide very robust hard real time. Table 1 below discloses a pseudo code for the synchronization operation of the device control module 200.

Table 1 shows the water code of one step cycle for generating the real-time motion of the joint device 100, the device control module 200 delivers the sensor data (s.data) received from the sensor (sensor) The joint reference calculated from the sensor data of the previous step may be obtained and transferred to each joint device. In Table 1, the device control module 200 may be represented as a daemon as a general control of each device 100, each agent 400 has a connection structure such as the eggs of grapes It may be expressed as AL. The algorithm of each agent 400 may be implemented by users and developers.

As shown in Table 1, when the device control module 200 (Deamon) operates a real time thread, i) sensor data is acquired in the current step, ii) a sync (SYNC) signal with each sensor data is generated. Iii-A) Each agent 400 (AL) updates the reference calculated in the previous step to the shared memory 300 to transmit an ACK signal, and iii-B) each agent. (400, AL) calculates the next step reference corresponding to the sensor data of the current step, iv) the device control module 200 (Deamon) delivers the updated reference to the articulation device (100). And, v) the sensor data of the next step may be requested to each device 100.

For example, as shown in Table 1, each of the agents 400 performs a task thread according to its own algorithm, and based on the sensor data of the current step, next agent (next refrence) Handles the operation. When the process is completed, it can wait until the next step. Here, when the next step is assumed to be the current step, the operation for the reference may be an operation preprocessed in the previous step.

In addition, when the agent 400 receives a synchronization start signal sync_flag through a sync watching thread, the agents 400 share a joint reference of a previous step generated in advance according to owner information of each agent. The memory device may be updated in the memory and the ack signal ack_flag according to the update result may be transmitted to the device control module 200.

In this synchronization process, the device control module 200 may use a one step delay to synchronize the control signal according to the joint reference with the sensor data and the process thread. This allows for robust and uniform device control, resulting in strict hard real-time. In addition, each agent 400 can fully utilize the time of a given control step period.

FIG. 10 is a timing diagram illustrating an operation timing between a device 100, a device control module 200, and an agent 400 according to an exemplary embodiment.

As shown in FIG. 10, first, the device control module 200 acquires sensor data for 5 ms, that is, a control step period, and obtains a reference value calculated by the agent 400 in a previous step. 100). In addition, data sharing between the device control module 200 and the agent 400 may be performed through input / output to the shared memory 300.

As shown in FIG. 10, the sensor data in the previous step n-1 is processed in the agent 400, and the working time may be shorter than the control step period. Accordingly, each agent 400 may have a suspend time. Then, when the control cycle of the current step starts, each agent 400 transfers the reference calculated in the previous step (n-1) to the device control module 200 according to the synchronization signal received from the device control module 200. The device control module 200 transmits the sensor data of the current step (n) to the agent 400. Each of the agents 400 processes an operation based on the sensor data of the current step n to generate reference data for the next step n + 1.

Each of the device control module 200 and the agents 400 may be implemented as a thread, and only the device control module 200 may include a thread having the same period of 5 ms as the control step period. This is because the role of the device control module 200 is a constant, the highest priority thread, so the time taken to take a given action is almost constant. 5 ms, which is a given control period, may be a very long time from the point of view of the device control module 200, and strict hard real-time may be implemented using this characteristic.

Accordingly, the operation of the device control module 200 may perform only a function of minimizing the variation of the time offset value from the start point of the thread of all actions. Accordingly, it is possible to increase the quality of the hard real-time control of various systems by ensuring the validity of the values due to each action.

11 illustrates three temporal components measured for system performance testing according to an embodiment of the invention.

First of all, 1) the period of the real-time thread of the device control module 200 was measured, and through this, it can be checked whether the real-time thread repeatedly guarantees a uniform period. Next, 2) the offset from the start of the thread of time the articulation reference is delivered to device 100 is measured. Through this, we can check whether we implemented strict hard real-time with almost no action delay. Finally, 3) Sync between device control module 200 and agent 400. Send a signal, Ack. The time taken for thread synchronization to receive a signal was measured. It indicates system responsiveness and you can see the effect of this value on the second measure of action delay. The agents 400 running this test were run when the number was 1, 3, 5, or 10 and obtained 4000 samples in each experiment.

12 shows a histogram graph in the case of 1) above.

Referring to FIG. 12, it can be seen that the concentration is 5000 us regardless of the number of agents 400 and AL that operate. The maximum jitter is around 10 us. That is, the real-time thread of the device control module 200 can confirm that the user starts and ends the given task exactly at the desired period.

FIG. 13 shows a histogram of the time taken for the joint reference to be output in the case of 2) above.

Referring to FIG. 13, it can be seen that a reference value is transmitted as a control signal between 30 us and 100 us after the real time thread of the device control module 200 starts. In addition, it can be seen that there is no tendency according to the number of operating agents 400 and AL. The amount of change is about 70us, which is about 1.5% of the time compared to the control period, about 7 times the measured jitter, and the period is 1 ms (1000 us) because it operates at 1 kHz for the experimental motor controller. Thus, the variation in the time it takes for the reference to be output is less than 10% of the motor control period of the device 100, which means that the system configuration of the present invention is a multi-rate system with different operating frequencies between the device and the control system. This means that it guarantees enough performance to run the Multi-Rate System.

Meanwhile, FIG. 14 shows a histogram of the time taken for synchronization between the device control module 200 and the threads of the agents 400 and AL to occur.

You can see that this time varies from a minimum of 10 us to a maximum of 70 us. Comparing with the above experimental results, it can be seen that the time taken for synchronization in case 3) is the largest among the time until the device control module 200 outputs the reference. This phenomenon occurs because the thread of the device control module 200 has the highest priority, and Sync. If the thread of the agent (400, AL) that captures the signal runs too fast, it consumes the CPU resources of the PC, so the trade-off is at the proper level because the performance of synchronization and the CPU resources to invest in the computation for motion generation are in conflict with each other. Indicates that is possible. As a result, it is possible to allocate enough CPU resources to the user while ensuring sufficient performance to operate the system.

As described above, according to an embodiment of the present disclosure, a plurality of agents having mutually independent processes and a shared memory in which references generated according to operations of the plurality of agents are stored are provided and provided to the hardware device using the references. By providing a separate device control module for controlling the, in a robot system requiring real-time, while several independent processes for the same hardware control can coexist, the operation of the robot can be stably controlled accordingly. In addition, even if each agent is independently developed, the reference memory can be synthesized and selected through the shared memory, thereby reducing the possibility of mutual collision and securing robust real-time. In addition, since agent replacement and real-time debugging are facilitated when an error occurs, collaboration convenience and scalability can be brought.

Meanwhile, the above-described methods according to the present invention may be stored in a computer-readable recording medium that is produced as a program for execution on a computer, and examples of the computer-readable recording medium include ROM, RAM, CD-ROM, Magnetic tapes, floppy disks, optical data storage devices, and the like, and also include those implemented in the form of carrier waves (eg, transmission over the Internet).

The computer readable recording medium can be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion. In addition, functional programs, codes, and code segments for implementing the method can be easily inferred by programmers in the art to which the present invention belongs.

In addition, although the preferred embodiment of the present invention has been shown and described above, the present invention is not limited to the specific embodiments described above, but the technical field to which the invention belongs without departing from the spirit of the invention claimed in the claims. Of course, various modifications can be made by those skilled in the art, and these modifications should not be individually understood from the technical spirit or the prospect of the present invention.

Claims

In the system control method,

Operating a plurality of agents using a shared memory and having mutually independent processes;

Obtaining hardware control data for controlling one or more devices from each reference generated from the plurality of agents and stored in the shared memory; And

Transmitting a control signal according to the reference to the one or more devices selected from the hardware control data.

How to control your system.
The method of claim 1,

Receiving sensor data from the one or more devices;

Updating the shared memory using the received sensor data; And

Outputting, by the plurality of agents, the reference updated according to the sensor data to the shared memory;

How to control your system.
The method of claim 1,

The shared memory includes a motion owner variable corresponding to the joint device,

The motion owner variable represents one agent having authority to the joint device among the plurality of agents,

Stopping an operation of the agent in which an error occurs when an agent in error exists among the plurality of agents; And

Changing the motion owner variable corresponding to the agent in which the error occurred to another agent;

How to control your system.
The method of claim 1,

Acquiring the hardware control data,

Generating hardware abstraction data according to each reference;

Generating a control signal corresponding to the hardware abstraction data by using a hardware database including hardware information about the one or more devices; And

Communicating the control signal to the one or more devices;

How to control your system.
The method of claim 1,

Operating, acquiring, delivering, receiving and updating are all processed within a first period corresponding to a real-time operating period of the system.

How to control your system.
In the system control unit,

A plurality of agents having mutually independent processes;

A shared memory storing a reference generated according to the operations of the plurality of agents; And

From each reference generated from the plurality of agents and stored in the shared memory, hardware control data for controlling one or more devices is obtained, and the control signal according to the reference is sent to the one or more devices selected from the hardware control data. Including a device control module for transmitting

System control unit.
The method of claim 6,

The device control module receives sensor data from the one or more devices, updates the shared memory using the received sensor data, and stores the reference in which the plurality of agents are updated according to the sensor data. Output to

System control unit.
The method of claim 6,

The shared memory includes a motion owner variable corresponding to the joint device,

The motion owner variable represents one agent having authority to the joint device among the plurality of agents,

If there is an agent having an error among the plurality of agents, the user system for stopping the operation of the error-prone agent, and further comprises a user system for changing the motion owner variable corresponding to the agent with the error to another agent doing

System control unit.
The method of claim 6,

Further comprising a hardware database containing hardware information for the one or more devices,

The device control module generates hardware abstraction data according to each reference, generates a control signal corresponding to the hardware abstraction data by using a hardware database including hardware information about the one or more devices, and controls the Transmit a signal to the one or more devices,

The hardware information includes at least one of a list of the device, a reduction ratio, an encoder pulse, a driver channel number, and communication protocol information.

System control unit.
In a real time control system,

One or more hardware devices including a joint or sensor device of the system;

A plurality of agents having processes independent of each other in association with the control system;

A shared memory storing a reference generated according to the operations of the plurality of agents;

From each reference generated from the plurality of agents and stored in the shared memory, hardware control data for controlling one or more devices is obtained, and the control signal according to the reference is sent to the one or more devices selected from the hardware control data. The state information of the current step is acquired from the controlling device, and the state information is transmitted to the plurality of agents processing a reference operation using the state information of the current step. A device control module which obtains reference information calculated in advance based on the information, and transmits the reference information to the device as a control signal corresponding to the current step; And

And a user system for managing real-time operating cycles and execution operations of the device control module and the agents.
In the system control method,

Obtaining state information of the current step from the controlling device;

Transferring the status information to one or more agents processing a reference operation using the status information of the current step;

Obtaining, from the agent, reference information calculated beforehand based on state information of a previous step; And

Transmitting the reference information to the device as a control signal corresponding to the current step.

Real time system control method.
The method of claim 11,

Delivering the state information,

Updating state information of the current step in a shared memory; And

If the status information is updated, transmitting a synchronization signal to the one or more agents,

The at least one agent acquires the state information from the shared memory when the synchronization signal is received and computes a reference for a next step

How to control your system.
The method of claim 11,

Acquiring the pre-calculated reference information,

Receiving, by the one or more agents, a response signal for updating and transmitting pre-calculated reference information to shared memory; And

Obtaining reference information of an agent from which the response signal is received from the shared memory;

How to control your system.
The method of claim 11,

Transmitting request information for sensor data for a next step to the controlling device;

How to control your system.
The method of claim 11,

The step period is uniquely synchronized with a thread of a device control module of the device system.

How to control your system.
In the system control unit,

A device communication unit obtaining state information of a current step from a controlling device and transferring the state information to one or more agents which process a reference operation using the state information of the current step; And

A reference obtaining unit which obtains, from the agent, reference information previously calculated based on state information of a previous step,

The device communication unit transfers the reference information to the device as a control signal corresponding to the current staff.

System control unit.
The method of claim 16,

The device communication unit updates the state information of the current step in the shared memory through the shared memory access unit,

If the status information is updated, further comprising a synchronization signal processor for transmitting a synchronization signal to the at least one agent

System control unit.
The method of claim 16,

When the synchronization signal is received, the one or more agents obtain the state information from the shared memory to calculate a reference for a next step,

The reference acquisition unit,

Receiving a response signal for the at least one agent updates and transmits the pre-calculated reference information to the shared memory, and obtaining the reference information of the agent received the response signal from the shared memory

System control unit.
The method of claim 16,

The device communication unit transmits request information for sensor data for a next step to the control target device.

System control unit.
The method of claim 16,

The step period is uniquely synchronized with a thread of a device control module in the device system.

System control unit.