Institute for robotics and process control

 

Institute for
Robotics and Process Control

Open PC-Based Robot Control Systems

Project Description


Since the beginning of the 90s, the Institute for Robotics and Process Control deals with the development of robot control systems. Common industrial ones only provide inflexible interfaces, which hamper or inhibit possibilities of multi-sensor integration. For instance, Cartesian force/torque control requires low control cycle times - not only for appropriate control behavior but for short response times on control variable leaps (e.g. caused by sudden hard environmental contact). A transputer-based system was a first open control system approach, which was applied to work on manutec manipulators. The latest version is a PC-based system, which focuses on a modular software architecture ( MiRPA, hybrid robot control architectures). Based on these experiences, new architectures for the Stäubli RX-series and the TX-series are provided. Here, the excellent mechanical properties are combined with an open and flexible control architecture to create a base for venerable research results.
The picture below shows our open robot control system based on the Stäubli TX series with an CS8c controller employing a low-level interface (LLI) developed by Stäubli.

txcontrol.jpg
Open robot control architecture based on the Stäubli TX-series featuring CS8c controllers and the low-level interface (LLI).


Since the advent of affordable multi-core processors, our entire robot control system including demanding sensor data acquisition and processing can be incorporated into one powerful PC running the real-time operating system QNX. An architecture sketch of our system is shown in the figure below. In contrast to our open control architecture for the RX-series, no modifications of the controller hardware are necessary. This feature enables using our control architecture with virtually any TX-series manipulator providing a low-level interface - merely a USB flash drive containing the software for the CS8c controller and a PC equipped with the RTOS QNX are necessary.

CS8architecture.jpg
Open robot control architecture consisting of the CS8c controller, a QNX PC, and a USB flash drive.


In the following paragraphs our open robot control architecture for the Stäubli RX-series is described. An RX60 manipulator executing a sensor guided assembly task is shown on the picture below. Its control cabinet stands in the background. The human machine interface (HMI) is mounted with a pivot arm to the control cabinet and provides excellent possibilities for manual control and teach-in programming. Besides the most important operational controls a touch screen panel and a Space Mouse is available.

p3235032.jpg
Example setup for a force-guided assembly task.




terminal.jpg
Control terminal with Space Mouse.


However, this work does not focus on a convenient HMI, but moreover on open interfaces, extensibility, scalability, robustness for external disturbances, and especially on the requirement to keep the usage of sensors entirely open. The above-mentioned possibilities for manual control a teach-in programming are only the necessary base.

Unless parts of the original control system match the new requirements, the parts are reused in the new system. Essentially, the power electronics, i.e. the frequency inverters and their accessories, have to be mentioned here. For manipulator operation, at least one control PC is required. A motion control board within this PC constitutes the interface to an electronic adaptation assembly, which delivers velocity set points to each frequency inverter on the one hand side, and which provides the joint positions to the control computer. In addition, I/O ports for further periphery (brakes, inverter control, valves, etc.) are provided. One major requirement is to make the absolute manipulator position permanently available to the controller. Therefore, an electronic assembly group for position tracking was developed. This one transmits the current position with an accuracy of 13 Bit via the serial port to the control PC. To guarantee the board's functionality at AC power failures and during transports, the position-tracking unit supplied by accumulators if required. The communication vie the serial port occurs only at during the start-up procedure, afterwards, die control computer receives the position information from the motion control board.

To assure an excellent dynamic behaviour, signal quality loss and signal latency have to be kept small, such that the necessary bandwidth of the components for signal adaptation between PC and drive electronic was determined preliminary to the development. Interfaces to the control computer are such open, that the user can freely decide how to design his software architecture. The signals for basic manipulator control are only the joint velocity set points for the drive controllers and the position inputs.

controlcabinet.jpg
Robot control cabinet containing three PCs and the power electronics


The proposed architecture uses a joint position controller, which is provided as interface to higher software layers. Besides the possibility to run all processes (joint position control, hybrid control, trajectory planning, user application, etc.) on one machine, is it useful to swap further processes out to other PCs and to take a real time communication system between them.

To follow the introductory mentioned aim of multi-sensor integration, the following figure depicts a suggestion for a possible architecture:

abb1.gif
Exemplary control architecture with three PCs.


Further information may be found in our Publications. In case of any further questions please contact Daniel Kubus.

It took 0.89s to generate this page.

### 机器人学、力学和控制的入门教程或书籍 对于希望进入机器人学、力学以及控制系统领域的人来说,可以从以下几个方面入手: #### 基础理论 机器人学涉及多个学科的知识融合,包括机械设计、传感器技术、嵌入式系统开发等。Player Project 提供了一系列开源工具用于支持机器人应用中的硬件接口与软件框架构建[^1]。 在电路基础知识方面,掌握基本元件的功能及其相互作用规律是非常重要的前提条件之一,这有助于理解如何搭建实际运行环境下的电子线路板(PCB)[^2]。 关于计算机视觉资源,则可以访问由爱丁堡大学提供的在线图书馆链接地址(http://homepages.inf.ed.ac.uk/rbf/CVonline/books.htm),其中包含了大量经典教材PDF版本下载服务;特别是Andrew Zisserman所著的作品,在模式识别等领域具有深远影响价值[^3]。 以下是几本推荐给初学者阅读的经典著作列表: 1. **Introduction to Robotics: Mechanics and Control** - 这本书全面介绍了机器人的运动学、动力学及控制原理等内容,适合刚接触该领域的读者作为入门指南。 2. **Modern Robotics: Mechanics, Planning, and Control** - 它不仅涵盖了传统意义上的机械臂操作分析方法论,还深入探讨了现代移动平台规划策略等方面的新进展趋势。 3. **Robot Modeling and Control** - 主要聚焦于建模过程中的数学描述方式,并结合具体实例讲解不同类型的控制器设计方案。 4. **Control Systems Engineering** - 针对自动化的反馈机制进行了详尽阐述,能够帮助学生建立起扎实可靠的工程实践能力基础之上进一步探索高级话题。 通过上述资料的学习积累,相信会对整个行业的概貌形成较为清晰的认识并激发更深层次的兴趣追求下去! ```python # 示例代码展示简单的PID控制器实现 class PIDController: def __init__(self, kp=0.0, ki=0.0, kd=0.0): self.kp = kp self.ki = ki self.kd = kd self.previous_error = 0.0 self.integral = 0.0 def update(self, error, dt): proportional_term = self.kp * error integral_term = self.integral += error * dt derivative_term = self.kd * ((error - self.previous_error) / dt) output = proportional_term + integral_term + derivative_term self.previous_error = error return output ```
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值