伟军 张 - 上海交通大学机器人研究所
志成 万 - 上海交通大学机器人研究所
俊 陶 - 上海交通大学机器人研究所
Jianzheng Zhang - Shanghai Jiao Tong University, School of Mechanical Engineering
The parallel robot is a popular topic among researchers and in applied industries, with advantages such as high durability, high load-carrying capacity, less error, high precision, low deadweight duty ratio, and good dynamic performance. Parallel robots are used on airplane simulators, during micro-operation, in surgery, and on large radio telescopes. However, most of those robots require a long development cycle, have complicated system maintenance and upgrades, and are expensive. Further development and use of parallel robots is also encumbered by limits to the construction of a stable, fast, and accurate open numerical control system. The key to overcoming this challenge is in using a real-time, multi-axis motion controller that offers precision, lower cost, a reduced development cycle, and easy maintenance. In addition, a highly functional software development platform with an intuitive user interface is critical to the product’s marketability in the industry.
By choosing PXI and LabVIEW as our development platform, we rapidly developed a six-DOF parallel robot control system. Using the NI PXI-7346 multiaxis motion controller enabled functionality such as multimotor synchronization, multiaxis-coordinated trajectory control, real-time trajectory curve display and selection, dynamic loading, and reconstructing of panels and datasharing. With the selected products, we created parallel robot control systems in less time for less money and acquired ideal results. For example, the 25 KHz to 25.6 MHz filter range for feedback signals generated by the encoder ensured stability, and real-time needs were more easily met using a 250 µs running cycle of six-axis proportional integral derivative (PID) control system compared to the traditional 1 ms. When six-axis coordinated motion finishes, steady state error of the end-effecter is less than 1 µm.
General Control System Design
The parallel robot drivers consist of six high-precision servo motors, each axis of which has a forward limit, a reverse limit, and home with 18 I/O. To ensure the end effecter arrives at the predetermined position along a fixed trajectory and works along that trajectory, we needed a powerful computer for the parallel robot trajectory planning and the inverse solution to compute and store data. Furthermore, data must be transmitted to the controller and drive in real time to generate current or voltage to drive motors. The PXI platform is ideal for handling the large data transfer, synchronization, and various I/O signals required, and offers advantages in platform bus bandwidth latency and modularized I/O from DC to 6.6 GHz RF. We selected the high-performance 8-slot chassis with the NI PXI-8186 embedded controller with 2.2 GHz Intel Pentium 4 CPU to reach high-speed trajectory planning, inverse solution, and data analysis. The NI PXI-6511 industry-level digital input interface module, as a peripheral module, provides up to 64-channel isolated digital input.
Hardware Design for the Control System
As a subsystem for chromosome incision devices, this parallel robot requires high positioning accuracy and a large work space. The basic mechanism is a six-PPPS decoupled mechanism with six DOF in space with six high-precision servo motors accomplishing six-dimension motion (displacement in x, y, z direction and rotation about x, y, z axis). The NI PXI-7356 multiaxis motion control module achieves microaccuracy requirements for the end platforms and coordinated controls for six motors. The card’s buffed breakpoint technology improves integral speed substantially: for common breakpoints, the max buffered trigger rate is 2 KHz, and the periodic rate can reach 4 MHz. The two-axis PID update rate can reach 62.5 to 250 µs for eight-axis, which is faster than 1 ms under ordinary systems. Such high computing efficiency satisfies this system’s requirements for quick response. Synchronization accuracy of PXI-7356 is less than one sampling period, and the position accuracy of encoder feedback is less than one quadrature count. The internal 8-channel 16-bit AI greatly improves analog-to-digital conversion resolution to ensure position accuracy is no more than one least-significant bit (LSB).
High-precision NI products meet the needs of this system, and the PXI-7356 multiaxis motion controller’s safety standard, s-curve adjustment, dual PID control loops, and electronic gear ensure system stability. The PXI-7356 multiaxis motion controller with the paired NI UMI-7774 motion interfaces uses a 64-channel digital I/O, which can control solid-state relay and read data from digital encoders/decoders. This makes signal read/write more convenient, including an 18-channel limit, a 12-channel enable signal, and the number of alarm signals.
Control System Software
Due to the complexity of the control system, we divided the software into an application layer, a core layer, and a drive layer. Each is divided into several functional modules based on their requirements, illustrated in Figure 2.
Operating the robot requires switches to control motors and brakes, interfaces to change motor parameters or piezoceramics, indicator lights for normal or alarm signals, and trajectory curves to monitor operation in real time, as well as seamlessly switching between user interfaces. Using LabVIEW we developed a user-friendly, convenient, and flexible human-machine interface.
The system uses a master/slave design to solve problems in data sharing between cycles which loop in different frequencies. Inverse solution models and control algorithms are embedded in the system. We implemented information interchanges between different modules using global, local, and shared variables. Switching between user interfaces involves the user event, notifier, queue technology, and synchronization technology, which avoids conflicts when two threads visit the same object at the same time. Adopting dynamic VI load technology saves memory and simplifies the framework and the front panel.
Core Layer: The procedure set is focused on robot trajectory control and I/O logic control and includes returning to home, continuous running, single-axis adjustment, trajectory curve selection, system self-test, and more. On one hand, this layer is responsible for implementing accurately synchronized motion control for every joint drive motor and trajectory control for end-effecter in operating space. On the other hand, to satisfy complex requirements for control tasks, it must also accomplish a set of I/O controls for motion and coordinated control for peripheral devices.
Drive Layer: We developed drivers as a set of functions to implement single-axis and multiaxis motion control, digital-to-analog conversion (DAC), and hardware I/O control. These include operational functions such as axis configuration, motion typesetting, and motor run and stop. This layer is responsible for setting motion axis parameters, motor acceleration and deceleration controls, run/stop control, DAC, and motion I/O settings and controls. Developing this system in LabVIEW provided efficiency gains in a fully functional, user-friendly system.
Overall System Performance and Testing
By implementing a new design for a parallel robot control system, we shortened the development cycle, reduced costs, improved system performance, and increased system functionality. Compared to earlier robot control systems, this solution is also easier to upgrade and maintain, resulting in higher cost performance. Advantages in system performance include stability, quickness, and accuracy. The 25 KHz to 25.6 MHz filter range for feedback signal generated by the encoder ensures stability. The real-time requirement is more easily achieved through the 250 µs control cycle of the 6-axis PID control system, much faster than the traditional 1 ms models. When 6-axis coordinated motion is complete, steady state error of the end effecter is less than 1 μm.
Basing the six-DOF parallel robot control system on the LabVIEW and PXI development platform allowed us to achieve advantages such as high reliability, high accuracy, high computing speed, high intelligence, and friendly HMI.
By using LabVIEW and the LabVIEW Real-Time Module, we established a module-based, open numerical motion control system with superior human-machine interaction, intuitive framework, and complete functionality. The solution also reduced development cycle and cost, and provides a good foundation for general application of parallel robot technology.
For more information on this case study, contact:
Shanghai Jiao Tong University, School of Mechanical Engineering