Existing actuator control systems usually involve only one control parameter: current or voltage and utilize limited sensor information such as position or velocity.  This simple format results from the use of electrical motors to drive constant speed systems such as production machinery.  Despite the  vast array of new technology that now exists such as sensors, communication busses, control electronics, computer chips, etc., actuator design and actuator control systems have remained virtually the same until now.  For the past ten years, The Robotics Research Group at UT Austin has actively pursued the development of intelligent actuators. This includes the mechanical design and architecture of such systems and the performance based control of these actuators.  These actuators are fully programmable and can operate under demanding conditions and with reduced maintenance.

In the continuing effort to advance the technology of these actuators the Robotics Research Group (RRG) has created the software test bed.  This setup will play a vital part in the development of decision-making software that will ultimately control the intelligent actuator.  The critical components of the test bed are the control system, a personal computer, a motor, and real time and field programmable gate hardware.  The test bed will be used to help a number of students who are presently performing research in the areas of condition-based maintenance, performance mapping, sensor fusion, switch reluctance motor design, and control synthesis. 

Return to Top of Page

	Motion Control Hardware Motion control hardware has been provided to the RRG by National Instruments of Austin, Texas.  The hardware comes with the software developer environment LabView.  This virtual instrument is powerful in integrating hardware components such as sensors to output voltage or current output signals while performing control functions in between.  Students will be able to create user-defined test setups, add unlimited control schemes, record test data, and perform high-level computations by using this product.
	Real-Time Operating Most computer-based control applications run on general-purpose operating systems like Microsoft Windows.  Nevertheless for more robust systems there is a need to for deterministic processing that non-real-time operating systems like Windows do not provide. National Instruments created LabVIEW Real-Time (RT) to address the need for deterministic real-time performance using LabVIEW.  LabVIEW RT couples the user-friendly benefits and functionality of LabVIEW with the power of real-time high update rates so students will be to generate deterministic applications using graphical programming.
	FPGA A field programmable gate array (FPGA) is a large collection of cells that contain configurable logic and memory elements.  Cells can be connected to each other using a large number of programmable switches in a variety of ways to fit a variety of applications.  FPGA will be used in the software test bed since hardware will frequently be reconfigured to satisfy various control arrangements.  Additionally, FPGAs lend themselves to the implementation of custom algorithms in hardware, they offer precise timing and synchronization, rapid decision making, and simultaneous execution of parallel tasks.

Return to Top of Page

Return to Top of Page

The project is currently in phase III and uses LabView FPGA to control a switch-reluctance motor.  The ultimate goal is to develop a test bed that will accommodate various types of device(s) under test (DUT).  In this case DUTs will be motors including the switch reluctance,  DC Brush, and DC Brushless.  The test bed will lend itself to quick changeover among the various types of motors through its versatile hardware and software.  Ultimately the software will feature a decision making system capable of controlling the DUT based on sensor feedback, CBM, performance maps, and user-defined criteria (e.g. the user may set priorities among parameters such as speed or torque).  In the previous phases the following has been accomplished:

Return to Top of Page

In the first phase of this research project it was necessary to learn about the hardware, determine the types of signals, and debug the system.  This was accomplished by implementing a PID controller in Labview, using the PID Control Toolkit. The parameters of the PID controller are proportional gain, integral time, derivative time, output range and cycle time.  In order to assess the controller performance two “.vi” files were created.  The first file was built determine the system response by analyzing the overshoot, rise time, and steady-state position error.  The second file showed a Bode diagram of the system to measure the gain in amplitude and the phase shift in the system.

In the second phase functions were implemented in Labview including a quadrature counter to read encoder data, generation of the PWM signal dictated by the desired PWM frequency and duty cycle, and a PID loop to control speed.  For the quadrature counter application, two signals (A and B) are read through digital channels with both rising and falling edges detected.  Possible counter sizes using the FPGA module in Labview are 8-bits, 16-bits or 32-bits.  In order to improve accuracy and guarantee perfect timing between the host computer and the FPGA software, a synchronization signal was also added.

Below is the block diagram of the vi that generates the PWM output to the OSMC. The digital output channel alternates between high and low values for time periods as specified.  The desired direction of movement determines the signals sent to the controller.

 In the current phase a switched reluctance motor is being controlled by the LabView software.  Basically this can occur through a technique which uses Hall sensor or an encoder to determine the position of the stator pole in relation to the rotor.  This becomes critical since switching on and off of the four phases at the appropriate time is necessary to cause rotation of the motor.  The method used here involved four Hall-sensors (one per phase) and a corresponding truth table, provided by Motorsoft, indicating which phase triggering corresponded to the various combinations of Hall-effect sensors.  Ultimately “vi.” files where constructed according to the sensor-logic for each phase.  In addition, a program that provided the PWM was developed and interacts with the aforementioned logic.   

Return to Top of Page

The hardware used in this project has changed over a period of time to integrate new technology into the Software test bed setup.  During phase I the test bed components consisted namely of parts previously used for the on a test bed previously used to measure actuator endurance and reliability.  The plan here was to operate the test bed in a PID control loop using a data acquisition board in a PC.  The purpose in this was to familiarize ourselves with the hardware, learn what type of signals flowed from each component, and to detect and remedy any component problems prior to the receipt of the National Instrument RT/FPGA controller. 

For this task the controlled component was ½ hp brush DC motor.  A Heidenhain encoder gathered analog data to determine the motor position.  To convert the analog signal, an interpolation and digitizing electronic unit was used between the encoder and a Vibrac sensing system.  The sensing system processes the square wave signal and outputs the shaft position to a National Instruments DAQ board.  The PC then executed a PID loop and sends a voltage signal to an AMC amplifier, which in turn delivers the appropriate amount of voltage to control the motor.  The motor and sensing device are mounted to a stiff frame and is supported by a steel table which resists vibration.  The motor is mounted to an adaptable fixture that can accommodate various types and sizes of motors.  

In the second phase of this research the AMC controller used for the single-phase brush DC motor was replaced with a Open Source Motor Control (OSMC) controller that does not perform commutation.  Alternatively LabView FPGA was introduced to the test bed which provided the means to generate the PWM speed commands.  As such, this project involved the integration of new hardware and software into the test bed.  In addition to the new controller, other components needed to accommodate these changes were a DC power supply, optoisolators to separate the logical components from the high power components and a Computer Optical Products high resolution encoder.

In the third phase four amplifiers are being used to trigger each phase of the switch reluctance motor.  The software is written in LabView and is used to trigger the phases on and off in accordance to the hall-sensor data.  The test bed fixturing was also modified to accommodate the mounting plate on the SRM.  Currently the motor is being run using a low-level control technique.  In future phase the motor will be controlled at a higher level (e.g. forward and reverse, higher speeds that require timed triggering).  Other types of sensors will also be integrated into the set-up including Magne magnetic brake and a Vibrac torque sensor. 

Motor Specifications

Motorsoft Switch Reluctance Motor

• Part Number:  RA165157

• Output 1.13Kw
  6000 rpm

• 0.73Kw
  6000 rpm

• 1.8 N-m

• 4 Phases

• 24 v

KollMorgern Servo Disc DC Motor (presently offline)

• Part Number: JR12M4CH

• 0.53 hp

• 60.8 v

• 8.42 A

• 3000 RPM Max

• Constant Duty

Return to Top of Page

For more information, please contact  Pradeep Ashok, Ganesh Krishnamoorthy, or  John Hall

Return to Top of Page