Computational Intelligence - August 2012 - 19

processing." Moreover, the interval output
It was found that a type-2 FLC is more suitable for
of the type-2 fuzzy system gives very useful
information to machine tool controllers in
Linear Microstepping Motor control under parameter
order to maximize material removal while
variation and external disturbances.
controlling tool wear or tool failure to
maintain part quality specifications."
The resulting system was tested at the
E. Type-2 FLSs for the Control of Modular
Micro-machining Laboratory at the Mondragón University
and Reconfigurable Robots
in Spain, where the actual setup is shown in Fig. 5.
In [15], a design methodology was presented for interval typeD. Type-2 TSK FLS for Neuro-Fuzzy
2 TSK FLSs for Modular and Reconfigurable Robots (MRRs)
Control of Linear Microstepping Motor Drives
followed by experiments evaluating the proposed type-2 conIn [14], a supervisory Takagi-Sugeno-Kang (TSK) neural
troller for a two-degree-of-freedom robot (shown in Fig. 7).
fuzzy network control system was proposed and demonMRRs allow the flexible assembly of modular components
strated for the control of Linear MicroStepping Motors
including physical components such as joints, electrical hard(LMSMs). As discussed in [14]-LMSMs have the merits of
ware as well as control software for rapid and flexible deploypermanent magnet synchronous machines and are widely
ment in a variety of commercial and industrial settings. The
used in robotic manipulators, industrial process machinery,
authors demonstrate that their system provides guaranteed staand machine tools where they must perform high-precision,
bility (system is Uniformly Ultimately Bound (UUB), see [15])
high-speed positioning operations. However, the high speed
while exhibiting the additional advantage of requiring minimal
of operation in combination with their construction can
tuning effort. The devised controller was compared and shown
result in high levels of vibration which needs to be remeto considerably outperform a well-known linear controller in
died by a control system that is able to deal with a signifiterms of tracking performance. In addition, the authors also
cant number of different sources of
uncertainties. The control system proposed in [14] was designed to handle
uncertainties in LMSMs, including sensor noise, payload variation and external disturbances.
It was shown through simulation
Tool
Tool
and real world experiments (the configuration and experiment setup are
Workpiece
shown in Fig. 6) that the interval type3
2
2 fuzzy system is less sensitive to dis1
Dynamometer
AE
Sensor
turbances when compared to type-1
Tool Path
fuzzy logic based systems. Hence it is
more suitable for LMSM control
(a)
(b)
under parameter variation and external
FIGURE 5 The high-precision milling setup at Mondragón University (Spain) [13].
disturbances [14].

DSP Servo Control Board
Pentium IV
PC

LMSM
Va

TMS320
VC33

SVPWM
Unit

Real-Time
Interface

A /D
Converter

Memory

Encoder
Interface

IGBT
Inverter

Vb
ia
ib

Current
Sensing Circuit

Comparator
Circuit

Linear
Encoder

FIGURE 6 Configuration of the experimental system [14].

AUGUST 2012 | IEEE COMPUTATIONAL INTELLIGENCE MAGAZINE

Table of Contents for the Digital Edition of Computational Intelligence - August 2012