Data processing: measuring – calibrating – or testing – Measurement system – Speed
Reexamination Certificate
1999-04-27
2004-03-09
Hoff, Marc S. (Department: 2857)
Data processing: measuring, calibrating, or testing
Measurement system
Speed
C702S145000, C324S165000
Reexamination Certificate
active
06704683
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to velocity determination of a moving object based on sensed position.
Measurements of velocity for electromechanical systems may be obtained in a number of ways, including the use of dedicated sensors such as tachometers and the taking of derivatives of signals from position sensors such as potentiometers and optical encoders. Hardware and software approaches can calculate these derivatives. For example, a simple operational amplifier circuit can differentiate an analog potentiometer signal to output a velocity signal, or the same potentiometer signal may be differentiated in software. Both approaches introduce lag into the velocity estimate. Filtering in hardware or software to cope with the derivative's sensitivity to noisy sensor data further worsens the lag, sometimes to the point that the velocity data becomes useless.
Due to their optical, non-contact design, encoders have a long service life, low friction, excellent repeatability, and inherently digital outputs with low noise. These attributes have made them popular position sensors. In applications requiring velocity as well as position information, designers often would like to obtain velocity information from an optical encoder rather than accept the added expense of an analog velocity sensor and its associated analog-to-digital conversion electronics. Software techniques for doing so have existed for some time. Two possible approaches include measuring the change in position (number of encoder pulses) over a fixed period of time, and measuring the period of time required to attain a fixed position change (the period of one encoder pulse, for example).
The first technique, pulse counting, can be useful at higher encoder speeds where pulses occur frequently, leading to acceptable quantization error and dynamic range in the velocity measurement. For slower speeds seen in electromechanical applications such as haptic feedback or “force feedback” devices, this technique becomes problematic. Consider the extreme case where the encoder moves so slowly that only one pulse count occurs in the given time period. If speed gradually increases until two pulse counts are measured, estimated velocity will suddenly double. The quantization error in this case is almost 100%. The application of smoothing algorithms may help (or simply create lag-induced instability in some systems), but cannot eliminate this fundamental problem.
The second technique, period counting, appears more attractive for measurements of slower velocities. With the ability to measure an infinitely long period, the slowest velocity may be measured. The resolution of the time measurement will determine the fastest measurable velocity. The equation for period counting:
ω
=
2
π
T
=
2
π
period
where &ohgr; is the angular speed of a rotating shaft. This technique has some implementation difficulties. Accurate period measurement requires an interrupt-driven time counter. Measuring time in this fashion with frequent encoder pulses and one or more sensing channels can severely tax a computer that must simultaneously handle other tasks (graphics, kinematics, modeling, etc.). It can also be costly on an embedded processor of more humble means. In addition to frequent interrupts, it requires an inversion to convert from period to velocity—this operation can be time-intensive or not available on low-end microcontrollers.
Principles of Optical Encoder Operation
Optical encoders use paired light sensors and light sources with mechanical interruptions to measure rotary or linear position.
FIG. 1
is a diagram of an incremental rotary optical encoder
10
attached to a motor shaft
12
of a motor
14
to measure the angle of rotation of the shaft. A light source
16
emits a beam of electromagnetic energy toward a light sensor or detector
18
through an encoder disk
20
which is transparent or includes open slots. An encoder hub
22
couples the disk
20
to the shaft
12
. See R. D. Klafter, T. A. Chmielewski, and M. Negin, “Robotic Engineering: An Integrated Approach,” Prentice-Hall, Englewood Cliffs, N.J., 1989, incorporated herein by reference.
FIG. 2
shows the face of the optical encoder disk
20
, with a striped pattern
24
that provides periodic interruptions between the light source and light sensor as the motor shaft rotates.
The encoder
10
of
FIG. 1
preferably uses two emitter-detector pairs (both emitters included in the light source
16
and both detectors included in the light sensor
18
). If only one pair is used, the single output signal indicates motion, but cannot indicate which direction the encoder disk/shaft is turning. A practical encoder requires the addition of a second emitter-detector pair slightly offset from the first so that it produces a square wave pulse stream that is 90 (electrical) degrees out of phase with the first pulse stream.
FIG. 3
a
is a diagrammatic illustration of the two signals
26
(channel A) and
28
(channel B) output by two detectors included in the light sensor
18
during clockwise rotation of the encoder disk
20
.
FIG. 3
b
is a similar illustration of the two signals
26
and
28
output during counterclockwise rotation of the encoder disk. Thus, the relative phases of the two signals provides an indication of the direction of rotation. The logic level outputs of the A and B channels of the incremental encoder are that A leads B when clockwise rotation occurs, and A lags B when counterclockwise rotation occurs.
FIG. 4
is a block diagram illustrating the components of a signal processing system
30
for an incremental optical encoder signal. Incremental optical encoders count pulses from the emitter-detector pairs in light source
16
and light sensor
18
. A quadrature decoder state machine
32
receives the two channels from the detector pair and outputs a direction signal and a count strobe to a counter
34
. Rotation of the disk
20
in one direction causes the counter to count up, and rotation in the other direction causes the counter to count down. Count and least significant bits are provided to a position register
36
which provides a position signal to a component that can process it, such as a controller or microprocessor. Because this arrangement can only measure a relative change from a starting position, if the application requires knowledge of absolute position the system must be powered up (or the counter reset) with the shaft in a known position. Inherently absolute optical encoders that use several pairs of light sensors and sources are also available. An absolute encoder has a more complicated Grey-code-based codewheel (or strip, in the case of a linear sensor) to interrupt the light beams. Though the encoder velocity techniques presented in this work can easily be applied to absolute encoders, a detailed description of their function is not presented. For further detail, see Klafter et al.
Encoders are useful in many applications. More specifically, computer interface devices such as joysticks, mice, track balls, steering wheels, etc. make use of encoders to determine the position of a user manipulatable object (manipulandum) in a workspace of the user object, and provide the position information to a host computer that is connected to the interface device. An encoder can be used to sense position of the manipulandum in one or more degrees of freedom. Force feedback interface devices are a form of interface device in which motion of the manipulandum is sensed and forces are output on the manipulandum using actuators such as motors. In some preferred force feedback interface device implementations, a local microprocessor, separate from the host computer, is included in the interface device to offload computational burden from the host computer. The local microprocessor can handle force computations and position processing, allowing the host to concentrate on running a host application and displaying environments on a display screen in part based on position information from the interface device.
The digita
Charioui Mohamed
Hoff Marc S.
Immersion Corporation
LandOfFree
Direct velocity estimation for encoders using nonlinear... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Direct velocity estimation for encoders using nonlinear..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Direct velocity estimation for encoders using nonlinear... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3253768