Modeling and Canceling Tremor in Human-Machine Interfaces
需积分: 0 116 浏览量
2024-04-05
21:40:38
上传
评论
收藏 1.41MB PDF 举报
Modeling and Canceling Tremor
in Human-Machine Interfaces
ero-phase modeling and canceling of
Z
tremor can improve precision in hu-
man-machine control applications. Past
methods of tremor suppression have been
hindered by feedback delays due to phase
lag and by the inability to track tremor
frequency over time. The weighted-fre-
quency Fourier linear combiner (WFLC)
is an adaptive noise canceller that pre-
cisely models tremor with zero phase lag.
This article briefly presents the WFLC
algorithm and describes its application to
computer input filtering, clinical tremor
quantification, and active tremor cancel-
ing for microsurgery.
Overview
Tremor is any involuntary, roughly
sinusoidal movement
[
11. Pathological
tremor is caused either by brain injury or
diseases such as essential tremor, Parkin-
son’s disease, and multiple sclerosis [l].
It is a persistent nuisance in mild cases,
and in severe cases it can be completely
debilitating [2]. Physiological tremor ex-
ists in all human motion [3]. Studies have
shown that tremor corrupts voluntary mo-
tion with an additive oscillatory distur-
bance, or noise [2]. Therefore, during
voluntary motion, tremor causes an
un-
wanted oscillatory disturbance, or noise.
An example of tremor disability is pre-
sented in Fig. 1, in which a subject with
cerebellar tremor due to brain trauma is
unsuccessful in attempting to draw an
AI-
chimedes spiral on a digitizing tablet.
The importance of tremor suppression
during computer input is increasing as
input interfaces that sense human motion
become more widespread. Computer
mice, joysticks, trac-balls, and digitizing
tablets are used to control applications
from word processing to computer-aided
design (CAD). Palmtop computers such
as the Newton (Apple Computer, Cu-
pertino, CA) use pen input. Sensory
gloves are used as input to virtual environ-
ments, which are being investigated for
various design and analysis functions, as
well as for educational and entertainment
uses
[4]. Use of these devices affects em-
ployability. For example, a vocational
CAD training program for disabled clients
Moy/June
1996
at Maryland Rehabilitation Center is cur-
rently inaccessible to persons with severe
pathological tremor, as they lack adequate
control over the interfaces used. For these
persons, effective tremor suppression dur-
ing computer input would enhance quality
of life, personal independence, and work-
place competitiveness
[5].
Pathological
tremor
is most commonly treated with
medication
[I],
but this fails to suppress
many tremors adequately-as many as
50% by some estimates [6]-necessitat-
ing development of assistive computer in-
terfaces. Past approaches to assistive
interfaces include viscous damping of a
joystick
[7]
or mouse (The MouseTRAp,
Michaelis Engineering, Southampton,
UK)
and lowpass filtering of input signals
[2]. These techniques involve phase lag,
which causes a time delay in the user’s
visual perception of motion via the com-
puter monitor. Studies have shown that
Cameron
N.
Riviere
The Robotics Institute, Cornegie Melon University
Nitish
V.
Thakor
Deportment
of
Biomedical Engineering,
Johns Hopkins School
of
Medicine
IEEE ENGINEERING IN MEDICINE AND BIOLOGY
feedback delays result in decreased mo-
tion accuracy and degraded handwriting
legibility [SI. Clearly, tremor canceling
without time delay is needed during com-
puter input.
Quantification of tremor is of interest
as an aid to clinical diagnosis and objec-
tive evaluation of treatment
[9].
Tremor is
typically examined via handwriting and
drawing specimens, often recorded using
digitizing tablets
[
lo].
Fast-Fourier trans-
form (FFT)-based spectral analysis is the
most popular method of tremor quantifi-
cation [1,3]. The FFT models the input
signal as a stationary periodic signal, i.e.,
one whose statistical characteristics do not
change with time. Yet tremor amplitude
and frequency are time-varying
[
11, often
making power spectra difficult to interpret
[ll].
A method to track modulation in
tremor frequency and amplitude over time
would allow a better clinical under-
standing of individual cases of tremor.
Physiological hand tremor causes im-
precision
in
microsurgery
[
121. Despite
the development of teleoperated surgical
systems that suppress tremor, direct, i.e.,
non-teleoperated, microsurgery retains its
appeal for the surgeon, due to its more
natural feel
[
131 and lower equipment cost
[
141.
Effective tremor suppression in
handheld microsurgical instruments
would result in greater precision, leading
to better outcomes and lower costs. The
ideal tremor-canceling system for direct
microsurgery would have a natural, unob-
trusive feel. One promising method is an
active noise control approach. Rather than
suppressing the actual hand tremor, the
surgical instrument tip is actuated with an
equal but opposite motion, effectively
subtracting the tremor out of the tip mo-
tion. Since this compensating signal can
only be effective in the absence of time
delay, a zero-phase tremor estimation sys-
tem is needed.
To model and cancel tremor during
purposeful movement, we have developed
the WFLC
[15].
This algorithm is well
suited
in
several ways to the applications
listed above. It is based
on a dynamic
truncated Fourier series model of tremor,
0739-51 75/96/$5.0001996
29
资源评论
wd1879804
- 粉丝: 0
- 资源: 15
最新资源
- 基于matlab实现本份代码能对图像进行gabor滤波处理，结合指纹方向图以及指纹沟壑频率特性，对指纹图像进行增强.rar
- 基于matlab实现RBM神经网络实现了手写数字体识别的GUI程序.rar
- 基于matlab实现蝙蝠算法优化相关向量机建模对数据进行建模和预测.rar
- 基于matlab实现编写的禁忌搜索算法，解决了TSP问题，对初学者有重要的参考价值.rar
- 基于matlab实现SOH关于IMU进行姿态求解的C代码，里面包含了两套代码，分别是重力约束法求解和梯度下降法求解.rar
- 1_简单电子邮件客户端.zip
- 基于matlab实现powell优化搜索算法，适合于多参数优化且目标函数中不包含参数的情况.rar
- 基于matlab实现PID神经网络前向和反向学习算法的matlab的原程序代码.rar
- 基于matlab实现nsga-2的多目标优化算法，有注解.rar
- AIR-AP1815-K9-ME-8-5-182-0.zipFor 1815 1830 1840 1850 2700 3700
资源上传下载、课程学习等过程中有任何疑问或建议，欢迎提出宝贵意见哦~我们会及时处理！
点击此处反馈
