Académique Documents
Professionnel Documents
Culture Documents
ABSTRACT
The snag of common linear filtered filtered-∈ Adaptive (𝑘)], 𝑓(0,0) = 0
𝑥(𝑘 + 1) = 𝑓[𝑥(𝑘), 𝑢(𝑘
(1)
Inverse Control (AIC) method is modify to hold with 𝑦(𝑘) = ℎ[𝑥((𝑘)]
the characteristic of non-linear
linear object with time delay
and corresponding filtered ∈-adaptive
adaptive Algorithm Where 𝑓( . ) is a non-linear
linear function, 𝑢(𝑘) and 𝑦(𝑘)
based on Real-Time
Time Recurrent Learning (RTRL) is respect output and input variables, respectively. The
presented to identify the parameters and ddesign the principles of inverse control can be extended to deal
controller. The simulation result on a non
non-linear ship with non-linear
linear system, through non-linear
non system has
model of “The R.O.V. Zeefakkel” and Adaptive PID no strict inverse model. Whereas a linear system
control keeps the same dynamic response possesses a unique inverse, non-linear
non systems have
performance, and also mathematical model for ship only local inverses, valid in a bounded region of the
Course keeping be discussed. A neural network signal space. The non-linear
linear system is supposed to has
adaptive
ive filter is introduced for the removal of Bounded-Input
Input Bounded Output(BIBO) stability, i.e.,
impulse noise in digital images. there are existing constants 𝐶 and 𝐶 , and for the
input 𝑢(𝑘) ∈ 𝑠 , 𝑠 = {𝑢: ∣𝑢(𝑘 𝑘) ∣≤ 𝐶 , ∀𝑘 ≥ 𝑘 }, the
Keyword: Filtered-∈ Adaptive Inverse Control, Ship system output 𝑦(𝑘) satisfies: ∣y(k)∣≤ 𝐶 . If its
course keeping, impulse noise detection and removal. linearized system is observable at balance point
INTRODUCTION 𝑥 = 0, the non-linear
linear system described by (1) has
Adaptive Inverse Control (AIC) is novel approaches local NARMA (Non-linearlinear Auto-Regressive
Auto Average)
which can make a plant track the input command model as follows
signal with a controller whose transfer function 𝑦(𝑘 + 𝑑) = 𝐹[𝑦(𝑘),
𝐹 …
approximates the inverse of plant transfer function. 𝑦 𝑘 − 𝑛 , 𝑢(𝑘), … , 𝑢((𝑘 − 𝑛 )] (2)
Compared with traditional methods, AIC can achieve
specified
fied dynamic responses more easily and has Where 𝑛 and 𝑛 are the orders of input time series
better ability of disturbance rejection. Simulation and output time series respectively: 𝑛 < 𝑛 , 𝑑 is
comparison with previous scheme and adaptive PID system delay which can ensure the existence of
control are performed on the non non-linear ship system inverse model, and 𝐹(( . ) is certain non-linear
maneuvering to test the effect of improved algorithm. function. If 𝐹( . ) is monotonic change with (𝑘) ,
The impulse noise
oise detection and removal of impulse system (1)is invertible in the input spectrum 𝑆 . If the
noise while preserving the integrity of an image input signal is 𝑦 (𝑘 + 1)) at time(𝑘 + 1), the
expressed mathematically be discussed. controller can be expressed as
𝑢(𝑘) = 𝑔[𝑦 (𝑘 + 1), 𝑦(𝑘),( ) ……..,𝑦 𝑘 − 𝑛 +
The Improved Filtered-∈ Adaptive Inverse 1, 𝑢𝑘−1,………𝑢𝑘−𝑛𝑢+1(3)
Control
The Single Input Single Output (SISO) discrete
discrete-time
Function 𝐹( . ) and 𝑔( . ) are both unique. Under the
nonlinear system to be considered
idered is described by n
dreive of 𝑢(𝑘), the plant output will track the input
dimension state equations as below
signal of the inverse controller.
@ IJTSRD | Available Online @ www.ijtsrd.com | Volume – 2 | Issue – 5 | Jul-Aug 2018 Page: 1993
International Journal of Trend in Scientific Research and Development (IJTSRD) ISSN: 2456-6470
each input pattern. Each set of 𝑀 features is CONCLUSION
considered as a vector in the M-dimensional feature In this paper, an improved nonlinear filtered-𝜀 method
space. Pattern classification is accomplished by a self- is introduced and adaptive PID control are performed
organising neural network which is capable of cluster on the control of non-linear ship maneuvering and
analysis for image data. also mathematical model for ship Course-Keeping and
impulsive noise be expressed.
The neural network proposed in this paper is a
simplified version of ART I (Carpenter & Gross berg, REFERENCE
1987). However, it is capable of grey – level 1. Widrow B, Walach E. Adaptive inverse control.
classification. It has two layers: the input layer and the Englewood cliffs, N J: Prentice-Hall press, 1996.
output layer. These layers are connected by feed
2. Zhang Y H, Liu S X. An adaptive automatic ship
forward paths. The nodes of output layer are lateral
steering system. Shipbuilding of china, 1991.
inhibitive to one another. They are designed as a
competitive network capable of choosing the winning 3. Nieminen, A., Heinonen, P., & Neuvo, Y. (1987).
node. The set of weights 𝑢 for any specific node A new class of detail-preserving filters for image
𝑖 constitute the prototype pattern for cluster𝑖. The processing. IEEE Transactions pattern Analysis
clustering algorithm starts from isolated patterns and Machine Intelligence.
coalesces the nearest patterns or groups according to a 4. Zhang Xianku, Lu Xiaofei, Guo Chen, Yang
threshold from the bottom up to form hierarchies. The Yansheng, Robust Neural Network Control for
essential point of this algorithm is to build up the Ship’s Course-Keeping, Journal of Ship
clusters using the Euclidean distance measure Mechanics, Vol. 10, Issue 5, 2006.
between the input 𝑠 and the weights𝑢 . Assume a
sequence of pattern samples 5. Du Gaang, Zhan Xingqun, Zhong Shan, Zhang
Weiming, Adaptive inverse control of non-linear
𝑠 = (𝑠 , 𝑠 ,….𝑠 ) 𝜖 𝑅
ship manueuvering based on Neural Networks,
and a set of weight vectors
Measuring and Testing Technique, Vol. 24 Issue
𝑢 = (𝑢 , 𝑢 , … , 𝑢 ) ∈ 𝑅
4, 2005.
It is well known that similarities exist between ART
and Kohonen’s self-organising map.
@ IJTSRD | Available Online @ www.ijtsrd.com | Volume – 2 | Issue – 5 | Jul-Aug 2018 Page: 1994