Multilayer Perceptrons to Approximate Quaternion Valued Functions

In this paper a new type of multilayer feedforward neural network is introduced. Such a structure, called hypercomplex multilayer perceptron (HMLP), is developed in quaternion algebra and allows quaternionic input and output signals to be dealt with, requiring a lower number of neurons than the real MLP, thus providing a reduced computational complexity. The structure introduced represents a generalization of the multilayer perceptron in the complex space (CMLP) reported in the literature. The fundamental result reported in the paper is a new density theorem which makes HMLPs universal interpolators of quaternion valued continuous functions. Moreover the proof of the density theorem can be restricted in order to formulate a density theorem in the complex space. Due to the identity between the quaternion and the four-dimensional real space, such a structure is also useful to approximate multidimensional real valued functions with a lower number of real parameters, decreasing the probability of being trapped in local minima during the learning phase. A numerical example is also reported in order to show the efficiency of the proposed structure. Copyright 1997 Elsevier Science Ltd. All Rights Reserved.

[1]  James L. McClelland,et al.  Parallel distributed processing: explorations in the microstructure of cognition, vol. 1: foundations , 1986 .

[2]  Toshimichi Saito,et al.  Torus doubling and hyperchaos in a five dimensional hysteresis circuit , 1994, Proceedings of IEEE International Symposium on Circuits and Systems - ISCAS '94.

[3]  Luigi Fortuna,et al.  On the capability of neural networks with complex neurons in complex valued functions approximation , 1993, 1993 IEEE International Symposium on Circuits and Systems.

[4]  Francesco Piazza,et al.  On the complex backpropagation algorithm , 1992, IEEE Trans. Signal Process..

[5]  T. Nitta,et al.  A back-propagation algorithm for complex numbered neural networks , 1993, Proceedings of 1993 International Conference on Neural Networks (IJCNN-93-Nagoya, Japan).

[6]  Giovanni Muscato,et al.  An hypercomplex neural network platform for robot positioning , 1996, 1996 IEEE International Symposium on Circuits and Systems. Circuits and Systems Connecting the World. ISCAS 96.

[7]  Ken-ichi Funahashi,et al.  On the approximate realization of continuous mappings by neural networks , 1989, Neural Networks.

[8]  Luigi Fortuna,et al.  Neural networks for quaternion-valued function approximation , 1994, Proceedings of IEEE International Symposium on Circuits and Systems - ISCAS '94.

[9]  William Rowan Hamilton,et al.  Lectures on quaternions , 1853 .

[10]  Keiichi Abe,et al.  Towards a Hierarchical Contour Description via Dominant Point Detection , 1994, IEEE Trans. Syst. Man Cybern. Syst..

[11]  Francesco Piazza,et al.  Non linear satellite radio links equalized using blind neural networks , 1991, [Proceedings] ICASSP 91: 1991 International Conference on Acoustics, Speech, and Signal Processing.

[12]  Cris Koutsougeras,et al.  Complex domain backpropagation , 1992 .

[13]  K. Spring.  Euler parameters and the use of quaternion algebra in the manipulation of finite rotations: A review , 1986 .

[14]  Jack C. K. Chou,et al.  Finding the Position and Orientation of a Sensor on a Robot Manipulator Using Quaternions , 1991, Int. J. Robotics Res..

[15]  Luigi Fortuna,et al.  Predicting complex chaotic time series via complex valued MLPs , 1994, Proceedings of IEEE International Symposium on Circuits and Systems - ISCAS '94.

[16]  Kazuhiro Ueda,et al.  Digital filters with hypercomplex coefficients , 1993, 1993 IEEE International Symposium on Circuits and Systems.

[17]  Kurt Hornik,et al.  Multilayer feedforward networks are universal approximators , 1989, Neural Networks.

[18]  David Hestenes,et al.  Invariant body kinematics: I. Saccadic and compensatory eye movements , 1994, Neural Networks.

[19]  Luigi Fortuna,et al.  Multilayer perceptrons to approximate complex valued functions , 1995, Int. J. Neural Syst..

[20]  C. Deavours.  The Quaternion Calculus , 1973 .

[21]  Francesco Piazza,et al.  A COMPARISON BETWEEN REAL AND COMPLEX VALUED NEURAL NETWORKS IN COMMUNICATION APPLICATIONS , 1991 .

[22]  P. Arena,et al.  Neural networks in multidimensional domains , 1998 .

[23]  Henry Leung,et al.  The complex backpropagation algorithm , 1991, IEEE Trans. Signal Process..

[24]  J.C.K. Chou,et al.  Quaternion kinematic and dynamic differential equations , 1992, IEEE Trans. Robotics Autom..

[25]  J. Wen,et al.  The attitude control problem , 1991 .

[26]  George Cybenko,et al.  Approximation by superpositions of a sigmoidal function , 1992, Math. Control. Signals Syst..

[27]  T. Nitta.  A backpropagation algorithm for neural networks based an 3D vector product , 1993, Proceedings of 1993 International Conference on Neural Networks (IJCNN-93-Nagoya, Japan).

[28]  K. Gürlebeck,et al.  Quaternionic Analysis and Elliptic Boundary Value Problems , 1989 .

[29]  L. Shen,et al.  Linear Algebra , 1968 .

[30]  Clark C. Guest,et al.  Modification of backpropagation networks for complex-valued signal processing in frequency domain , 1990, 1990 IJCNN International Joint Conference on Neural Networks.

[31]  R. G. Fenton,et al.  Computational Analysis of Robot Kinematics, Dynamics, and Control using the Algebra of Rotations , 1994, IEEE Trans. Syst. Man Cybern. Syst..

[32]  Russell H. Taylor,et al.  Planning and execution of straight line manipulator trajectories , 1979 .

[33]  Kang G. Shin,et al.  Intelligent Coordination of Multiple Systems with Neural Networks , 1991, 1991 American Control Conference.

[34]  Luigi Fortuna,et al.  Chaotic time series prediction via quaternionic multilayer perceptrons , 1995, 1995 IEEE International Conference on Systems, Man and Cybernetics. Intelligent Systems for the 21st Century.