Automotive Engine Idle Speed Control with Recurrent Neural Networks

This paper describes the development of recurrent neural network controllers for an automotive engine idle speed control (ISC) problem. Engine ISC is a difficult problem because of troublesome characteristics such as severe process nonlinearities, variable time delays, time-varying process dynamics and unobservable system states and disturbances. We demonstrate that recurrent neural network controllers can be trained to handle these difficulties gracefully while achieving good regulator performance for a representative model of 4-cylinder, 1.6 liter engine. Empirical results clearly illustrate that neural network controllers with relatively large amounts of internal feedback provide more robust performance for the ISC problem than do neural network controllers that are static or contain limited internal recurrent connections.

[1]  Michael I. Jordan,et al.  Generic constraints on underspecified target trajectories , 1989, International 1989 Joint Conference on Neural Networks.

[2]  Lee A. Feldkamp,et al.  Neural control systems trained by dynamic gradient methods for automotive applications , 1992, [Proceedings 1992] IJCNN International Joint Conference on Neural Networks.

[3]  Kumpati S. Narendra,et al.  Identification and control of dynamical systems using neural networks , 1990, IEEE Trans. Neural Networks.

[4]  Maurizio Abate,et al.  Use of fuzzy logic for engine idle speed control , 1990 .

[5]  Richard S. Sutton,et al.  Adaptive State Representation and Estimation Using Recurrent Connectionist Networks , 1995 .

[6]  Lee A. Feldkamp,et al.  Practical considerations for Kalman filter training of recurrent neural networks , 1993, IEEE International Conference on Neural Networks.

[7]  J.A. Cook,et al.  Modeling of an internal combustion engine for control analysis , 1988, IEEE Control Systems Magazine.

[8]  Lee A. Feldkamp,et al.  Model reference adaptive control with recurrent networks trained by the dynamic DEKF algorithm , 1992, [Proceedings 1992] IJCNN International Joint Conference on Neural Networks.

[9]  Lee A. Feldkamp,et al.  Recurrent network training with the decoupled-extended-Kalman-filter algorithm , 1992, Defense, Security, and Sensing.

[10]  George Vachtsevanos,et al.  A systematic design method for fuzzy logic control with application to automotive idle speed control , 1992, [1992] Proceedings of the 31st IEEE Conference on Decision and Control.

[11]  Ronald J. Williams,et al.  A Learning Algorithm for Continually Running Fully Recurrent Neural Networks , 1989, Neural Computation.

[12]  B. K. Powell,et al.  Modern control applications in idle speed control , 1983, 1983 American Control Conference.

[13]  Lee A. Feldkamp,et al.  Neural network modeling and control of an anti-lock brake system , 1992, Proceedings of the Intelligent Vehicles `92 Symposium.

[14]  Alexander G. Parlos,et al.  Nonlinear dynamic system identification using artificial neural networks (ANNs) , 1990, 1990 IJCNN International Joint Conference on Neural Networks.

[15]  B. K. Powell,et al.  Modern control applications in idle speed control , 1983 .

[16]  Lee A. Feldkamp,et al.  Decoupled extended Kalman filter training of feedforward layered networks , 1991, IJCNN-91-Seattle International Joint Conference on Neural Networks.

[17]  Kumpati S. Narendra,et al.  Gradient methods for the optimization of dynamical systems containing neural networks , 1991, IEEE Trans. Neural Networks.