Tustin neural networks: a class of recurrent nets for adaptive MPC of mechanical systems

The use of recurrent neural networks to represent the dynamics of unstable systems is difficult due to the need to properly initialize their internal states, which in most of the cases do not have any physical meaning, consequent to the non-smoothness of the optimization problem. For this reason, in this paper focus is placed on mechanical systems characterized by a number of degrees of freedom, each one represented by two states, namely position and velocity. For these systems, a new recurrent neural network is proposed: Tustin-Net. Inspired by second-order dynamics, the network hidden states can be straightforwardly estimated, as their differential relationships with the measured states are hardcoded in the forward pass. The proposed structure is used to model a double inverted pendulum and for model-based Reinforcement Learning, where an adaptive Model Predictive Controller scheme using the Unscented Kalman Filter is proposed to deal with parameter changes in the system.

[1]  Boris Murmann,et al.  Long Short-Term Memory Neural Network Equilibria Computation and Analysis , 2018 .

[2]  Colin N. Jones,et al.  Recurrent Neural Network based MPC for Process Industries , 2019, 2019 18th European Control Conference (ECC).

[3]  Christopher D. Manning,et al.  Improved Semantic Representations From Tree-Structured Long Short-Term Memory Networks , 2015, ACL.

[4]  Jürgen Schmidhuber,et al.  LSTM: A Search Space Odyssey , 2015, IEEE Transactions on Neural Networks and Learning Systems.

[5]  S. Hochreiter.  Recurrent Neural Net Learning and Vanishing , 1998 .

[6]  Frank Allgöwer,et al.  Discrete-time Incremental ISS: A framework for Robust NMPC , 2013, 2013 European Control Conference (ECC).

[7]  Jürgen Schmidhuber,et al.  Learning to forget: continual prediction with LSTM , 1999 .

[8]  J. Dormand,et al.  A family of embedded Runge-Kutta formulae , 1980 .

[9]  Christian Osendorfer,et al.  NAIS-Net: Stable Deep Networks from Non-Autonomous Differential Equations , 2018, NeurIPS.

[10]  Ryan P. Adams,et al.  On the Difficulty of Warm-Starting Neural Network Training , 2019, ArXiv.

[11]  Carl E. Rasmussen,et al.  Model based learning of sigma points in unscented Kalman filtering , 2010, 2010 IEEE International Workshop on Machine Learning for Signal Processing.

[12]  Rudolph van der Merwe,et al.  The unscented Kalman filter for nonlinear estimation , 2000, Proceedings of the IEEE 2000 Adaptive Systems for Signal Processing, Communications, and Control Symposium (Cat. No.00EX373).

[13]  Reddi Kamesh,et al.  Novel Formulation of Adaptive MPC as EKF Using ANN Model: Multiproduct Semibatch Polymerization Reactor Case Study. , 2016, IEEE transactions on neural networks and learning systems.

[14]  Marcello Farina,et al.  Model predictive control design for dynamical systems learned by Long Short-Term Memory Networks , 2019, ArXiv.

[15]  Robert R. Bitmead,et al.  Persistently exciting model predictive control , 2014 .

[16]  Bjarne A. Foss,et al.  MPC-based dual control with online experiment design , 2015 .

[17]  Razvan Pascanu,et al.  On the difficulty of training recurrent neural networks , 2012, ICML.

[18]  Karl Johan Åström,et al.  Adaptive Control , 1989, Embedded Digital Control with Microcontrollers.

[19]  Hao Wang,et al.  Traffic Speed Prediction Under Non-Recurrent Congestion: Based on LSTM Method and BeiDou Navigation Satellite System Data , 2019, IEEE Intelligent Transportation Systems Magazine.

[20]  Frank Allgöwer,et al.  Robust MPC with recursive model update , 2019, Autom..

[21]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[22]  Petros G. Voulgaris,et al.  On optimal ℓ∞ to ℓ∞ filtering , 1995, Autom..

[23]  Marcello Farina,et al.  Model Predictive Control Design for Dynamical Systems Learned by Echo State Networks , 2019, IEEE Control Systems Letters.

[24]  Boris Murmann,et al.  Some Local Stability Properties of an Autonomous Long Short-Term Memory Neural Network Model , 2018, 2018 IEEE International Symposium on Circuits and Systems (ISCAS).

[25]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[26]  武田 一哉,et al.  Recurrent Neural Networkに基づく日常生活行動認識 , 2016 .

[27]  Ofelia Begovich,et al.  PREDICTIVE CONTROL WITH CONSTRAINTS OF A MULTI-POOL IRRIGATION CANAL PROTOTYPE , 2007 .

[28]  Nadine Gottschalk,et al.  Computer Controlled Systems Theory And Design , 2016 .

[29]  Kenji Doya,et al.  Bifurcations of Recurrent Neural Networks in Gradient Descent Learning , 1993 .

[30]  田畑 昭久,et al.  Model Based Learning に基づく工学導入教育支援教材の開発 , 2003 .