Simplified Lqg Control with Neural Networks
نویسنده
چکیده
A new neural network application for non-linear state control is described. One neural network is modelled to form a Kalmann predictor and trained to act as an optimal state observer for a non-linear process. Another neural network is modelled to form a state controller and trained to produce a minimum variance controller for the non-linear process. After training, tuning possibilities for the observer as well as for the controller are introduced to improve the closed loop robustness and noise suppression. The advantage of this method is that tuning takes place after the time consuming training session. The method is illustrated by a simple, multi variable example. Résumé: Une nouvelle application de réseaux neuraux pour un contrôle non linéaire d’état est décrite. Un predicteur de Kalmann formé à l’aide d’un réseau neural est entraîné pour agir comme un observateur d’état optimal pour un processus non linéaire. Un autre réseau neural modelisant un contrôleur d’état est développé et entraîné comme un contrôleur à variance minimale pour le processus non linéaire. Après l’entraînement, les possibilités d’ajustement et pour l’observateur et pour le contrôleur sont introduites pour améliorer la robustesse et la suppression de bruit du système en boucle fermé. L’avantage de cette méthode est que la procédure d’ajustement n’intervient qu’aprés que le long processus d’entraînement soit expiré. Un exemple à multi-variables est utilisé pour l’illustration de la méthode.
منابع مشابه
Modeling and Control of Gas Turbine Combustor with Dynamic and Adaptive Neural Networks (TECHNICAL NOTE)
متن کامل
Optimal Sequence-Based Control of Networked Linear Systems
Digital data networks have been used in control applications for more than 20 years to connect sensors, actuators, and controllers of a control loop. Typically, highly specialized networks are applied, the so called fieldbuses. If sufficiently dimensioned, these networks ensure deterministic data transmission with guaranteed latency. However, the increasing availability, flexibility, and lower ...
متن کاملOptimal LQG Control over Lossy Packet Networks
This paper is concerned with control applications over lossy data networks. Sensor data is transmitted to a estimation-control unit over a network, and control commands are issued to subsystems over the same network. Sensor and control packets may be lost or delayed. This is modeled by allowing that sensor and control packets are successfully received with certain probabilities. In this context...
متن کاملBelief Space Planning Simplified: Trajectory-Optimized LQG (T-LQG) (Extended Report)
Planning under motion and observation uncertainties requires solution of a stochastic control problem in the space of feedback policies. In this paper, we reduce the general (n +n)-dimensional belief space planning problem to an (n)-dimensional problem by obtaining a Linear Quadratic Gaussian (LQG) design with the best nominal performance. Then, by taking the underlying trajectory of the LQG co...
متن کاملSeven-Level Direct Torque Control of Induction Motor Based on Artificial Neural Networks with Regulation Speed Using Fuzzy PI Controller
In this paper, the author proposes a sensorless direct torque control (DTC) of an induction motor (IM) fed by seven-level NPC inverter using artificial neural networks (ANN) and fuzzy logic controller. Fuzzy PI controller is used for controlling the rotor speed and ANN applied in switching select stator voltage. The control method proposed in this paper can reduce the torque, stator flux and to...
متن کامل