Design and Development of Control Systems for an Aircraft. Comparison of Performances Through Computational Simulations

Garrido

Kern

2021

Sensors

Self Cite

This paper presents the results of the design, simulation, and implementation of a virtual vehicle. Such a process employs the Unity videogame platform and its Machine Learning-Agents library. The virtual vehicle is implemented in Unity considering mechanisms that represent accurately the dynamics of a real automobile, such as motor torque curve, suspension system, differential, and anti-roll bar, among others. Intelligent agents are designed and implemented to drive the virtual automobile, and they are trained using imitation or reinforcement. In the former method, learning by imitation, a human expert interacts with an intelligent agent through a control interface that simulates a real vehicle; in this way, the human expert receives motion signals and has stereoscopic vision, among other capabilities. In learning by reinforcement, a reward function that stimulates the intelligent agent to exert a soft control over the virtual automobile is designed. In the training stage, the intelligent agents are introduced into a scenario that simulates a four-lane highway. In the test stage, instead, they are located in unknown roads created based on random spline curves. Finally, graphs of the telemetric variables are presented, which are obtained from the automobile dynamics when the vehicle is controlled by the intelligent agents and their human counterpart, both in the training and the test track.

Section: State Of Research On Intelligent Agentsmentioning

confidence: 99%

Design and Implementation of Intelligent Agent Training Systems for Virtual Vehicles

Garrido

Kern

2021

Sensors

Self Cite

“…These tools can be used for both face-to-face lessons and learning experiences out of the classroom, such as virtual laboratories (remote access to simulated systems) and remote laboratories (remote access to real equipment). To teach control systems in face-to-face environments, software packages such as MatLab or LabView have been the preferred option, since they contain several libraries and specialized tools (toolboxes) that even allow users to combine simulation environments with physical laboratory equipment [16][17][18][19].…”

Section: Introductionmentioning

confidence: 99%

Integrating ROS and IoT in a Virtual Laboratory for Control System Engineering

Journal of Applied Mathematics

Kern

2020

Self Cite

This article presents the implementation of a learning environment for the teaching of control systems. This environment integrates physical equipment and simulation, monitoring, and control through a network. A software platform based on ROS (Robotic Operating System) grants access to the system through intranet and Internet, facilitating the integration of new test equipment. The environment developed can be used in didactic experiences both inside and outside the classroom, enhancing the learning of four main study topics: modeling, analysis, parameter estimation, and controller design.

Design and Performance Analysis of Level Control Strategies in a Nonlinear Spherical Tank

Garcia-Garcia

2023

Processes

This work seeks to contribute to the study of techniques for level control considering a nonlinear plant model. To achieve this goal, different approaches are applied to classical control techniques and their results are analyzed. Fuzzy Logic Control (FLC), Artificial Neural Network (ANN), Adaptive Neuro-Fuzzy Inference System (ANFIS), Model Predictive Control (MPC) and Nonlinear Auto-Regressive Moving Average (NARMA-L2) controllers are designed for the level control of a spherical tank. Subsequently, several tests and scenarios similar to those present in industrial processes are established, while the transient response of the controllers, their performance indices for monitoring the reference value, the rejection of disturbances, the presence of parameter uncertainties and the effects of noise are analyzed. The results show good reference tracking, with a settling time of approximately 5 s for 5 cm and a rise time of less than 4 s. No evidence for steady-state error or overshoot was found and controllers behave positively in the diverse scenarios assessed. The FLC and ANN controllers showed the greatest limitations, while ANFIS, MPC and NARMA-L2 exhibited competitive results considering their transient response and the performance indices calculated.