Add Thesis

Using machine learning for control systems in transforming environments

Written by F. Barkrot, M. Berggren

Paper category

Bachelor Thesis

Subject

Computer Science

Year

2020

Abstract

Thesis Control theory: Control theory involves automatic systems and open-loop and closed-loop systems. An automatic system refers to a system that can work without human supervision. Control theory is an interdisciplinary subject, which means that it is used in many different branches, such as vehicles, robotics, and space technology [27]. Without a control system, our technology today might be very different. The control system enables many of our machines to work as expected. The control system is usually based on the principle of feedback. An input signal is compared with a reference signal representing the goal to be achieved [9]. In order to be able to build a control system, a good understanding of the process to be adjusted is required. The most important factors are the output of the system and their response to changes in input. Input is some kind of command or stimulus applied to the system. The form of input and output can vary. The inputs and outputs need to be given in order to be able to identify the components of the system. A control system can have multiple inputs and outputs [8]. A process has static and dynamic characteristics. The static characteristic is the characteristic in the static state of the process, which means that static amplification is ubiquitous in the work of the process. The dynamic characteristics of the process will consider rigidity, time delay and transients. For example, when you step on the pedal of a car, it takes some time to get the desired speed. The static characteristics between different systems are usually very similar, while the dynamic characteristics may be different. Therefore, different processes can be divided into different types. There is a shutdown process, an overshoot process and so on. The inverted pendulum is an unstable process. This means that feedback is required to get the output close to the required reference signal. There are many steps to go through when designing a control system. The control system must first be studied. In order to be able to build the system, it is necessary to know what types of sensors are necessary, what actuators to use and where they should be placed. After all the necessities of the system have been determined, it can be modeled. After the system is modeled, the attributes of the generated model can be determined, and performance specifications can be set. According to the previous conclusions, the type of controller to be used is set, and the design of the controller meets the measured characteristics. There are different types of controllers available for different types of systems. Some of these controllers will be explained in more detail in the following sections. After determining the appropriate controller, the system can be simulated. 2.1.1 Open-loop and closed-loop control systems Control systems are generally divided into two categories: open-loop and closed-loop systems. These systems are separated according to the control actions responsible for activating the system and generating output. The word action in the word control action is not necessarily change, movement, or activity. In a system designed to allow an object to hit a target, the control action is the distance between the object and the target. Distance is not an action, but in this case, since the object hits the target, the action movement is implicit. In an open-loop control system, the control action has nothing to do with the output, while in a closed-loop control system, the control action depends on the output. The ability of an open-loop control system to perform accurately depends entirely on its calibration. When the system is calibrated, the input-output relationship is established to obtain the desired system accuracy. Closed loop control system is also called feedback control system. This article will use this system, and the feedback control system will be explained in more detail in the next section. [27] 2.1.1.1 Feedback control system Feedback is the main difference that separates the closed-loop control system from the open-loop control system. Feedback allows the output to be compared with the system input and allows appropriate control actions to be formed based on the output and input. The existence of feedback gives the system many different attributes. For example, a feedback control system improves accuracy, while feedback reduces the influence of external interference or noise. As shown in Figure 2.1, the output signal is fed back to the object to be controlled, which is called feedback. In addition to feedback, there are some external signals, such as external interference and sensor noise. These signals are called external signals [9]. The controller monitors the process and has an input variable, the command input, which is compared with the set point as the desired result. If the input is the same as the set point, the control system has reached the target, if they are different, the controller sends the object input to the object to tell it how to perform to reach the set point. 2.1.2 Controller The controller is the core of the control system. The task of the controller in the system is to use the information from the feedback to create a control signal in an attempt to reduce the error. The simplest form of the controller is the switch controller, in these controllers, the control signal can only take two different values. The value of the signal depends on the output, whether it is positive or negative. The switch controller is simple, but not always accurate enough. As a more advanced switch controller, there are multi-level controllers. Read Less