The following text was used as the abstract in for a government grant on this subject.
Networked control systems aim to overcome the disadvantages of conventional digital control systems at the application level. Disadvantages such as difficulty of modification, vulnerability to electrical noise, difficulty in maintenance and upgrades can be overcome by methods offered by networked control systems. If the networked control system idea, consisting of a sensor node, controller node and actuator node, each implemented as independent computers connected by a communication network, is applied in its simplest form, it will cause other problems while solving those just stated. The greatest problem is the variable and unbounded delay caused by communication protocols, and the next is the data loss over the network. Most of the research in the literature assumes lossless communication networks. Therefore, we can say that the proposed project aims to cover an important gap in the literature. The results of this project are directly applicable to the needs of the industry.
In a conventional control system designed according to digital control theory and working at the lowest possible sampling frequency, if the output of the plant is not measured, the control outputs calculated and applied to the plant by the actuators within each and every sampling period, instability may result. Unfortunately, general purpose communication networks cause just such problems with their random delay and data loss properties. Although communication networks with guaranteed delay upper bounds and loss rates exist, they are not flexible or widely used except for special applications.
In this project, we propose a novel networked control method where satisfactory control is possible even under random delay and data loss. The idea is to keep a model of the plant inside the controller node and at every sampling period reset the states of this model to the actual or predicted states of the plant. Then, not only the current control output, but also a number of control signal estimates into the future are calculated and sent to the actuator node within the same data packet. This way, if the communication between the sensor node and actuator node or the controller node and the actuator node is temporarily interrupted, previously estimated control outputs at the actuator are applied to the plant. If this method is augmented in some ways, it can produce satisfactory control even under a data loss rate of 80%. This has been the subject of a masterfs degree thesis advised by the principal investigator of this project. The results will be presented at an international conference.
The purpose of this project is to find the theoretical and application-specific aspects of the method explained above, and improve it. The foremost aim is to theoretically derive the stability criteria of the method and compare the results with simulations and experiments. Another aim is to investigate the effects of the difference between the plant and its model inside the controller node, and also of sensor noise. The effects of data loss and the choice of communication network protocol will also be investigated.
The theoretical results will first be verified first on the simulation test bed which has already been completed, and later, on a dedicated experimental platform, which has partially been built by the project team. After the project is completed, the purchased equipment can be used for a successor project or for educational purposes. It is estimated that the project will be completed in 24 months, and the project team be composed of one principal investigator, one doctoral degree candidate and one masterfs level student.
There is an urgent need in the industry for the application of the outcome of the project. There is no research going on on this subject in Turkey to the knowledge of the author and relevant research in the world is just starting, which suggests that this is a good field to work in. The main idea is an original idea proposed by the author. The project is the first step in the final ideal of developing methods for complex distributed control systems. The information and experience gained from this project will be the basis for further projects in this area.
The added value from this project overlaps well and with the goals of the requirements of key technologies to be developed as set out in TUBITAKfs Vision 2023 documents in several areas. As explained in detail further on in the document, it provides a basis for the technologies that are planned to be developed in Turkey.
After the main results of the project are obtained, new projects with industrial partners are targeted, since the outcomes of the project are suitable to be directly used in the industry.
Keywords: Real-time systems, embedded systems, automatic control, distributed control, lossy computer networks, variable time delay.