A method has been discovered to enhance the satellite automatic control system.

The automatic control system gathers data about an object and generates the appropriate response, ensuring the device operates smoothly. Information is collected through sensors, whose signals must be converted into digital code using an analog-to-digital converter. Within this converter are neurons responsible for this process. If one of these neurons fails, repairs are necessary, which may not always be done quickly, especially in artificial satellites orbiting Earth. Researchers at Perm Polytechnic University have proposed a solution to reduce the risk of equipment failure by enabling the measuring neurons to autonomously diagnose their own malfunctions and "decide" which of them will replace the faulty component.

2025-01-30 10:02:04https://naked-science.ru/article/column/avtomaticheskogo-upravlen

They have developed a prototype of a device that will enhance the fault tolerance of the entire system in the future.

Development is presented as part of a competition for research conducted in research laboratories and student design bureaus.

The analog-to-digital converter is responsible for data collection for the automatic control system. This device translates physical quantities (temperature, pressure, sound) into digital form using measuring components. The information is sent to sensors that convert it into an electrical signal, which is then transmitted to neurons in the converter, where the data is "translated" into a code understandable by the control system. At the core of this process is an integrated circuit (chip on the board) similar to the microprocessor found in every computer.

For instance, this is how climate control works: the analog signal in this case is the temperature level. Sensors read the temperature and send it to the converter, where its neurons "explain" to the control system what the room temperature is. If it exceeds the permissible level, air cooling is activated.

If one of the measuring components fails, the automatic control system stops functioning correctly. Therefore, there is a need to enhance its fault tolerance. This is especially important for applications in hard-to-reach maintenance areas, such as in unmanned aerial vehicles and artificial satellites.

Scientists from Perm Polytechnic University have developed a prototype technology featuring self-replacing measuring neurons. This will reduce the risk of failure in automatic control systems and increase the device's lifespan, regardless of its location.

“To extend the lifespan of the analog-to-digital converter, we propose using a network of neurons that can reorganize and replace each other in case of a failure in one of them. The network will conduct the reorganization itself, utilizing additional connections between them. This will allow for the exclusion of malfunctioning elements through embedded self-diagnosis algorithms,” explains Ilya Artemyev, a student from the Department of Automation and Telemechanics at PNIU.

“Currently, the prototype consists of a discrete board where the effectiveness of the technology can be tested. We believe that due to the reorganization of neurons, our development may take 50 percent more time to measure compared to a conventional converter. However, this drawback is effectively offset by the fact that in our case, the same signal is measured in parallel by several neurons. While one of them is reorganizing, the other continues to operate, ensuring the overall functionality of the converter. We have set up and debugged the software part, but ultimately, a complete integrated circuit should be developed for full operation,” comments Anton Posyagin, an associate professor at the Department of Automation and Telemechanics at PNIU and a candidate of technical sciences.

The prototype created by PNIU scientists serves as the foundation for further development, which can be utilized in various automatic control systems, such as unmanned aerial vehicles, artificial satellites, and other industrial applications. The implementation of the analog-to-digital converter with self-replacing neurons will reduce the risk of equipment failure.