CONCEPT OF INTELLIGENT MEASURING SYSTEM FOR ANALYZING THE ENERGY CONSUMPTION OF IOT MODULES

Abstract

The conducted analysis showed the shortcomings of known techniques for measuring MCU (microprocessor systems and IoT modules - hereinafter MCU) current consumption. In order to improve self-powered systems, designed to perform complex algorithms (learning ANNs, searching for signatures of suspicious software code, etc.), it is necessary to minimize the energy consumption of software, since hardware of modern MCU has low consumption. Optimizing MCU software in terms of energy consumption requires experimental studies of the energy consumption of the execution of instructions, commands and programs. Due to the lack of specialized measuring equipment that has sufficient accuracy (the error must be less than the difference in power consumption of instructions and commands) and immunity to interference, known models do not allow optimizing the software in terms of power consumption and thereby increasing the operating time of autonomously powered systems without recovery battery charge. And this is important for MCU and IoT modules operating in remote locations, dangerous for people, and critical infrastructure systems. However, a direct study of the energy consumption of MCU instructions, commands and programs as part of IoT modules is very time-consuming. RISC microcontrollers have few commands, but many modes of their execution, so it would be necessary to investigate 10-20 thousand options. It is proposed to use the methods of artificial intelligence for the classification of teams and the forecast of energy consumption of those teams that were not studied. The basis of the concept of a measuring system for the analysis of current consumption of smart devices and IoT modules can be the architecture developed in this paper. At the same time, a more detailed study of them and provision of noise protection, in particular, due to non-contact measuring devices, is required.