First of all, intelligent automation technology opens up a broad prospect for the application of related fields of instrumentation and measurement. Using intelligent software and hardware, each instrument or meter can accurately analyze and process the current and previous data information at any time, and appropriately abstract the measurement process from low, medium and high levels to improve the existing measurement system. The performance and efficiency of traditional measurement systems are extended, such as the use of intelligent technologies such as neural networks, genetic algorithms, evolutionary calculations, and chaos control, to enable instruments to achieve high-speed, high-efficiency, multi-function, high mobility and flexibility.
Secondly, you can also use microchip technology such as microprocessors and microcontrollers in different instruments of the decentralized system, design fuzzy control programs, set critical values for various measurement data, and use fuzzy rule fuzzy inference technology to Various fuzzy relationships make various types of fuzzy decisions. Its advantage is that it does not need to establish a mathematical model of the controlled object, nor does it need a large amount of test data. It only needs to summarize appropriate control rules based on experience, apply chip offline calculation and field debugging, and generate accurate according to our needs and accuracy. Analysis and punctual control actions.
Especially in sensor measurement, the application of intelligent automation technology is more extensive. Using software to implement signal filtering, such as fast Fourier transform, short-time Fourier transform, wavelet transform and other technologies, is an effective way to simplify the hardware, improve the signal-to-noise ratio, and improve the dynamic characteristics of the sensor , but it is necessary to determine the dynamic mathematical model of the sensor , and the higher order The filter has poor real-time performance. Using neural network technology, high-performance autocorrelation filtering and adaptive filtering can be realized. Among them, real-time and non-real-time, fast-changing and slow-changing, fuzzy and deterministic data information may support each other, or they may contradict each other. At this time, the extraction and fusion of object features until the final decision is made, and a correct judgment is made. It will be difficult. So neural network or fuzzy logic will be the most worthy method. For example, the gas sensing array is used for mixed gas identification. In the signal processing method, a combination of self-organizing mapping network and BP network can be used to first classify and then identify the components. To reduce the complexity of the algorithm and improve the recognition rate. For another example, the difficulty of detecting and identifying food taste signals was once the main obstacle for research and development units. Today, wavelet transforms can be used for data compression and feature extraction, and then inputting the data into a fuzzy neural network trained with a genetic algorithm, which greatly improves the recognition rate for simple compound flavors. For another example, in the evaluation of the quality of cloth fabrics, the use of flexible hands for tactile signal processing, and the fault diagnosis of machines, intelligent automation technology has also achieved a large number of successful examples. Make full use of the powerful self-learning, self-adaptation, self-organization capabilities of artificial neural network technology, associative and memory functions, and the black box mapping characteristics of inputs and outputs for non-linear complex relationships, both in terms of applicability and fast real-time Both will greatly exceed complex functional formulas, which can make full use of multi- sensor resources and obtain more accurate and credible conclusions comprehensively.