Analysis of output signals of angular position sensors for the use of neural networks

摘要

In many industrial applications the measurement of angles, rotational speed and currents is of prime importance. There are several different concepts in use, which focus on con-tactless and wear-free measurement. Especially magnetoresistive sensors possess a prime role for such applications due to their favourable characteristics like accuracy and robustness. However, the role of MR-based sensors is challenged by increasing requirements, which are not fulfilled by state-of-the-art sensing concepts completely. Hence the application of neural networks is examined, when the input data is generated by a sensor array. The focus of this work is the examination of the different output concepts and their feasibility with neural networks.