A novel real-time detection of orthogonal transient weak ELF magnetic signals

Abstract

Real-time detection of transient weak magnetic signals is a common problem in many areas, such as tracking of in-pipe robots by the transmitting and receiving of extremely low frequency (ELF) magnetic signals, which is a critical issue related to the safety of pipelines and attracts special attention. Owing to the limit of transmitting power and interruption in open air as well as the mobility of the robot, the received ELF magnetic signals are relatively weak and transient, and are even as low as 10 pico-tesla, and the related SNR is as low as −3 dB. In this paper, a novel transient weak ELF magnetic signal detection method is proposed based on orthogonal search coil sensors. First, a Morlet wavelet is employed to model received orthogonal transient signals in which the envelope decay rate represents moving velocity. Second, a least square criterion is used to construct orthogonal signal energy statistic, then a one-to-one relationship between the envelope decay rate and the moving velocity is established. Third, the ternary decision tree method is developed to maximize orthogonal signal energy statistic and realize optimal real-time signal detection under the Newman-Pearson criterion. The simulation and the experimental results both show that the detection performance is quite close to the theoretical upper boundary, and the developed instruments based on the orthogonal search coil sensors are effective in real-time tracking.