| Some military targets(command center, arsenal, etc.) have been moved to underground with the development of security and defense system, they are well sheltered and becoming more complex and stronger. Equipped with intelligent fuze, Earth penetration warhead(EPW) can not only attack simple structures(silo, underground tunnel, etc.), but also complex ones. Fuze has developed from delay fuze into modern forms with multi-function and multi-purpose. We studied components of penetration signal, proposed 3 filtering algorithms, analyzed signal overlap, and designed detonation control system and Hardware-in-the-Loop simulation system for fuze. The main novelties and contributions are:(1) We proposed Wavelet-PCA-ICA method for Single Channel Blind Source Separation(SCBSS). Penetration signals are typical super-Gaussian signals as they are usually obtained from one sensor and the kurtosis of them are positive. We proposed Wavelet-PCA-ICA method for SCBSS, analyzed the penetration signal and studied the relationship between separated independent components and vibration of substructures of the projectile-fuze system. Furthermore, we analyzed the interaction of substructures’ vibration based on the waveform of independent components.(2) We proposed an adaptive filtering algorithm for penetration signal. Penetration signals are non-stationary signals and statistics of them varies with time. It is difficult to realize optimal filtering with Wiener or Kalman Filter, but adaptive filter can. With less prior-knowledge needs of input signals, adaptive filter is a real-time signal processing method. The proposed nonlinear adaptive filtering algorithm has 3rd-order nonlinear weight, which ensures deceleration curve is smooth everywhere(not only continuous) and avoids sharp points, which is crucial for target detection. The algorithm is superior to others as it filters high-frequency part of the penetration signal while keeps the high-frequency part.(3) We analyzed penetration signal overlap when EPW penetrates into multi-layer targets. Penetration signal will overlap in both time and frequency domain when the penetration resistance and the natural frequency of the projectile-fuze system are not far away, and it is impossible to extract target characteristics even by filtering. The parameters of the projectile-fuze system are pulse width and gap between adjacent pulses of exciting signals, stiffness and damping. Pulse width and gap decide decay time and velocity of residual vibration while stiffness decides natural frequency of the projectile-fuze system. Pulse width and gap correspond to initial velocity and thickness of hard targets respectively. Penetration signal overlap can be 3 cases: no overlap, time overlap and time-frequency overlap. For the case of time overlap, filtering is an effective way; but detection of multi-layers is impossible when it comes to time-frequency overlap, in which case auxiliary methods are needed.(4) We designed a detonation control system based on DSP and a low-power system based on MCU. With state design theory, we designed the hardware and software of the system based on DSP. The system has 3 channels with 250 KHz & 16 bit ADC, 512K×16bit Flash and PCB350C02 piezoelectric accelerometers which could bear 50,000 g shock. The field test showed that power consumption is the biggest problem of the system, which is beyond our requirements. By analyzing power consumption in system level and the corresponding methods to lower the consumption, we proposed a new scheme based on MCU.(5) We designed a Hardware-in-the-Loop simulation system, which can be used to evaluate fuze. The system consists of 5 parts: simulation equipment, evaluation board, interface, control platform and supported software. The simulation error of the system is less than 5%, therefore it can avoid the difficulties in the modeling of some parts in mathematical simulation and the difficulties and danger of fuze test in field. |
PDF Full Text Request