Lattice-Boltzmann interactive blood flow simulation pipeline

A fast direct locator for radiation source based on composite convolution neural network

“…In [9], multilayer perceptron (MLP) neural network was adopted to realize mapping between objective positions and positioning parameters in two-step method for indoor WiFi-based localization. However, deep neural networks often require a large number of fully connected relationships, which puts great pressure on real-time computation and pretraining of nodes, making it difficult to achieve the real-time superiority of advanced optimization methods as described in [10][11][12]. The scheme based on LSTM network is introduced in the context of ultra wide band (UWB) technology, whose relative positioning error can reach 1.4% [13], but the disadvantage is also high complexity.…”