Abstract
To optimize the radial basis function (RBF), an improved genetic algorithm (IGA) was used for measuring high suspended sediment concentration (HSSC) in the Yellow River. The improved probabilities crossover and mutation were utilized to realize the twin goals of maintaining diversity in the population and sustaining the convergence capacity of the GA, and a new method was proposed to resolve the problem of deciding the optimal values of the probabilities of crossover and mutation. In this method, the effects of environmental factors, such as temperature, depth and flow rate on the capacitive differential pressure (CDP) sensors, were discussed. To compare the fusion effect of the IGA-RBF method, the sediment concentration data were fused and processed using unary linear regression (ULR), multiple linear regression (MLR), back propagation (BP) neural network, standard RBF (S-RBF) methods, adaptive GA-optimized RBF (AGA-RBF) and dynamic GA-optimized RBF (DAGA-RBF). The results show that the IGA-RBF method can effectively eliminate environmental influences and raise the measuring accuracy and stability of the system.