The Short-Term Traffic Flow Prediction Method Based on Detectors PSO Algorithm

According to the shortcoming that traffic detector configuration is not considered in traffic flow prediction, this paper proposes an algorithm for short-term traffic flow prediction based on traffic sensor optimization with particle swarm optimization. The paper takes the minimum mean square prediction error as the objective function, selects suitable detectors with PSO algorithm, and predicts the short term traffic flow with wavelet neural network. The test results with America I-84 highway measured data show that the proposed algorithm can ensure the accuracy of short-term traffic flow prediction, the investment costs can be reduced effectively, and can provide the certain theory instruction for the optimal layout of traffic detector in highway engineering.