Atmospheric PM_2.5 concentration prediction and noise estimation based on adaptive unscented Kalman filtering

Due to the randomness and uncertainty in atmospheric environment, accompanied by a variety of unknown noise. Accurate prediction PM 2.5 concentration is very important for people prevent injury effectively. In order predict more accurately this hybrid modelling method support vector regression adaptive unscented Kalman filter (SVR-AUKF) proposed case incorrect or Firstly, model was established regression. Secondly, state space framework combined with estimate uncertain noise through continuous updating when unknown. Finally, compared SVR-UKF method, simulation results show that accurate robust. The SVR-UKF, AR-Kalman, AR BP methods. has higher accuracy concentration.