From Fig. 9 and Fig. 10, it can be

From Fig. 9 and Fig. 10, it can be seen that the idealized system model of the unmanned helicopter has good pitch and yaw angle tracking under the condition of simulated noise, with flat tracking curve, fast response speed and small error.From Fig. 11 and Fig. 12, it can be seen that the semi-physical simulation model of the unmanned helicopter is affected by actual and simulated noise in the actual environment, and the actual trajectory of the pitch and yaw angles cannot track the given pulse signal well, and the steady-state error and overshoot are larger. After filtering by Kalman filter, the pitch and yaw angles can track the impulse signal well, and the steady-state error and overshoot are effectively reduced, and the external interference and noise are better suppressed. Compared with the idealized model, the simulation curve of the semi-physical model has a very small oscillation, the response speed is slow and the error is larger, but the results are more realistic and can restore the actual working environment of the unmanned helicopter to a great extent, which makes the results of this experiment more practical.