In view of the sounding needs of long-endurance highaltitude UAV

a temperature sensor carried by bomb-dropping sonde at an altitude of 0~16km is designed.The computational fluid dynamics(CFD)method is used to analyze external multiple physical fields of the sonde casing.After location and wind speed of the temperature sensor are obtained

CFD is further used to calculate the relationships among altitude

wind speed

solar radiation intensity and solar radiation error.Subsequently

the data are trained using the particle swarm optimization-back propagation(PSO-BP)algorithm and the predictive model is obtained.Finally

a low-pressure wind tunnel experimental platform is constructed to conduct comparative experiments to validate the radiation error of the temperature sensor.The experimental results show that the mean absolute error(MAE)of the temperature sensor for sonde carried on the long-endurance high-altitude UAV platform is 0.0123K

and the root mean square error(RMSE)is 0.0171K.