Pressure sensors are extensively applied for monitoring tasks under extreme working conditions such as aerospace due to their characteristics of miniaturization

low cost and high precision.However

the output is greatly affected by temperature

owing to the temperature sensitivity of the piezoresistive coefficient and the bridge resistance of the semiconductor materials which consequently restricts measurement accuracy under hightemperature conditions.To address this issue

a software compensation method based on a radial basis function(RBF)neural network optimized by the sparrow search algorithm(SSA)is proposed.This method optimizes the center points of the RBF in the RBF neural network to overcome the shortcomings of traditional clustering approaches such as K-means

which are sensitive to initial values and prone to local optima

thereby improving the convergence quality and generalization performance of the network.Experimental results demonstrate that over a wide temperature range from room temperature to 350℃

the SSA-RBF reduces the full-scale error to 0.25% FS.Without increasing hardware complexity

this method effectively improves the output consistency of the sensor at high temperatures

providing a feasible software compensation pathway for its industrial applications.