Objective  The highlight model method is widely used for predicting the acoustic scattering characteristics of underwater targets due to its high computational efficiency. However, its accuracy is limited when dealing with complex geometries and acoustic coatings. To improve prediction accuracy while maintaining computational speed, this paper proposes a modified highlight model method that incorporates equivalent reflection coefficients and a genetic algorithm (GA) for rapid and accurate target strength (TS) prediction of complex underwater vehicles.

Method The proposed method first decomposes the complex target into several geometric components, each assigned an equivalent reflection coefficient to construct a modified highlight model. A genetic algorithm is then employed to invert these reflection coefficients by minimizing the root mean square error (RMSE) between the model predictions and reference results obtained from the planar element method (PEM). The total target strength is obtained through coherent summation of the scattering contributions from all components. Finally, the model is validated using two benchmark structures: a double-hull and a single-double hybrid hull model, at frequencies of 1 kHz, 5 kHz, and 10 kHz.

Results  Compared with the PEM, the modified highlight model shows good agreement in TS predictions across all tested frequencies. The RMSE and the relative average error is kept below 4 dB. In terms of computational efficiency, the modified method achieves a speed-up of approximately 1 000 times, with calculation times at the second level, making it highly suitable for real-time applications.

Conclusion Studies show that the modified highlight model method can significantly improve the prediction accuracy of the acoustic scattering characteristics of complex underwater targets while maintaining computational efficiency; the proposed method can provide a new technical approach for the acoustic stealth design of complex underwater targets and the rapid prediction of target characteristics.