Abstract: To solve the problem that the feature weight cannot be automatically selected in the sampling algorithm of filament winding doffing point trajectory based on the feature function of the spatial feature curve, a bi objective optimization model is established, which takes the feature weight as a variable and the MAE and RMSE of the linear interpolation generated curve and the original curve of the sampling point obtained as the objective function. A bi objective optimization method based on improved NSGA-Ⅱ algorithm is proposed to optimize the feature weight. The example verification shows that the MAE and RMSE of the Pareto solution set obtained by the improved NSGA-Ⅱ algorithm are reduced by 0.002 and 0.105 on average compared with the traditional NSGA-Ⅱ algorithm, the MAE and RMSE of the feature weight selected by the algorithm are reduced by 12.9% and 8.5% respectively when the feature weight is (0.1, 0.3), and the bit feature weight is reduced by 20.6% and 11.4% respectively when the feature weight is (0.9, 0.1), effectively improving the accuracy of the doffing point trajectory sampling.

Key words: doffing point trajectory sampling, characteristic function of space curve, NSGA-Ⅱ, composites