Research on the construction of weaponry indicator system and intelligent evaluation methods

Area of Science:

• Engineering
• Computer Science
• Defense Technology

Background:

• Traditional weapon and equipment index systems suffer from subjective interference and low construction efficiency.
• Equipment test video data offers a rich source for objective and efficient assessment.

Purpose of the Study:

• To develop an intelligent index system construction and assessment method using target detection on equipment test video data.
• To improve the efficiency, objectivity, and stability of combat effectiveness assessment for weapon systems.

Main Methods:

• An optimized IPSO-BP network model with dynamic weights and adaptive learning factors was developed.
• An improved DS (Dempster-Shafer) evidence-parallel neural network assessment method was proposed, enhancing numerical stability.
• An integrated CNN (Convolutional Neural Network) based multi-view feature integration assessment model was constructed using improved residual network blocks.

Main Results:

• The proposed methods demonstrated efficient and intelligent construction of the indicator system.
• The intelligent assessment methods showed improved stability and accuracy compared to existing approaches.
• The study successfully enriched the evaluation of indicator data through advanced computational techniques.

Conclusions:

• The target detection-based approach significantly reduces subjective interference in index system construction.
• The developed intelligent assessment models offer a more stable and efficient alternative for evaluating weapon and equipment combat effectiveness.
• This research provides a valuable framework for advancing defense technology assessment through AI and machine learning.