In order to calibrate the magnetic measurement system used in guided munition on site, a two-stage calibration (TSC) scheme without reference is proposed in this paper. Analyzing the interfering magnetic field in the projectile and misalignment angles between the projectile coordinate system and measurement coordinate system establishes a proper mathematical equivalent model and derives a calibration method. The first stage is ellipsoid fitting to obtain the equivalent zero-offset, equivalent sensitivity and equivalent non-orthogonal angles of the sensor; the second stage is to calibrate the misalignment angles between the projectile coordinate system and the measurement coordinate system with the three-position calibration (TPC) method. Complete calibration is convenient to operate and does not need an additional reference, which has wide applicability. The simulation results show that the deviation in the measured value after compensation is within 100 nT. The experiment proves that the error of compensated magnetic value is about 150 nT, which meets the accuracy of requirements in guided munitions.

The dynamic evolution value assessment of maritime target groups is a complex problem. A key challenge lies in converting physical data into quantifiable indicators and dynamically adjusting weights based on situational changes. These capabilities critically influence assessment accuracy and subsequent decision-making. To accurately assess the dynamic evolution values, an assessment scheme for maritime target groups based on the attention mechanism and hybrid quantification methods is proposed. First, based on analyzing the equipment and characteristics of various types of targets, an indicator system with self-generating capabilities is constructed. Secondly, based on real-time data, quantitative methods for various indicators are proposed. Finally, the attention mechanism is introduced into the fully connected neural network and an intelligent weight learning model with adaptive capability is proposed. The results of the simulation example show that the hybrid quantification methods can achieve accurate quantification of indicators based on real-time information, and the weight learning model can achieve accurate weight distribution under different situations. The proposed assessment scheme is feasible for the assessment of the dynamic evolution value of the maritime target group. It provides an intelligent scheme for the assessment of the dynamic value of the target group.

Accurate assessment of member importance is essential for understanding contribution differences in complex group systems. However, existing methods often depend on predefined internal structures or relational information, making them unsuitable for scenarios where internal organization is unknown and only observable information is available. This paper proposes a Dynamic Contribution Assessment Model (DCAM) for importance assessment based on observable information. A multi-dimensional feature representation framework is first constructed by considering individual state characteristics, group interaction relationships, and behavioral dynamic evolution characteristics. Then, a dynamic contribution assessment mechanism is developed by integrating indicator-level information contributions and adaptive feature-layer contributions to evaluate the importance of heterogeneous members. A dynamic complex group simulation scenario is established to validate the proposed method. The results show that the proposed approach can effectively distinguish importance differences among members using only external observable information, and the obtained assessment results are consistent with the observed motion evolution and interaction characteristics of the group. The proposed method provides an effective approach for member importance analysis in complex group systems with unknown internal structures.