Search

Machine learning (ML) is well suited for the prediction of high-complexity, high-dimensional problems such as those encountered in terminal ballistics. We evaluate the performance of four popular ML-based regression models, extreme gradient boosting (XGBoost), artificial neural network (ANN), support vector regression (SVR), and Gaussian process regression (GP), on two common terminal ballistics' problems: (a) predicting the V50 ballistic limit of monolithic metallic armour impacted by small and medium calibre projectiles and fragments, and (b) predicting the depth to which a projectile will penetrate a target of semi-infinite thickness. To achieve this we utilise two datasets, each consisting of approximately 1000 samples, collated from public release sources. We demonstrate that all four model types provide similarly excellent agreement when interpolating within the training data and diverge when extrapolating outside this range. Although extrapolation is not advisable for ML-based regression models, for applications such as lethality/survivability analysis, such capability is required. To circumvent this, we implement expert knowledge and physics-based models via enforced monotonicity, as a Gaussian prior mean, and through a modified loss function. The physics-informed models demonstrate improved performance over both classical physics-based models and the basic ML regression models, providing an ability to accurately fit experimental data when it is available and then revert to the physics-based model when not. The resulting models demonstrate high levels of predictive accuracy over a very wide range of projectile types, target materials and thicknesses, and impact conditions significantly more diverse than that achievable from any existing analytical approach. Compared with numerical analysis tools such as finite element solvers the ML models run orders of magnitude faster. We provide some general guidelines throughout for the development, application, and reporting of ML models in terminal ballistics problems. [ABSTRACT FROM AUTHOR]

In this paper, an intelligent control method applying on numerical virtual flight is proposed. The proposed algorithm is verified and evaluated by combining with the case of the basic finner projectile model and shows a good application prospect. Firstly, a numerical virtual flight simulation model based on overlapping dynamic mesh technology is constructed. In order to verify the accuracy of the dynamic grid technology and the calculation of unsteady flow, a numerical simulation of the basic finner projectile without control is carried out. The simulation results are in good agreement with the experiment data which shows that the algorithm used in this paper can also be used in the design and evaluation of the intelligent controller in the numerical virtual flight simulation. Secondly, combined with the real-time control requirements of aerodynamic, attitude and displacement parameters of the projectile during the flight process, the numerical simulations of the basic finner projectile’s pitch channel are carried out under the traditional PID(Proportional-Integral-Derivative) control strategy and the intelligent PID control strategy respectively. The intelligent PID controller based on BP(Back Propagation) neural network can realize online learning and self-optimization of control parameters according to the acquired real-time flight parameters. Compared with the traditional PID controller, the concerned control variable overshoot, rise time, transition time and steady state error and other performance indicators have been greatly improved, and the higher the learning efficiency or the inertia coefficient, the faster the system, the larger the overshoot, and the smaller the stability error. The intelligent control method applying on numerical virtual flight is capable of solving the complicated unsteady motion and flow with the intelligent PID control strategy and has a strong promotion to engineering application. [ABSTRACT FROM AUTHOR]