Flight attack defense system

Abstract

This invention provides a flight attack defense system, comprising: a radar detection module that detects a flying target and generates flight information of the target; a visual acquisition module that captures images of the flying target and generates images of the target; a central processing unit that receives the flight information and images of the target; a trajectory prediction module that receives the flight information and generates first flight trajectory data; an image recognition module that receives the images of the target and generates target shape data; a central processing unit that receives the target shape data and obtains target feature data; a trajectory generation module that receives the target feature data and the first flight trajectory data and generates second flight trajectory data; an interception calculation module that receives the second flight trajectory data and generates a gun control command and a gun order command; and a defense platform that, according to the gun control command, adjusts the firing attitude and sets the detonation mode of at least one electronically controlled fuse according to the gun order command, thereby controlling the missile's flight direction and detonation time to automatically form an interception network defense, thereby reducing defense costs and reducing manual judgment to improve judgment accuracy.

Claims

1. A flight attack defense system, comprising: A radar detection module is used to detect the position and speed of a flying target and generate flight information of the target. The system includes a target acquisition module for tracking and capturing images of flying targets and generating target images; and a central processing unit (CPU) for receiving target flight information and target images. The CPU includes a CPU module, a storage module, a trajectory prediction module, an image recognition module, a trajectory generation module, and an interception calculation module. The storage module stores multiple target feature data. The CPU receives target flight information and transmits it to the trajectory prediction module. The trajectory prediction module processes the target flight information to generate first flight trajectory data, which is then transmitted from the CPU module to the target acquisition module. The target acquisition module captures images of the flying target based on the first flight trajectory data and generates... An image of a target object is transmitted to a central processing unit. The central processing unit then transmits the target object image to an image recognition module. The image recognition module generates target object shape data from the target object image. The central processing unit receives the target object shape data and extracts corresponding target object feature data from its storage module. The trajectory generation module receives the extracted target object feature data and first flight trajectory data and generates second flight trajectory data, which is then transmitted to the central processing unit. The interception calculation module receives the second flight trajectory data and generates a gun control command and a gun order command, which are sent to at least one defensive platform and an electronically controlled fuse. The defensive platform adjusts its firing attitude according to the gun control command, and the electronically controlled fuse sets its detonation mode according to the gun order command.

2. The flight attack defense system as described in claim 1, wherein the defense platform may be a mobile platform.

3. The flight attack defense system as described in claim 1, wherein the trajectory prediction module may include a machine learning model or a deep learning model to predict the future trajectory of the flight target based on the target's flight information. The machine learning model may be a temporal prediction model, such as a recurrent neural network, a long short-term memory network, a gated recurrent unit, a Transformer architecture or an equivalent model thereof, or a model architecture used for trajectory prediction / motion prediction (such as the model used in Trajectory Prediction or Motion Forecasting tasks).

4. The flight attack defense system as described in claim 1, wherein the image recognition module may employ a convolutional neural network (CNN) or other deep learning image classification / recognition model to classify and identify the target image to generate target morphology data.

5. The flight attack defense system as described in claim 1, wherein the trajectory generation module can utilize a machine learning model / deep learning model, taking the first flight trajectory data and the target feature data as input, and outputting the second flight trajectory data. The model can be a conditional sequence model or a probabilistic multimodal trajectory model, such as, but not limited to: a conditional prediction model of a recurrent neural network (RNN / LSTM / GRU), a sequence-to-sequence (Seq2Seq) prediction model of a Transformer, a Mixture Density Network (MDN) or its equivalent probabilistic model, a variational architecture (VAE), or a diffusion-based architecture. The multimodal trajectory generation model implemented by forecasting can be further explained by the fact that the model can be trained using historical motion data. The data may include: historical trajectory data measured by sensors, synthetic trajectory data generated by simulation systems, and motion constraints / morphological features established by public specifications or test data. Supervised learning, semi-supervised or hybrid training methods can be adopted to improve the model's generalization ability to different target morphologies or environmental conditions.

6. The flight attack defense system as described in claim 1, wherein the explosion mode set by the artillery command is selected from the group consisting of impact explosion, delayed, instantaneous, proximity, airburst, self-destruct, in-flight selection, and electronic security.

7. The flight attack defense system as described in claim 1, wherein the defense platform is selected from a rocket launcher system or a mortar system, and the electronically controlled fuze is a modular programmable rocket fuze or mortar fuze.

8. The flight attack defense system as described in claim 1, wherein the interception operation module may include a decision model, which may be implemented by a hybrid architecture of deep neural networks, gradient boosting trees, or rule-based / optimization modules and machine learning modules, and may be constructed through reinforcement learning, supervised learning, imitation learning or equivalent methods.

9. The flight attack defense system as described in claim 1 further includes an intercepting drone connected to the central processing unit, the central processing unit being configured to determine the relative distance between the flight target and one of the defense platforms, and when the relative distance is less than a danger line distance threshold, the central processing unit generates a control command containing a tracking path information to the intercepting drone to drive the intercepting drone to intercept the flight target.

Images