22results about How to "Enhanced confrontation ability" patented technology

The invention relates to the technical field of artificial intelligence and unmanned systems, and discloses an intelligent target allocation method and system based on deep reinforcement learning, and the method comprises the steps: collecting and classifying a security place state, a firepower unit state, an enemy target state and a historical allocation scheme through multi-element channels, and obtaining battlefield information; integrating various battlefield information into a dictionary in a key-value pair form, and taking a dictionary observation space form as situation input; performing target distribution on the input information through a decision action calculation model, and generating a decision instruction; converting the decision instructions into actions which can be executed in a simulation environment, executing the decision instructions, simulating an actual combat process, observing and recording action effects, evaluating the decision instructions according to a simulation result, generating evaluation feedback, adjusting decision actions according to the evaluation feedback, calculating model parameters, and performing iterative optimization; the intelligent level and the confrontation efficiency of the confrontation system are obviously enhanced.

The invention relates to the technical field of underwater robot formation, in particular to an underwater unmanned underwater vehicle cluster defense confrontation method, which utilizes an underwater unmanned cluster to make up the defect of insufficient functions of a single platform by means of the number advantage, and aggregates various underwater platforms with different functions to form a common body capable of completing a given task together; a traditional expensive large manned platform is replaced by a large number of small-size low-cost distributed unmanned platforms, and the distributed unmanned platform has the advantages of being high in efficiency, low in cost, reconfigurable and capable of executing multiple tasks, is used for underwater defense and resistance, and can greatly increase the underwater reconnaissance search range and improve the response speed and the confrontation effect.

The present disclosure relates to the field of artificial intelligence and cluster confrontation technology, and provides a training method and system for confrontation clusters, and a confrontation method and system for confrontation clusters. The training method includes: storing confrontation data in an online strategy sample memory in real time; whenever the amount of data in the online strategy sample memory exceeds a preset first threshold, transferring the data to an offline strategy sample memory; filtering out multiple data with the best confrontation indicators of an intelligent agent from the offline strategy sample memory to obtain filtered data; determining a policy gradient based on the current data and the filtered data in the online strategy sample memory; and updating the neural network parameters of the intelligent agent in the confrontation cluster based on the policy gradient to train the confrontation cluster. The present disclosure can solve the problems of weak confrontation ability of clusters, solidification of confrontation strategies and poor adaptability, can enhance the confrontation ability of clusters, improve the continuous learning and evolution of clusters, and improve the flexibility and adaptability of confrontation strategies.

The application relates to the technical field of artificial intelligence, and provides a method and device for training a battle strategy model of a multi-player battle game, a computer device, a storage medium and a computer program product. The method comprises the following steps: obtaining each historical state from battle data corresponding to a virtual object lineup for each virtual object lineup, determining a game target corresponding to each historical state, determining a battle strategy type, training a game target prediction model based on the historical state, the battle strategy type and the reward data, inputting a response state corresponding to each battle action of each virtual object lineup based on an initial battle strategy model and a battle strategy type corresponding to the corresponding virtual object lineup into the game target prediction model to obtain a predicted game target, and training the initial battle strategy model based on reward data corresponding to each predicted game target to obtain a battle strategy model, so that the confrontation ability of the virtual object lineup can be effectively improved.

The application relates to the technical field of artificial intelligence, and provides a method and device for training a battle strategy model of a multi-player battle game, a computer device, a storage medium and a computer program product. The method comprises the following steps: obtaining each historical state from battle data corresponding to a virtual object lineup for each virtual object lineup, determining a game target corresponding to each historical state, determining a battle strategy type, training a game target prediction model based on the historical state, the battle strategy type and the reward data, inputting a response state corresponding to each battle action of each virtual object lineup based on an initial battle strategy model and a battle strategy type corresponding to the corresponding virtual object lineup into the game target prediction model to obtain a predicted game target, and training the initial battle strategy model based on reward data corresponding to each predicted game target to obtain a battle strategy model, so that the confrontation ability of the virtual object lineup can be effectively improved.

This disclosure provides an interpretable air combat decision-making method and apparatus. The method includes: acquiring current situation information from an air combat environment; wherein, the air combat environment contains a first fighter jet and a second fighter jet hostile to the first fighter jet; determining, based on the current situation information, rules corresponding to the current situation information from a rule base corresponding to the first fighter jet, and constructing a matching set corresponding to the first fighter jet based on the rules; obtaining the predicted behavior of the second fighter jet through a behavior prediction model of the second fighter jet, determining alternative maneuvers for the first fighter jet based on the predicted behavior and the matching set, and constructing a set of maneuvers corresponding to the first fighter jet based on the alternative maneuvers and the matching set; and controlling the first fighter jet to execute the maneuvers corresponding to the set of maneuvers based on the set of maneuvers. By applying this disclosure, not only can the adversarial performance of air combat decision-making be improved, but also air combat decision-making rules with generalizability and interpretability can be obtained.

The invention relates to a low-detectability active and passive composite blinding system for an urban combat environment, and belongs to the technical field of electronic countermeasures. Comprising an active blinding module for emitting a directional energy beam to cause saturation, interference or permanent damage to an enemy optical, infrared or laser sensor; the passive blinding module is used under the condition that the electromagnetic energy is not actively emitted; the low detectability platform is used for bearing the active blinding module and the passive blinding module; and a cooperative control unit. According to the invention, active blinding, passive blinding and low detectability technologies are unprecedentedly deeply fused, and unified scheduling is carried out by an intelligent cooperative control unit, so that the risk of detection and countering by enemies is reduced to the greatest extent, and the characteristics of short sight distance, complex background and diversified threat sources of urban combat are fully considered. Active accurate blinding and passive hidden interference can be effectively combined, and meanwhile the method has the advantages of being high in survivability, high in confrontation efficiency and high in tactical flexibility.

The invention discloses an adaptive drilling Agent system and method based on a large language model and a storage medium, and the system comprises an instruction receiving and contextualization module which is used for receiving an unstructured natural language instruction of a performer and generating structured prompt information; the large language model reasoning core is used for outputting semantic reply oriented to participants and a structured action sequence for simulation execution based on the structured prompt information; the simulation and constraint engine is used for performing physical feasibility verification and logic / authority verification on the structured action sequence based on a physical evolution model and an environment logic rule and outputting a verification result; and the action verification and feedback module is used for generating a structured failure message when the verification result is failure and recharging the structured failure message to the big language model reasoning core as a hard constraint to trigger the regeneration of an alternative action sequence. And when the verification result is that the structured action sequence passes the verification, sending the structured action sequence passing the verification to a simulation and constraint engine for execution, and outputting an execution result.

The invention relates to a multi-agent target allocation method suitable for a complex task scene, and belongs to the technical field of agent clusters. The method comprises the following steps: in a complex task scene, defining an agent cluster target allocation behavior, optimizing a target, establishing a mathematical model of an agent cluster target allocation problem, and defining a constraint; splitting an agent cluster target allocation problem into a multi-agent behavior design problem and a task allocation problem; the method comprises the following steps: aiming at a multi-agent behavior design problem, orderly segmenting a scene environment, and respectively searching an approximately optimal path in each segmented sub-region; aiming at a task distribution problem, target distribution gains brought by different task distribution strategies are estimated, and a task combination generating the maximum target distribution gain is selected for task distribution. According to the strategy method provided by the invention, the approximately optimal target distribution gain can be obtained, and the balance between the target distribution effect and the calculation efficiency can be effectively achieved.

The present application relates to an image processing method, device, computer equipment and storage medium. The method includes: obtaining a target image of a target image category and a current adversarial image corresponding to an adversarial image category; performing feature extraction on the target image to obtain a target extracted feature, and obtaining a reference extracted feature based on the target extracted feature; performing feature extraction on the current adversarial image to obtain an adversarial extracted feature; calculating a feature difference value between the adversarial extracted feature and the reference extracted feature, and obtaining a target image loss value based on the feature difference value; obtaining an interference adjustment value based on the target image loss value, and adjusting the pixel value of the current adversarial image based on the interference adjustment value to obtain a target adversarial image corresponding to the target image. This method can improve the adversarial effect of adversarial samples. The image recognition model in the present application can be a neural network model based on artificial intelligence. The present solution is applied to the field of artificial intelligence to improve the reliability of the neural network model.

This invention belongs to the field of artificial intelligence algorithm technology and discloses an intelligent transfer adversarial method and system for electromagnetic signal recognition under incomplete information. First, multiple models of alternative target networks with incomplete information are constructed in parallel, and these alternative models are trained using an evolutionary strategy based on a prior knowledge matrix. Then, an ensemble network model is generated based on the trained alternative models, and a fully connected layer and a layer with feature-level weighted fusion are designed. Finally, an optimized adversarial method for electromagnetic spatial signal recognition tasks is designed, generating adversarial samples under the ensemble model, and transferring these adversarial samples to the target network to achieve adversarial capabilities against the target network. This invention can effectively achieve adversarial capabilities against the target network when the target network information is unknown, and it still exhibits good transfer adversarial performance under the same signal recognition task and different unknown target network conditions.

The invention provides an industrial control equipment-oriented flow trip line generation method and system based on honey points, and relates to the technical field of network security. The method provided by the invention comprises the following steps: generating an instruction based on a configuration file and a tripwire strategy; obtaining an original load sequence, identifying sensitive fields in the original load sequence based on a flow trip line load generation model, and replacing the sensitive fields with mask marks to form a mask load sequence; predicting a mask mark position in the mask load sequence to generate bait data; filling the bait data to a mask mark position, and combining the bait data with a non-sensitive field in the mask load sequence to generate camouflage load data; embedding the camouflage load data into a standard protocol frame based on the instruction, constructing a flow trip line data packet, and embedding the address information of the honey point into the flow trip line data packet; and when the generation time arrives, actively sending the traffic trip line data packet to the outside according to the sending interval. According to the invention, active trapping perception and dynamic cheating of potential attacks are realized by actively sending analog data traffic.

The invention discloses an impact-resistant PE water supply pipe and a preparation method thereof, and relates to the technical field of PE water supply pipes. Comprising an outer layer assembly, an inner layer assembly and a filling assembly fixed between the outer layer assembly and the inner layer assembly, the filling assembly comprises a composite layer, a plurality of supporting structures are arranged on the composite layer, the composite layer is obtained through a first method, and then the filling assembly is obtained. On the basis of the inner-layer PE pipe and the outer-layer PE pipe, the filling layer is designed inside, the outer surface of the inner-layer assembly is subjected to hot melting treatment, the outer wall of the inner-layer assembly is wrapped with the prepared filling assembly, and due to the fact that the outer wall of the inner-layer assembly is in the hot melting state, the filling assembly and the inner-layer assembly have good adhesiveness; the outer layer assembly in the plastic state is wrapped on the surface of the hardened filling assembly, thermalization treatment is carried out, the outer layer assembly is shrunk, the outer layer assembly, the inner layer assembly and the filling assembly form a whole, and the impact resistance effect is improved.

The application provides an adversarial training method, device and equipment based on adaptive group sample perturbation constraint, which comprises the following steps: inputting a training image into an initial network model to obtain a network output vector corresponding to the training image and a predicted category; if it is determined that the classification result of the training image is wrong based on the predicted category and an actual category of the training image, the training image is determined as a natural sample image; if it is determined that the classification result of the training image is correct based on the predicted category and the actual category, a target adaptive group corresponding to the training image is determined based on the network output vector; a target perturbation vector corresponding to the training image is determined based on a target sample perturbation constraint corresponding to the target adaptive group, and a perturbation sample image is generated based on the target perturbation vector and the training image; and the initial network model is trained based on the natural sample image and the perturbation sample image to obtain a target network model. Through the technical scheme of the application, the anti-interference ability of the target network model to attack samples is significantly improved.

The utility model discloses a driving hammer type fighting toy car, which comprises a push shovel, a car body, an attack hammer, a driving hammer power device and a magnetic attraction armor, the push shovel is arranged at the front end of the car body, the driving hammer power device is arranged on the car head of the car body, and the end part of a hammer handle of the attack hammer is arranged on the driving hammer power device. The attack hammer and the corresponding driving hammer power device are added on the basis of an original push shovel, and the attack hammer is driven by the driving hammer power device to turn over, withdraw or swing, so that the toy car has two attack modes of shoveling and hammering, the toy car is more aggressive in battle, playability of the toy car is remarkably improved, and the toy car is worthy of popularization and application. The interactivity and the confrontation among the toy cars are also increased, players are attracted to put more energy to participate in fight, and the interestingness and the irritation in wartime are greatly improved; and when the impact force of the magnetic armor is large enough, a magnetic armor falling mechanism can be triggered, so that an intuitive and clear victory or defeat judgment standard is provided for fight, and joyful feeling and fight passion during playing can be improved.

The application provides an underwater unmanned underwater vehicle formation dynamic area search system and relates to the field of unmanned underwater vehicle cooperative operation, and comprises the following steps: generating a virtual marine environment according to seabed depth information and land obstacle distribution information; arranging virtual particles in a preset search area to form a gravitational field, and giving the virtual particles an initial mass; constructing a detection pressure field according to the detection activities of the unmanned underwater vehicle formation, determining the probability distribution of the target appearance position based on the reverse distribution of the detection pressure field, and generating a counter target according to the probability distribution; dynamically updating the mass value of the virtual particles according to the detection coverage range of the unmanned underwater vehicle formation; and calculating the gravity and repulsion that each unmanned underwater vehicle in the unmanned underwater vehicle formation receives according to the mass distribution of the virtual particles in the gravitational field, and updating the position and heading of each unmanned underwater vehicle based on the gravity and repulsion. The application improves the efficiency and authenticity of dynamic target search.

The present invention discloses a traffic shaping method and system based on deep learning and adversarial strategies. The shaping method includes: monitoring network traffic in real time through a traffic monitor, capturing information of traffic data, and performing data preprocessing, extracting traffic features through the preprocessed traffic data, and training a DNN classifier; inputting the trained DNN classifier and traffic data into a patch generation module, preliminarily generating adversarial patches according to a dual simulated annealing algorithm with a multi-objective optimization algorithm, and optimizing the adversarial patches through an adaptive adjustment strategy and a bandwidth overhead adjustment strategy; patch injection: injecting the optimized adversarial patch into the original traffic data through an injection controller to shape the original traffic data. The present invention solves the problem that the adversarial capability and privacy security protection capability of IoT devices in the face of device identification and traffic analysis attacks need to be strengthened.

The invention belongs to the technical field of aircraft underwater launching and carrying, and particularly relates to an aircraft launching device. An underwater multi-shot type aircraft launching device comprises four cabin sections and a fixing device. An aircraft launching cylinder is arranged in the first cabin section, and a bottom channel of the aircraft launching cylinder is opened and closed through an electromagnetic valve A; the second cabin section is a pressure chamber, and the airbag is mounted on the outer side of the second cabin section and communicated with an inner chamber of the second cabin section through a gas transmission pipeline; a through hole for conveying gas into the second cabin section is formed in the top of the third cabin section; a high-pressure air pump A is arranged in the third cabin section; a high-pressure air pump B is arranged in the fourth cabin section; exhaust passages are arranged at the lower part of the high-pressure air pump B, and an electromagnetic valve E for controlling a switch is arranged on each exhaust passage; the high-pressure gas tank supplies gas to the high-pressure gas pump A and the high-pressure gas pump The fixing device is used for fixing the fourth cabin section in an initial state; and after receiving the signal, releasing the fixation. The unmanned aerial vehicle has the capability of launching multiple aircrafts and is high in stability.

The invention discloses a patient leg postoperative intelligent personalized rehabilitation training method based on the Internet, and the method comprises the steps: data collection, which is used for collecting patient condition information and leg rehabilitation training data uploaded by a patient through the Internet; the data preprocessing module is used for performing data preprocessing on the patient training detection original data; patient rehabilitation training feature extraction, which is used for extracting patient rehabilitation training features; the incorrect training action detection is used for intelligently detecting incorrect training actions of the patient; the patient rehabilitation training guidance correction module is used for generating guidance correction indexes and generating patient leg personalized rehabilitation training grade suggestion schemes in combination with the medical knowledge graph; in order to avoid the limitation of time, geographic position and leg walking inconvenience of the patient, the Internet technology and deep learning image analysis are introduced, the doctor remote leg training recovery doctor seeing guidance efficiency is improved, the on-site follow-up diagnosis cost of the patient is saved, and the treatment enthusiasm of the patient and the leg rehabilitation speed are improved.

The invention discloses an adversarial sample generation method and device for modulation recognition in a real scene, and the method comprises the steps: constructing a modulation signal recognition model which simulates a recognition device of a modulation signal; a disclosed modulation signal data set RML2016.10b is adopted to train the modulation signal identification model, and a high-precision modulation signal identification model is obtained; constructing a physical confrontation sample generation model used for generating pulse type disturbance; training an adversarial sample generation model by adopting the constructed modulation signal recognition model and the disclosed modulation signal data set RML2016.10b to obtain an effective adversarial sample generation method; and inputting the normal signal sample into an adversarial sample generation model to obtain an adversarial sample with an attack effect. The confrontation sample with a physical attack effect is generated, and modulation signal recognition model confrontation in a real environment is realized.

The present specification provides a script detection method and device, which are applied to a security detection engine that adopts an inference framework. The security detection engine is pre-loaded with an inference strategy library; the inference strategy library contains several inference strategies for inferring the adversarial means used by the adversarial statements used to counter the detection contained in the script to be detected. The method includes: in the process of running each statement contained in the script to be detected in a simulated running environment, determining whether the target statement to be executed is a suspected adversarial statement. If the target statement is a suspected adversarial statement, the inference strategy in the inference strategy library is executed to infer the adversarial means used by the target statement, and obtain adversarial information used to describe the adversarial means used by the target statement. The target statement is corrected based on the adversarial information, and the corrected target statement is executed, and whether the target statement is an adversarial statement is further determined based on the execution result.

The invention discloses a text adversarial attack method based on antonym replacement, which comprises the following steps of: 1) performing word segmentation on an original text through LAC, HanLP and LTP word segmentation and part-of-speech tagging tools, and tagging adjectives in the original text; 2) further screening unlabeled texts by using OpenHowNet, generating a vocabulary library related to positive or negative adjectives, and labeling words of which the similarity is greater than a threshold value; 3) sorting the words according to the importance scores by adopting a search algorithm based on a greedy algorithm and word importance sorting, and determining a disturbance sequence of each word in the text; and 4) based on screening and manual labeling results, constructing a complete antonym library, and disturbing labeled adjectives to generate a high-similarity and readable adversarial text. According to the method, based on an antonym replacement strategy, text perturbation is performed through a special structure, so that the generated adversarial text keeps the semantic similarity of the original text, and judgment of the sentiment classification model can be effectively interfered.