6results about How to "Reasonable decision" patented technology

This invention relates to the field of road construction safety monitoring technology, specifically to an intelligent identification and early warning method for road construction areas from a low-altitude perspective. This method aims to address the technical problem that existing technologies cannot accurately identify risks in construction areas when visual perception is degraded, such as in the presence of blurred images. The method includes: collecting multimodal data of the construction area using multiple sensors mounted on a drone; real-time evaluation of visible light image clarity, triggering a multimodal risk inference mode when the clarity falls below a preset threshold; extracting residual visual and non-visual features, inputting them into a risk inference model that integrates a construction safety knowledge graph and a probabilistic graphical network, and inferring scene completion results and their posterior probabilities; calculating the current comprehensive risk probability, generating early warning information and pushing it to the management terminal when the probability exceeds a threshold. This invention improves the environmental adaptability and robustness of road construction inspection and early warning systems, enabling all-weather, all-scenario construction safety management.

The application discloses a kind of brain-like reinforcement learning method and system based on layered experience playback, it is related to reinforcement learning and brain-like computing technical field.The method comprises the following steps: S1, collecting observation data and pre-processing;S2, initialize experience buffer pool and actor network, critic network and corresponding target network, parameter initialization;S3, initialize exploration noise and select action from actor network according to current state and execute, store the obtained experience sample to experience buffer pool;S4, obtain new sample from experience buffer pool, carry out short-term memory pool update;S5, use attention discrimination module to determine whether part of experience in short-term memory experience pool is transferred to long-term memory experience pool;S6, the parameters of actor network, critic network and corresponding target network are updated.The application uses the above method, improves the experience utilization rate of intelligent agent, improves the performance of reinforcement learning, and has wide application potential in many fields.

The application discloses a kind of based on artificial intelligence's consumer goods harmful substance intelligent screening method, it is related to artificial intelligence technical field, it is difficult to effectively identify and analyze the technical problem that a large number of natural language, commercial name or commonly known as etc. non-standardized description exists in ingredient list, it is also difficult to accurately convert these fuzzy text information into the standardized terminology that can be accurately matched by regulation database;Including the following steps: the fuzzy ingredient list containing natural language or commercial name description is preprocessed by multi-task text regularization model;Then, utilize embedding semantic similarity model and the multi-evidence fusion presumption unit of improved attention mechanism, fuzzy component is intelligently analyzed and matched, and standard substance identification and confidence are output;Finally, through multi-agent risk decision framework combines real-time regulation database, dynamic compliance screening and risk assessment are carried out;The application improves the intelligent level of consumer goods safety compliance management.

The application provides a bus side knife switch operation risk assessment method and device, and relates to the technical field of electric power engineering. The method comprises the following steps: if the knife switch in the power system is a bus side knife switch, the wiring type of the power system is obtained; then, according to the wiring type, the operation type corresponding to the wiring type is determined; then, according to the wiring type and the operation type, the simulation fault equipment of the power system is determined; finally, according to the simulation fault equipment, the power grid accident risk level of the power system is determined. Through the method, the accuracy of the risk assessment of the bus side knife switch operation can be improved.

The disclosed vehicle autonomous car-following control method for uneven road surface extracts car-following events from unstructured road driving data, constructs a multi-modal state space containing vehicle kinematic state and road geometric and physical characteristics, and a two-dimensional action space of continuous acceleration and steering coordination; designs a road surface adaptive reward function; uses a double experience replay buffer and target network soft update for training, and deploys the obtained road surface perception strategy model to an unstructured road automatic driving vehicle to output continuous acceleration and steering control instructions. The invention embeds road geometric and physical properties into the reinforcement learning state space by introducing a road feature encoder and designs a road surface adaptive reward weight, which alleviates the model mismatch problem of traditional car-following methods on low adhesion and high undulating uneven road surfaces, and improves the car-following efficiency and comfort under the premise of zero collision.

The application relates to a time window graph model-based timing circuit critical node positioning method and equipment, which comprises the following steps: 1) analyzing a circuit netlist, initializing a criticality calculation matrix according to a circuit node type, wherein the criticality calculation matrix comprises a single fault calculation matrix and a multiple fault calculation matrix; 2) based on a time window graph model, aggregating circuit node states through circuit simulation simulation to obtain circuit node slices; 3) calculating the criticality of all nodes in the slices to a target node based on probability and the criticality calculation matrix; and 4) obtaining critical nodes of the timing circuit according to node criticality sorting. Compared with the prior art, the application has the advantages of small time and space overhead, and the result obtained based on the method has greater reliability improvement for reinforcement of the timing circuit.