4results about How to "Improve accurate judgment" patented technology

The invention provides an engine combustion mode switching method and device, a vehicle and a storage medium, and belongs to the technical field of tail gas aftertreatment, engine combustion modes comprise a normal combustion mode and a low urea consumption mode, and the urea consumption of the low urea consumption mode is smaller than that of the normal combustion mode; the method comprises the steps that the urea-fuel ratio and the nitrogen oxide specific emission within the preset duration are determined; when the urine-fuel ratio is larger than a urine-fuel ratio threshold value and the nitrogen oxide specific emission amount is larger than a specific emission amount threshold value, the duration when the urine-fuel ratio is larger than the urine-fuel ratio threshold value and the nitrogen oxide specific emission amount is larger than the specific emission amount threshold value is obtained; and when the duration is longer than the calibrated duration, the combustion mode of the engine is switched from a normal combustion mode to a low urea consumption mode. According to the invention, the consumption of urea is reduced while the fuel economy is considered.

This invention discloses a torque and durability testing system for bolts used in new energy vehicles. The system includes a clamping module for clamping and testing sample bolts. The clamping module includes a clamping tool comprising a main board and a sub-board. It also includes a durability fatigue testing module for simulating the environmental conditions encountered by the battery pack during actual operation of new energy vehicles, thereby improving testing accuracy. The durability fatigue testing module includes a vibration unit, a heating unit, a resonant unit, and a problem collection unit. This solution improves the accuracy of bolt torque and durability testing by simulating the environment encountered by the battery pack during actual operation of new energy vehicles. Furthermore, by separating and integrating these tests, it further improves the accurate identification of bolt defects. The added environmental simulation does not require excessive cost, effectively improving the overall efficiency and accuracy of the testing system.

This invention discloses an automatic clamping control system for plant specimens, belonging to the field of automatic control technology. The invention acquires high-resolution color images of plant specimens through a data acquisition module; a model building module constructs a color point cloud model based on the high-resolution color images; a pressure monitoring module acquires real-time pressure distribution data between the clamping actuator and the plant specimen; a data processing module analyzes the color point cloud model and pressure distribution data to obtain a fused feature sequence, and identifies candidate abnormal regions and their types based on multimodal joint analysis; a decision execution module generates differentiated clamping control strategies based on the abnormal region types. This system realizes multi-source perception and collaborative decision-making during the plant specimen clamping process, enabling early identification of pre-damage risks and dynamic optimization of clamping behavior, effectively avoiding damage to plant specimens and significantly improving the safety of digital processing of plant specimens.

The invention belongs to the technical field of federated learning, and discloses a method and system for carrying out gradient inversion attack on privacy image data of a user in a federated learning system, and the method comprises the steps: firstly, pre-training a diffusion model and a coder-decoder based on a small amount of prior data similar to the user data through a honest and curious federated server; through gradient information uploaded by a user, a server calculates to obtain a user image tag and a high-dimensional feature. The server constructs a diffusion sampling schedule and a decreasing iteration schedule. And in the progressive decrease round, the generation process of diffusion model sampling is disassembled, gradient matching optimization is carried out on an intermediate result generated by sampling of each time step, and a feature search space is limited through high-dimensional feature regularization until all time steps are completed. According to the method, the gradient inversion attack capability on the user image in the federated learning scene is effectively improved, the similarity and visual quality of the reconstructed image are improved, and more user privacy features are revealed.