6results about How to "Avoid violations" patented technology

The invention provides an Internet of Vehicles key information classification early warning method and a vehicle-mounted device. The Internet of Vehicles key information classification early warning method comprises the steps: acquiring Internet of Vehicles information of a target vehicle; determining current driving state information of the target vehicle according to the Internet of Vehicles information; and determining the type of information sent to the target vehicle according to the driving state information. According to the invention, the information with different emergency degrees can be sent to the user according to the current driving state information, multi-channel grading early warning is realized according to the information type, and the effects of assisting driving, promoting safety and improving user experience are achieved.

The utility model provides a remote drainage container monitoring device which comprises a container which is located in a ward and used for containing patient drainage liquid, and further comprises a fixing base which is installed at the bottom of a sickbed and used for containing the container, a camera which is installed on the fixing base and used for monitoring the state of the container, and a server which is arranged at a nurse station and used for receiving and analyzing camera information. The monitor is used for outputting an analysis result of the server, the operation table is used for operating display content of the monitor, and the alarm device is used for giving an alarm according to a preset value of the server. According to the utility model, real-time monitoring can be realized in a nurse station, the workload of nurses is greatly reduced, a quiet environment can be provided for a ward, and meanwhile, medical personnel can intensively know the state of drainage liquid of each patient and timely process the state of the drainage liquid in the patient.

This invention provides a scaling method and system for a vGPU-based SLO-aware model in a computing power network. The method includes: designing and deploying a Prometheus-based monitoring system to collect load metrics from the GPU and instance sides of the Kubernetes cluster, obtaining monitoring metrics; constructing an output length predictor based on a fine-tuned BERT-base lightweight model, and combining it with the real-time system status to build a joint prediction model for first-token latency, inter-token latency, and end-to-end latency, obtaining latency prediction results; based on the monitoring metrics, adopting a two-stage scheduling strategy to achieve dual-level load balancing at the GPU and instance levels; and designing an SLO-aware elastic scaling mechanism based on the monitoring metrics and latency prediction results, dynamically selecting scaling strategies according to SLO violation risks. This invention achieves an optimized balance between operating costs and service performance.

The present invention relates to the filed of automatic drive and intelligent transportation system, an adaptive cruise control system and an adaptive cruise control method supporting traffic light recognition are provided. The adaptive cruise control system in the present invention includes: a wireless receiving module adaptive with a wireless transmission module of the traffic light and configured to receive first traffic light state information about a traffic light on the current lane sent by the wireless transmission module; and a main controller being in communication with the wireless receiving module and configured to obtain the first traffic light state information received from the wireless receiving module, and perform a stop control or adaptive cruise control based on the first traffic light state information. In the present invention, the wireless receiving module adaptive with the wireless transmission module of the traffic light is introduced to obtain the state information of the traffic lights, and an adaptive cruise control for avoiding a driver from violation of traffic light is further realized based on the state information of the traffic lights, so that a safety of driving can be improved.

The application discloses an audio material auditing and warehousing method and system, relates to the technical field of audio auditing, and has the technical scheme as follows: obtaining a target audio material to be audited, and performing compliance screening on the target audio material; if the compliance of the target audio material cannot be determined, determining a propagation scene label after analyzing metadata information and audio waveform features of the target audio material; if the propagation scene label belongs to a cross-channel propagation type, extracting cross-channel audit records of audio of the same type from a historical cross-channel audit database, and calculating a feature correlation coefficient between the historical audio and the target audio material; separating a to-be-determined audio segment from the target audio material, generating an audit report after judging the multi-channel compliance of the to-be-determined audio segment according to the feature correlation coefficient and a real-time audit rule library of each channel, and formulating an adaptive strategy to complete warehousing determination; and the effect is to provide strong support for the standardized management and effective propagation of audio materials.

The present application relates to the technical field of large models and business management, and particularly relates to a procurement document detection method and system based on necessary response item proportion detection; the procurement document analysis method comprises the following steps: obtaining a multi-source input format file; converting the multi-source input format file into a structured text; tokenizing the structured text to obtain minimum semantic units; and encoding the minimum semantic units based on BERT to output the category label of each Token; based on the Token after the category label is labeled, at least the following features are extracted: a pre-constructed structured synonym knowledge graph, synonym retrieval features of a first target Token in the Token are obtained based on dense vector retrieval; artifacts are generated based on end-to-end detection AI of the Token, and text generation authenticity features of all Tokens are obtained by combining the synonym retrieval features; and information confidence features of a second target Token in the Token and a website entity attribute vector are extracted according to the pre-extracted website entity attribute vector.