7 results about "Software diagnosis" patented technology

The invention discloses a multi-agent dynamic diagnosis method and system for a vehicle-mounted software system, and relates to the technical field of software diagnosis. Collecting multi-source operation data of the vehicle-mounted software in real time; performing document analysis and skill processing on the internal technical document of the original equipment manufacturer to obtain an executable skill set; constructing a dynamic dependency graph based on the multi-source operation data and the executable skill set; obtaining a current fault prompt, calling a multi-skill collaborative agent, substituting multi-source operation data and the fault prompt into the dynamic dependency graph, arranging action layer operation in an executable skill set through the agent, executing log analysis, dependency chain verification, abnormal link identification and fault scene restoration, and outputting abnormal associated data; based on the dynamic dependency graph, the abnormal associated data, the rule layer information of the executable skill set and the historical fault cases, root cause reasoning is executed to obtain an executable troubleshooting scheme, and the fault diagnosis accuracy and efficiency of the vehicle-mounted software system are improved.

The application discloses a software diagnosis system, which comprises at least one occurrence domain for recording events triggered by software activities and performing diagnosis on the events; the occurrence domain comprises an event processing source for recording and processing event sources of the software activities; an information block generator connected to the event processing source for generating recording information blocks or event information blocks and sending the recording information blocks or the event information blocks to the event processing source; a primary diagnosis device for performing diagnosis on the recording information blocks or the event information blocks and generating a diagnosis report based on the diagnosis result; and a primary capture device connected to the event processing source and the primary diagnosis device for realizing information docking between the event processing source and the primary diagnosis device and performing information filtering on the recording information blocks or the event information blocks and / or the diagnosis report. Through the above method, the software diagnosis system provided by the application can realize dynamic collection and real-time diagnosis.

The application discloses an AUTOSAR-based CAN-to-LIN diagnosis method and device, equipment and a storage medium, and relates to the technical field of automobile gateway software diagnosis. The application comprises the following steps: after a gateway receives a CAN diagnosis message request sent by a diagnosis instrument, the CAN diagnosis message request transmitted through a CAN driver is received; message conversion is performed on the CAN diagnosis message request, and a LIN diagnosis message request is obtained; the LIN diagnosis message request is transmitted to a LIN driver, and a LIN diagnosis result fed back by the LIN driver according to the LIN diagnosis message request is received; message conversion is performed on the LIN diagnosis result, a CAN diagnosis result is obtained, and the CAN diagnosis result is transmitted to the diagnosis instrument through the CAN driver, so that the development cost of a LIN diagnosis host computer is saved, and CAN and LIN controller diagnosis functions are realized through a CAN host computer.

The invention relates to the technical field of software detection, and discloses a malicious software static searching and killing method and device, equipment and a medium, and the method comprises the following steps: obtaining a permission information file and a calling information file of to-be-detected software; performing feature extraction on the permission information file to obtain a permission feature vector; performing feature extraction on the calling information file to obtain a calling feature vector, and performing character scanning on the calling information file to obtain a character feature vector; constructing a call graph according to the call information file to obtain a function call graph, and performing node feature statistics and global structure feature extraction in the function call graph to obtain call statistical features; and according to the permission feature vector, the calling feature vector, the character feature vector and the calling statistical feature, performing software diagnosis classification on the to-be-detected software, and determining a diagnosis result. The method has the beneficial effects that the multi-scale calling features and the multi-dimensional features are combined, deep analysis of the to-be-detected software is realized, and the detection rate and the detection accuracy of the malicious software are remarkably improved.

The invention relates to an abnormity recovery method and system for a vehicle-mounted camera. The method comprises the steps that after the vehicle-mounted system is started, the running state of a driver in the vehicle-mounted system is determined; the method comprises the following steps: after a sensor driver in a vehicle-mounted system is started, acquiring hardware diagnosis data of a vehicle-mounted camera through a first diagnosis interface in the sensor driver; after an image processing drive in the vehicle-mounted system is started, acquiring software diagnosis data of image data acquired by a vehicle-mounted camera through a second diagnosis interface in the image processing drive; when the vehicle-mounted system runs to a stage of interacting with an upper-layer framework, creating an exception diagnosis and recovery thread, and acquiring hardware diagnosis data and software diagnosis data through the exception diagnosis and recovery thread; and according to the hardware diagnosis data or the software diagnosis data, calling a first recovery interface in the sensor driver or calling a second recovery interface in the image processing driver, and executing recovery operation on the vehicle-mounted camera. The technical problem that the abnormal vehicle-mounted camera cannot be quickly recovered is solved.

The invention discloses a method for out-of-band automatic detection and automatic isolation of AI card faults, which comprises the step of continuously monitoring the state of an AI card through a special hardware module by means of independent network link when a server operation system is paralyzed and AI card software runs abnormally. The method does not depend on a monitoring method of an operating system and application software, realizes a'discovery-processing-recovery 'closed loop of AI card faults through out-of-band automatic detection and isolation through hardware monitoring, software diagnosis and intelligent analysis of independent channels, and is a core technology for efficient operation and maintenance of a data center AI cluster.

This application discloses a relay intelligent control method and system for AC loads. The method includes the following steps: based on the multimodal data of the relay, predicting and updating the mechanical action delay of the relay; and issuing a drive command before the AC zero-crossing point based on the predicted mechanical action delay after receiving a switching command; using parallel hardware fast protection channels and software diagnostic channels to monitor and protect the relay coil in real time; and constructing a switching control strategy library optimized based on intelligent algorithms learning historical experience to identify the load type in real time and dynamically switch the optimal switching control strategy accordingly. The system is used to implement the above method. The beneficial effects of this application are: by combining the hardware fast protection channel with the software protection channel, the circuit can be immediately cut off in the event of a severe short-circuit fault. The hardware protection operates independently of the controller, and combined with the intelligent diagnostics of the software channel, a dual protection redundancy is formed, significantly improving the system's safety level.