5results about How to "Implement security detection" patented technology

This invention discloses a method and system for mind chain hijacking security analysis of VLA models, belonging to the fields of artificial intelligence security and embodied intelligence. The invention first acquires the VLA model under test and collects normal trajectories to construct an offline test dataset; then, it initializes an adversarial patch and overlays it onto the original visual observation target position to generate adversarial observation; subsequently, it constructs a joint optimized loss function including mind chain adversarial loss and action adversarial loss, constraining the model to generate a preset target mind chain and target action when inputting unaltered instructions; then, iterative optimization using an optimization algorithm yields the optimal adversarial patch; finally, it uses this optimal patch to test the mind chain hijacking vulnerability in the model under test. This invention has strong concealment and wide applicability, achieving covert and targeted hijacking of robot behavior without tampering with user instructions, effectively detecting security vulnerabilities in the inference layer of VLA models, and providing support for the secure deployment of embodied intelligence systems.

The utility model discloses an anode assembly, an X-ray tube with the anode assembly and an X-ray imaging device, and the anode assembly comprises a target assembly which is arranged in a vacuum chamber and is used for being bombarded by electron beams emitted by a cathode assembly to generate X-rays; the anode cylinder is arranged on the periphery of the target assembly in a sleeving mode and integrated with the target assembly, the end, facing the tube shell, of the anode cylinder is connected with the tube shell in a sealed mode, and a first through hole used for X-ray emitting is defined in the circumferential direction of the anode cylinder; the anode shielding cover is arranged outside the anode cylinder in a sleeving manner and connected with the anode cylinder, and a second through hole which corresponds to the first through hole and communicates with the vacuum cavity is defined in the circumferential direction of the anode shielding cover; and the outer layer window seals the second through hole. According to the anode assembly disclosed by the embodiment of the utility model, the target assembly and the anode cylinder are integrated into a whole, so that the anode assembly with a smaller diameter can be obtained, and the anode assembly is convenient to assemble into an ultrathin X-ray source and a miniaturized X-ray imaging device, so that the safety detection of a narrow area is realized.

This application provides a security protection method and a security resource pool. The method includes: obtaining service traffic from a service switch through a traffic splitter deployed on a third server and forwarding it to an orchestrator deployed on a second server; using the orchestrator, sequentially forwarding the service traffic to corresponding security devices for security detection according to a preset security service chain, wherein the security devices are deployed on a first server; the traffic splitter, the orchestrator, and the security devices are all virtual machines; using the orchestrator, the service traffic that passes the security detection is returned to the service switch via the traffic splitter. Through a full-dimensional redundancy design from the underlying platform to virtual network elements, the single point of failure problem is completely solved, the availability of the security resource pool is improved, and continuous security protection of service traffic is achieved.

This invention discloses a message security detection method, device, and storage medium. The method includes: acquiring a message to be detected based on an intermediate system-to-intermediate system routing protocol; determining a first data structure for the message content based on the message to be detected; if a pre-determined message detection rule determines the existence of a historical message to be detected, analyzing the historical message and the first data structure to determine the message's legitimacy. This solves the problem of not being able to detect abnormal ISIS messages in a timely manner. By analyzing the message to be detected to determine the first data structure for the message content, and determining whether a historical message to be detected exists according to the pre-determined message detection rule, if so, analyzing the historical message and the first data structure to determine the legitimacy of the message to be detected, this invention achieves security detection of messages formed by intermediate system-to-intermediate system routing protocols, enabling timely detection of abnormal messages and improving data security.

The utility model belongs to the technical field of conveyor control, and particularly relates to an automobile seat line production and transportation system. Comprising a transfer roller way, a waiting roller way and a controller, the transfer roller way is arranged at the front end of the waiting roller way in the conveying direction, a plurality of to-be-transferred stations are arranged on the transfer roller way, a transfer roller way stopper is arranged at the front end of the transfer roller way in the conveying direction, a to-be-transferred station detection switch is arranged on the side face of each to-be-transferred station, and the to-be-transferred station detection switch is connected with the controller. A plurality of waiting stations are arranged on the waiting roller way, a waiting roller way stopper is arranged in front of the conveying direction of each waiting station, and a waiting station detection switch is arranged on the side face of each waiting station. The probability of problems is greatly reduced, and the safety of equipment is ensured.