Abstract: A target detection method based on fusion of vision, lidar and millimeter wave radar comprises: obtaining original data detected by a camera, a millimeter wave radar, and a lidar, and synchronizing the millimeter wave radar, the lidar, and the camera in time and space; performing a calculation on the original data detected by the millimeter wave radar according to a radar protocol; generating a region of interest by using a position, a speed, and a radar reflection area obtained from the calculation; extracting feature maps of a point cloud bird's-eye view and the original data detected by the camera; projecting the region of interest onto the feature maps of the point cloud bird's-eye view and the original data detected by the camera; fusing the feature maps of the point cloud bird's-eye view and the original data detected by the camera, and processing a fused image through a fully connected layer.

Abstract: A device may receive behavior information that identifies a first user, of a first set of users, in association with a behavior. The behavior may relate to one or more requests, from a client device being used by the first user, to access a network resource. The device may determine, based on a model, whether the behavior is normal. The model may include a normal behavior pattern based on behavior information associated with the first set of users. The device may provide an instruction to allow the client device to proceed with the behavior or provide an instruction to disallow the client device from proceeding with the behavior based on determining whether the behavior is normal. The device may update the model based on the behavior information that identifies the first user and that identifies the behavior.

Abstract: A device may receive behavior information that identifies a first user, of a first set of users, in association with a behavior. The behavior may relate to one or more requests, from a client device being used by the first user, to access a network resource. The device may determine, based on a model, whether the behavior is normal. The model may include a normal behavior pattern based on behavior information associated with the first set of users. The device may provide an instruction to allow the client device to proceed with the behavior or provide an instruction to disallow the client device from proceeding with the behavior based on determining whether the behavior is normal. The device may update the model based on the behavior information that identifies the first user and that identifies the behavior.

Abstract: A device may receive behavior information that identifies a first user, of a first set of users, in association with a behavior. The behavior may relate to one or more requests, from a client device being used by the first user, to access a network resource. The device may determine, based on a model, whether the behavior is normal. The model may include a normal behavior pattern based on behavior information associated with the first set of users. The device may provide an instruction to allow the client device to proceed with the behavior or provide an instruction to disallow the client device from proceeding with the behavior based on determining whether the behavior is normal. The device may update the model based on the behavior information that identifies the first user and that identifies the behavior.

Abstract: A method and device may estimate the accuracy of position data using kernel density estimator. The method may include receiving, from a plurality of user devices, network requests having embedded position data representing locations of the plurality of user devices. The method further includes extracting, from the network requests over a time period, the embedded position data of a user device associated with the plurality of user devices; and receiving baseline position data representing the locations of the user device over the time period. The method included generating a probability density estimate of the locations of the user device based on a kernel density estimator using the baseline position data, determining accuracy scores for the embedded position data using the probability density estimate of the locations, and filtering the embedded position data to remove outliers from the embedded position data.

Abstract: A device may receive performance information for a traffic flow assigned to a quality of service (QoS) class. The device may determine an overall packet delay, associated with the traffic flow, based on the performance information. The device may determine a radio access network (RAN) delay, associated with the traffic flow, based on the performance information. The device may determine a target packet delay associated with the QoS class. The device may identify, based on the target packet delay, the RAN delay, and the overall packet delay, a QoS sub-class to which the traffic flow is to be assigned. The QoS sub-class may be associated with the QoS class. The device may cause packets, associated with the traffic flow, to be marked for treatment in accordance with the QoS sub-class.

Abstract: A device may receive behavior information that identifies a first user, of a first set of users, in association with a behavior. The behavior may relate to one or more requests, from a client device being used by the first user, to access a network resource. The device may determine, based on a model, whether the behavior is normal. The model may include a normal behavior pattern based on behavior information associated with the first set of users. The device may provide an instruction to allow the client device to proceed with the behavior or provide an instruction to disallow the client device from proceeding with the behavior based on determining whether the behavior is normal. The device may update the model based on the behavior information that identifies the first user and that identifies the behavior.

Abstract: A device may receive performance information for a traffic flow assigned to a quality of service (QoS) class. The device may determine an overall packet delay, associated with the traffic flow, based on the performance information. The device may determine a radio access network (RAN) delay, associated with the traffic flow, based on the performance information. The device may determine a target packet delay associated with the QoS class. The device may identify, based on the target packet delay, the RAN delay, and the overall packet delay, a QoS sub-class to which the traffic flow is to be assigned. The QoS sub-class may be associated with the QoS class. The device may cause packets, associated with the traffic flow, to be marked for treatment in accordance with the QoS sub-class.

Abstract: A device may receive traffic information associated with traffic flows assigned to a group of quality of service (QoS) treatment levels and travelling via a network device. The device may identify a group of target performance metrics corresponding to the group of QoS treatment levels. The device may determine a group of weight factors based on the traffic information and the group of target performance metrics. The group of weight factors may be determined such that a group of predicted performance metrics is optimized with respect to the group of target performance metrics. The device may output information identifying the group of weight factors to cause a parameter, associated with the network device, to be updated based on the group of weight factors. The parameter may relate to a manner in which the traffic flows are processed by the network device.

Abstract: A device may receive traffic information associated with traffic flows assigned to a group of quality of service (QoS) treatment levels and travelling via a network device. The device may identify a group of target performance metrics corresponding to the group of QoS treatment levels. The device may determine a group of weight factors based on the traffic information and the group of target performance metrics. The group of weight factors may be determined such that a group of predicted performance metrics is optimized with respect to the group of target performance metrics. The device may output information identifying the group of weight factors to cause a parameter, associated with the network device, to be updated based on the group of weight factors. The parameter may relate to a manner in which the traffic flows are processed by the network device.

Abstract: A system may receive raw information associated with a network and may prepare the raw information to create optimized information. The optimized information may include the raw information that has been sorted. The system may correlate the optimized information to create a set of correlated information. The system may aggregate at least two sets of correlated information to create aggregated information. The system may determine that network analytics are to be performed using the set of correlated information or the aggregated information. The system may determine information associated with performing the network analytics, including the set of correlated information or the aggregated information. The system may perform the network analytics based on the information associated with performing the network analytics. The system may provide a result associated with performing the network analytics. The result may indicate a manner in which to improve a performance of the network.

Abstract: A vaginal drug delivery device includes a tubular base of an inert rubber composition, a first layer having a thickness up to 3 mm composed of a mixture of a drug to delivered, at least one surfactant and at least one dispersing agent applied to said outer surface of the tubular base, and a second layer of silicone rubber having a thickness up to 1 mm encapsulating the first layer on the tubular base whereby said drug will diffuse through said second layer when the device is inserted into the vagina to treat the patient with the drug in the first layer.