Abstract: The embodiments are directed towards providing personalized federated learning (PFL) models via sharable federated basis models. A model architecture and learning algorithm for PFL models is disclosed. The embodiments learn a set of basis models, which can be combined layer by layer to form a personalized model for each client using specifically learned combination coefficients. The set of basis models are shared with each client of a set of the clients. Thus, the set of basis models is common to each client of the set of clients. However, each client may generate a unique PFL based on their specifically learned combination coefficients. The unique combination of coefficients for each client may be encoded in a separate personalized vector for each of the clients.

Abstract: A display substrate, a manufacturing method thereof, and a display device are provided. The display substrate includes a base substrate and a white OLED display unit on the base substrate, and further includes: an optical adjustment structure on a light emitting side of the white OLED display unit, where the optical adjustment structure is in a peripheral region of each pixel region. The optical adjustment structure is configured to absorb light in a first wavelength range or convert light in a first wavelength range into light in a second wavelength range.

Abstract: A semiconductor package including one or more dam structures and the method of forming are provided. A semiconductor package may include an interposer, a semiconductor die bonded to a first side of the interposer, an encapsulant on the first side of the interposer encircling the semiconductor die, a substrate bonded to the a second side of the interposer, an underfill between the interposer and the substrate, and one or more of dam structures on the substrate. The one or more dam structures may be disposed adjacent respective corners of the interposer and may be in direct contact with the underfill. The coefficient of thermal expansion of the one or more of dam structures may be smaller than the coefficient of thermal expansion of the underfill.

Abstract: The present invention discloses a large-scale MIMO satellite mobile communication method and system. A satellite or gateway station uses spatial angle information of user terminals to group the users to be served in the coverage area to form space division user groups, wherein the user terminals in the same group use the same time-frequency resources to communicate with the satellite, while the user terminals in different groups use different time-frequency resources to communicate with the satellite. For the user terminals in the same space division user group, the satellite or gateway station uses statistical channel information of each user terminal to calculate a downlink precoding vector and an uplink receiving processing vector corresponding to each user terminal, and then uses the obtained vectors to perform downlink precoding transmission and uplink received signal processing.

Abstract: The present invention discloses a satellite-based blockchain architecture, including a terrestrial blockchain miner network, a constellation system, and a consensus protocol coordinating the constellation system and the terrestrial blockchain miner network. In each round, a satellite generates an oracle, and satellites broadcast the oracle to the terrestrial blockchain miner network. The oracle selects a terrestrial miner as a winner of the current round based on a specific rule. The winning terrestrial miner has the right to generate a new block in the round, and broadcasts the new block to other miners by using the terrestrial blockchain miner network. Other miners receiving the new block check the validity of the block, and if the check succeeds, the block is broadcast to other miners by using the terrestrial blockchain miner network.

Abstract: A lane stripe detecting method based on a three-dimensional LIDAR is for detecting a lane stripe of a road surface around a vehicle. A data acquisition transforming step is for obtaining a plurality of vehicle scan point coordinates. The vehicle scan point coordinates are divided into a plurality of scan lines, and each of the scan lines has a plurality of scan points. A horizontal layer lane stripe judging step is judging whether each of the scan points is a horizontal stripe point or a non-horizontal stripe point according to a threshold value of each of the scan lines. At least two of the threshold values of the scan lines are different from each other. A vertical layer lane stripe judging step is for judging whether each of the horizontal stripe points is a same lane stripe point or a different lane stripe point.

Abstract: A large-scale MIMO (Multiple-Input Multiple-Output) wireless transmission method for millimeter wave/Terahertz networks is provided. In order to reduce the interruption of propagation in the millimeter wave/Terahertz band, a plurality of cells are combined into a wireless transmission network, the base station in each cell is equipped with a large-scale antenna array, and a unitary transformation matrix is used to achieve large-scale beam coverage for user terminals in the entire network. Moreover, in order to reduce the influence of the multipath and Doppler effects on transmission performance, received signals are synchronized for time and frequency in each receiving beam of a user terminal. The method allocates power for signal transmission according to the statistic information of synchronized equivalent channels, and gives an optimal power allocation matrix by iterative solution based on the CCCP (concave-convex procedure) and the deterministic equivalent method.

Abstract: A video system, a video processing method, a device and a computer readable medium are disclosed. The system includes: a front-end device and a cloud server; the front-end device is configured to collect video stream data, and set a video identifier and a service scenario identifier for the video stream data, upload the video identifier, the video stream data and the service scenario identifier to the cloud server; the cloud server is configured to generate a video file corresponding to the video identifier according to the service scenario identifier, the video identifier and the video stream data; and store the video file.

Abstract: A large-scale MIMO (Multiple-Input Multiple-Output) wireless transmission method for millimeter wave/Terahertz networks is provided. In order to reduce the interruption of propagation in the millimeter wave/Terahertz band, a plurality of cells are combined into a wireless transmission network, the base station in each cell is equipped with a large-scale antenna array, and a unitary transformation matrix is used to achieve large-scale beam coverage for user terminals in the entire network. Moreover, in order to reduce the influence of the multipath and Doppler effects on transmission performance, received signals are synchronized for time and frequency in each receiving beam of a user terminal. The method allocates power for signal transmission according to the statistic information of synchronized equivalent channels, and gives an optimal power allocation matrix by iterative solution based on the CCCP (concave-convex procedure) and the deterministic equivalent method.

Abstract: The invention discloses a beam domain optical wireless communication method and system. A base station is equipped with an array of optical transceiver ports or transmitter/receiver ports and a lens, each optical transceiver port forms a beam with centralized energy through the lens, and the base station generates beams in different directions by using the optical transceiver port array and the lens, thereby realizing multi-beam coverage or large-scale beam coverage in a communication region. The base station transmits/receives signals of multiple or a large number of user terminals by using channel state information of each user terminal, and different optical transceiver ports transmit/receive signals in different directions, thereby realizing simultaneous communication and bidirectional communication between the base station and different user terminals.

Abstract: The present invention discloses a large-scale MIMO satellite mobile communication method and system. A satellite or gateway station uses spatial angle information of user terminals to group the users to be served in the coverage area to form space division user groups, wherein the user terminals in the same group use the same time-frequency resources to communicate with the satellite, while the user terminals in different groups use different time-frequency resources to communicate with the satellite. For the user terminals in the same space division user group, the satellite or gateway station uses statistical channel information of each user terminal to calculate a downlink precoding vector and an uplink receiving processing vector corresponding to each user terminal, and then uses the obtained vectors to perform downlink precoding transmission and uplink received signal processing.

Abstract: The invention discloses a beam domain optical wireless communication method and system. A base station is equipped with an array of optical transceiver ports or transmitter/receiver ports and a lens, each optical transceiver port forms a beam with centralized energy through the lens, and the base station generates beams in different directions by using the optical transceiver port array and the lens, thereby realizing multi-beam coverage or large-scale beam coverage in a communication region. The base station transmits/receives signals of multiple or a large number of user terminals by using channel state information of each user terminal, and different optical transceiver ports transmit/receive signals in different directions, thereby realizing simultaneous communication and bidirectional communication between the base station and different user terminals.

Abstract: An active safety assistance system for pre-adjusting speed and a control method using the same detect whether there is one other vehicle around a host vehicle. The trajectory of the other vehicle neighboring the host vehicle is estimated when the other vehicle exists. After fitting the other-vehicle trajectory to a lane to determine the intention of the other vehicle, the method determines whether the other vehicle is used as a target vehicle that influences the movement of the host vehicle and calculates a control parameter according to the fitted result and the intention of the other vehicle. The method calculates a target speed and a steering-wheel angle of the host vehicle and controls a steering wheel, a throttle pedal and a brake force of the host vehicle according to the trajectory, the control parameter, and the target speed of the host vehicle.

Abstract: A nighttime vehicle detecting method is for capturing an image by a camera and driving a computing unit to compute the image and then detect a highlight point of the image. The nighttime vehicle detecting method is for driving the computing unit to perform a communicating region labeling algorithm to label a plurality of highlight pixels connected to each other as a communicating region value, and then performing an area filtering algorithm to analyze an area of the highlight pixels connected to each other and judge whether the highlight pixels connected to each other are a vehicle lamp or not according to a size of the area. The nighttime vehicle detecting method is for driving the computing unit to perform an optical flow algorithm to obtain a speed of the vehicle lamp, and then filtering the vehicle lamp moved at the speed smaller than a predetermined speed.

Abstract: A video system, a video processing method, a device and a computer readable medium are disclosed. The system includes: a front-end device and a cloud server; the front-end device is configured to collect video stream data, and set a video identifier and a service scenario identifier for the video stream data, upload the video identifier, the video stream data and the service scenario identifier to the cloud server; the cloud server is configured to generate a video file corresponding to the video identifier according to the service scenario identifier, the video identifier and the video stream data; and store the video file.

Abstract: A lane stripe detecting method based on a three-dimensional LIDAR is for detecting a lane stripe of a road surface around a vehicle. A data acquisition transforming step is for obtaining a plurality of vehicle scan point coordinates. The vehicle scan point coordinates are divided into a plurality of scan lines, and each of the scan lines has a plurality of scan points. A horizontal layer lane stripe judging step is judging whether each of the scan points is a horizontal stripe point or a non-horizontal stripe point according to a threshold value of each of the scan lines. At least two of the threshold values of the scan lines are different from each other. A vertical layer lane stripe judging step is for judging whether each of the horizontal stripe points is a same lane stripe point or a different lane stripe point.

Abstract: The lidar detection device uses four two-dimensional lidars to scan an obstacle, to obtain original point-cloud data corresponding to the obstacle, and the original point-cloud data includes a relative distance, a relative angle and a relative speed of the obstacle relative to the vehicle, Next, the point-cloud data is classified into at least one point-cloud group corresponding to the obstacle, and a border length of the obstacle is obtained according to a contour of the point-cloud group. Kalman filter and extrapolation are used to estimate and track a movement path of a dynamic obstacle, and transmit the relative speed and the border length of the dynamic obstacle to an automatic driving controlling device. According to the relative distance, a coordinate of the dynamic obstacle nearest the vehicle can be obtained and transmitted to the automatic driving controlling device efficiently.

Abstract: A nighttime vehicle detecting method is for capturing an image by a camera and driving a computing unit to compute the image and then detect a highlight point of the image. The nighttime vehicle detecting method is for driving the computing unit to perform a communicating region labeling algorithm to label a plurality of highlight pixels connected to each other as a communicating region value, and then performing an area filtering algorithm to analyze an area of the highlight pixels connected to each other and judge whether the highlight pixels connected to each other are a vehicle lamp or not according to a size of the area. The nighttime vehicle detecting method is for driving the computing unit to perform an optical flow algorithm to obtain a speed of the vehicle lamp, and then filtering the vehicle lamp moved at the speed smaller than a predetermined speed.

Abstract: Signal transmitting/receiving electronic devices or terminals in the present disclosure are configured to conduct per-beam signal synchronization in massive MIMO communication with a signal receiving/transmitting device or base station. During the massive MIMO communication, the devices or terminals are configured to transmit/receive signals via a set of beams to or from the signal receiving/transmitting device or base station. For beam domain signal of each individual beam of the plurality beams, the devices or terminals are configured to determine target time adjustments based on time shifts of the beam domain signals induced by multipath effect and target frequency adjustments based on frequency offsets of the beam domain signals induced by the Doppler effect; adjust time independent variables of the beam domain signals by the time adjustments; and adjust frequency independent variables of the beam domain signals by the frequency adjustments.

Abstract: A method is to be implemented by a processor connected to a storage device and an image capturing device, and includes steps of: performing a first and a second classifications on an image received from the image capturing device for respectively obtaining first and second results of obstacle detection, and storing the same in the storage device; determining whether there is a distinct obstacle, which is an obstacle indicated by one of the first and second results of obstacle detection, and not indicated by the other one of the first and second results of obstacle detection; and calculating a penalty score based on the distinct obstacle, and updating a credibility score by subtracting the penalty score therefrom.