Abstract: A multi-intelligence federated reinforcement learning (FRL)-based vehicle-road cooperative control system and method at the complex intersection use a vehicle-road cooperative control framework based on the Road Side Unit (RSU) static processing module and the vehicle-based dynamic processing module. The historical road information is supplied by the proposed RSU module. The Federated Twin Delayed Deep Deterministic policy gradient (FTD3) algorithm is proposed to connect the federated learning (FL) module and the reinforcement learning (RL) module. The FTD3 algorithm transmits only neural network parameters instead of vehicle samples to protect privacy. Firstly, FTD3 selects only specific networks for aggregation to reduce the communication cost. Secondly, FTD3 realizes the deep combination of FL and RL by aggregating target critic networks with smaller Q-values. Thirdly, RSU neural network participates in aggregation rather than training, and only shared global model parameters are used.

Abstract: A method for executing computation, a computing device, a computing system, and a storage medium are provided. The method includes: confirming, via a compiler, whether there is a call instruction related to a thread block modification request in a kernel function to be compiled; in response to confirming that there is the call instruction related to the thread block modification request in the kernel function to be compiled, determining a corresponding program segment associated with the call instruction; configuring a required thread block and thread local register for the corresponding program segment; and inserting a control instruction into the corresponding program segment to enable the thread block configured for the corresponding program segment to execute relevant computation of the corresponding program segment, and an unconfigured thread block not to execute the relevant computation. The disclosure can improve overall performance, make coding and maintenance easy and reduce error rate of code.

Abstract: An intelligent driving system (10). A vehicle-mounted subsystem (100), a road subsystem (200), and a platform subsystem (300) are provided, wherein traffic information of each road section is acquired in real time by using the road subsystem (200), and traveling paths of vehicles in each road section are obtained according to the traffic information; the platform subsystem (300) comprehensively obtains, according to the traffic information of each road section, traffic information of the entire administrative area, and the optimal traveling path; and the vehicle-mounted subsystem (100) provides a traveling path on the basis of the road subsystem (200) and the platform subsystem (300), controls an operation state of the current vehicle in view of a traveling state of itself, and performs traveling path planning according to actual road conditions of each road section, thereby improving the accuracy of traveling path planning and the safety of automatic driving.

Abstract: This application provides a surgical robot control method, a computer device, and a surgical robot system. The surgical robot control method includes: receiving a user demand, and generating an interactive control command; generating a motion control command according to the interactive control command; and controlling a terminal of a mechanical arm to perform the motion control command. The motion control command includes controlling the terminal of the mechanical arm to act in accordance with a plurality of motion modes. The surgical robot control method may control the terminal of the mechanical arm to act in accordance with the plurality of motion modes through the motion control command, realize different motion schemes in various clinical application scenarios, and realize controlling the terminal of the mechanical arm to be flexibly switched between the plurality of motion modes in any application scenario through the interactive control command.

Abstract: Various embodiments of the present disclosure are directed towards an image sensor. The image sensor comprise a substrate having a first region and a second region. A first gate overlies the first region. A second gate overlies the second region. A deep trench isolation (DTI) structure is in the substrate and laterally between the first region and the second region. A first floating diffusion node is in the first region. A second floating diffusion node is in the second region. An interlayer dielectric (ILD) structure is over the substrate. A dielectric structure is between the ILD structure and the substrate. The dielectric structure is laterally between the first and second floating diffusion nodes. The dielectric structure is laterally spaced from the first and second gates. The dielectric structure overlies the DTI structure. A width of the dielectric structure is greater than a width of the DTI structure.

Abstract: Disclosed herein are methods, structures, apparatus and devices for the termination of unused waveguide ports in planar photonic integrated circuits with doped waveguides such that free-carrier absorption therein may advantageously absorb any undesired optical power resulting in a significant reduction of stray light and resulting reflections.

Abstract: A slurry blending tool may include a blending tank to receive and blend one or more materials into a slurry, and at least one inlet pipe connected to the blending tank and to provide the one or more materials to the blending tank. The at least one inlet pipe may vertically enter the blending tank and may not contact the blending tank. The slurry blending tool may include a blending pump partially provided within the blending tank and to blend the one or more materials into the slurry. The slurry blending tool may include an outlet pipe connected to the blending pump and to remove the slurry from the blending tank.

Abstract: A slurry dispersion system is provided, and includes a slurry source system, an in-line analyzer and a controller. The slurry source system provides a slurry for a chemical mechanical polishing (CMP) process. The in-line analyzer measures at least large particle count (LPC) value of a sampled slurry sampled from the slurry dispersion system, and generates an indication signal based on the LPC value, in which the indication signal indicates at lease one characteristic of the slurry. The controller receives the indication signal, and generates a control signal based on the indication signal for performing a real time control on the slurry dispersion system for controlling quality of the slurry.

Abstract: The present invention relates to a citrus oil extract comprising a reduced amount of oxidative byproduct compounds, wherein the amount of oxidative byproduct compounds is less than 2500 ppm. Preferably oxidative byproduct compound is selected from the group consisting of: p-cymene, t-p-menth-2-en-1-ol, t-p-mentha-2,8-dienol, t-limonene oxide, camphor, p-cymen-8-ol, t-limonene 8,9-epoxide, c-limonene 8,9-epoxide and t-carveol. The invention also relates to methods for the preparation of an citrus oil extract, and a flavored or perfumed consumer product comprising the citrus oil extract of the invention.

Abstract: A multi-die QFN hybrid package includes a carrier having flip-chip leads and wire-bonding leads. A first die and a second die are mounted on the flip-chip leads, respectively, in a flip-chip manner. The first die is spaced apart from the second die. A third die is stacked over the first die and the second die. The third die is electrically connected to the wire-bonding leads around the first die and the second die through bond wires. A mold cap encapsulates the first die, the second die, the third die, the bond wires, and partially encapsulates the carrier. The flip-chip leads and the wire-bonding leads are exposed from a bottom mold cap surface.

Abstract: A slurry blending tool may include a blending tank to receive and blend one or more materials into a slurry, and at least one inlet pipe connected to the blending tank and to provide the one or more materials to the blending tank. The at least one inlet pipe may vertically enter the blending tank and may not contact the blending tank. The slurry blending tool may include a blending pump partially provided within the blending tank and to blend the one or more materials into the slurry. The slurry blending tool may include an outlet pipe connected to the blending pump and to remove the slurry from the blending tank.

Abstract: A multi-intelligence federated reinforcement learning (FRL)-based vehicle-road cooperative control system and method at the complex intersection use a vehicle-road cooperative control framework based on the Road Side Unit (RSU) static processing module and the vehicle-based dynamic processing module. The historical road information is supplied by the proposed RSU module. The Federated Twin Delayed Deep Deterministic policy gradient (FTD3) algorithm is proposed to connect the federated learning (FL) module and the reinforcement learning (RL) module. The FTD3 algorithm transmits only neural network parameters instead of vehicle samples to protect privacy. Firstly, FTD3 selects only specific networks for aggregation to reduce the communication cost. Secondly, FTD3 realizes the deep combination of FL and RL by aggregating target critic networks with smaller Q-values. Thirdly, RSU neural network participates in aggregation rather than training, and only shared global model parameters are used.

Abstract: A solid component for preparing an olefin polymerization catalyst, and a preparation method therefor and an application thereof are provided. The solid component contains: (i) a magnesium compound as represented by the following formula (1), (ii) a Lewis base (LB), and optionally, (iii) metal components other than magnesium. The LB is a compound as shown in general formula (II) or an amide compound as shown in general formula (II?). The solid component has a better particle morphology, and a catalyst prepared by using the solid component as a carrier is less likely to be crushed, and has better stereostructural orientation in olefin polymerization, particularly in propylene polymerization or copolymerization.

Abstract: A spherical carrier for olefin polymerization catalysts has at least one magnesium-containing compound having a structure represented by formula (1). The spherical carrier has a relatively good particle morphology, and substantially no abnormally morphological particles will appear. A method for preparing the spherical carrier can be used to prepare a carrier having a small particle size and greatly expands the particle size range of the preparable carrier. When the catalyst prepared by using the carrier is used for olefin polymerization, polymerization activity is good, substantially no abnormally morphological material is present, and hydrogen response is good.

Abstract: A parallel remote control driving system for an intelligent network vehicle, comprising an intelligent network vehicle control device, a parallel driving control and a remote control driving device. The remote control driving device generates, according to a remote control driving request signal from the parallel driving control device, a remote control driving instruction signal, and generates, according to user operations, a driving mode signal and a vehicle control signal and transmits the vehicle control signal to the intelligent network vehicle control device through the parallel driving control device, for remote control of the intelligent network vehicle. With the parallel remote control driving system, the human driver is no longer necessary when the intelligent network vehicle is on the road. Therefore, the manpower cost including training cost, the technical requirement, and the safety cost of the human driver is significantly reduced, thus facilitating the promotion of intelligent network vehicle.

Abstract: The present disclosure discloses a simple preparation method of green acid-modified cyclodextrin, belonging to the field of cyclodextrin modification. Cyclodextrin and succinic acid are mutually modified by an esterification reaction in the presence of catalyst, and acid-modified cyclodextrin with different modification degrees is obtained by controlling reaction time at a high temperature or under a microwave action. The prepared modified cyclodextrin does not involve the use of toxic and harmful reagents, the modification reaction is simple, green and controllable, and the loading capacity of the modified cyclodextrin to guest molecules is remarkably improved compared with the loading capacity of the original cyclodextrin, so that the modified cyclodextrin has great application potential in the health fields such as food, medicines, and cosmetics.