30results about How to "Improve autonomous driving performance" patented technology

The invention relates to a motorcade driving method based on roadside equipment. The method comprises steps of enabling the roadside equipment to be disposed at a roadside, enabling at least one following vehicle to follow one pilot vehicle, forming a motorcade, enabling the following vehicle to be a low-grade automatic driving vehicle, and enabling the pilot vehicle to be a high-grade automatic driving vehicle; enabling the pilot vehicle receive the high-precision map information sent by roadside devices and carry out planning driving, enabling each vehicle in the motorcade establish communication with at least two roadside devices and periodically obtain position coordinates of the vehicle based on a signal propagation power loss model, and forming a driving track of the vehicle; enabling each vehicle share the driving track with an adjacent following vehicle behind through wireless communication, and enabling each following vehicle drive along the driving track of a previous vehicle. Compared with the prior art, the method has advantages of high safety, high sensitivity, high error-tolerant rate and the like.

The invention discloses a longitudinal/lateral/vertical personified cooperative control method for an off-road electric drive unmanned vehicle, and relates to the vehicle control. The method comprises the steps: taking a driver as a research object, and analyzing the steering/acceleration and deceleration operation behavior features of a driver under the off-road condition; constructing a personified control system based on multi-performance target optimization according to the parameter uncertainty, nonlinear coupling and time lag of an unmanned vehicle in a non-structure road environment; building a longitudinal/lateral/vertical coupling dynamic personified control system for the off-road electric drive unmanned vehicle, and achieving the dynamic coupling control of the off-road electric drive unmanned vehicle through the transmission, conversion and evolution of the energy and information of dynamic coordination of the coupling dynamic control system, thereby effectively improving the autonomous running comprehensive energy of the electric drive unmanned vehicle, and remarkably improving the performances of the control system.

The invention provides a logistics vehicle and a trailer pose determination, pre-collision detection and automatic driving method thereof, and belongs to the technical field of automatic driving of logistics vehicles. The trailer pose determination method comprises the steps that the actual pose, the actual front wheel turning angle and the actual speed of a trailer are obtained through a sensor; inputting the actual front wheel turning angle and the actual speed of the trailer into the logistics vehicle kinematic model to obtain a trailer pose estimation value and a trailer angle estimation value based on the logistics vehicle kinematic model; comparing the actual pose of the trailer with the estimated value of the pose of the trailer, multiplying the deviation between the actual pose and the estimated value of the pose of the trailer by a coefficient K to obtain a pose error, and correcting the estimated value of the angle of the trailer by using the pose error to obtain an accurate angle of the trailer; and combining the actual pose of the trailer, the accurate trailer angle and the geometric parameters of the trailer and each trailer to obtain the pose of each trailer. The method can accurately determine the pose of the trailer in the logistics vehicle.

The invention discloses a cloud server, a vehicle-mounted terminal and a vehicle, and the cloud server comprises: a user feature receiving module which is used for receiving at least one user featuresent by the vehicle-mounted terminal; The prompt information generation module which is used for generating prompt information through a user model according to at least one user feature; The prompt information sending module which is used for sending the prompt information to the vehicle-mounted terminal, so that the vehicle-mounted terminal interacts with a user according to the prompt information to generate interaction content; The interactive content receiving module whch is used for receiving the interactive content sent by the vehicle-mounted terminal; And the correction module which isused for correcting the user model according to the interaction content. The cloud server can correct the user model according to the interaction content through interaction with the vehicle-mountedterminal, so that the thinking of a developing person can be presented, the human-computer interaction capability is enhanced, the uniqueness of a driver is embodied, more memory computing services can be provided for the driver and passengers, and the auxiliary driving and automatic driving capabilities are enhanced.

The invention relates to a light beam separation module, a laser radar system and a control method thereof. The light beam separation module comprises a first light splitting assembly, an optical rotation assembly and a second light splitting assembly. The first light splitting assembly is used for separating the incident light beam into first P polarized light and first S polarized light; the optical rotation assembly is used for rotating the first P polarized light into second S polarized light and rotating the first S polarized light into second P polarized light; the second light splittingassembly is used for mixing the second P polarized light and the second S polarized light into a first emergent light beam emitted to a target object; the second light splitting assembly is also usedfor separating the echo light beam into third P polarized light and third S polarized light; the optical rotation assembly is also used for maintaining the polarization directions of the third P polarized light and the third S polarized light; and the first light splitting assembly is also used for mixing the third P polarized light and the third S polarized light into a second emergent light beam which is emitted to a laser detection module. According to the embodiment of the invention, the laser detection module can receive the P polarized light, and the energy received by the laser detection module is increased.

The invention discloses a data processing method and device, equipment and a storage medium. The method comprises the steps that first position information and first sensing data of a target vehicle and second position information and second sensing data of a current vehicle are acquired; based on the first position information and the second position information, determining relative position information between the target vehicle and the preceding vehicle and a sensing data detection blind area corresponding to the current vehicle; determining blind area standby sensing data corresponding to the sensing data detection blind area from the first sensing data based on the sensing data detection blind area; according to the relative position information, the second sensing data and the sensing data detection blind area, performing segmentation processing on the blind area standby sensing data to obtain target blind area sensing data corresponding to the sensing data detection blind area; updating blind area sensing data in the second sensing data based on the target blind area sensing data to obtain target sensing data of the current vehicle; according to the invention, the ability of bicycle automatic driving in automatic driving can be improved.

The invention provides an automatic driving method based on conditional imitation learning and reinforcement learning, and the method comprises the steps: S1, collecting driving data, and marking a corresponding decision action as an expert decision; s2, according to the driving data, calculating to obtain a course angle, and obtaining a feature vector and a feature map; s3, constructing a training data set based on the feature vector, the feature map, the course angle and the expert decision, and training a model by using the training data set; and S4, applying the trained model to the vehicle to realize the automatic driving decision of the vehicle. According to the method, by combining conditional imitation learning and reinforcement learning, the dependence of the model on a training data set is reduced, and meanwhile, the problem of low reinforcement learning exploration efficiency caused by random initialization is solved; real vehicle data are collected and applied to model training, so that the trained model is closer to a real driving scene.

The invention discloses a method and device for determining a vehicle preview point and electronic equipment. The method comprises the following steps: acquiring a target trajectory point in an expected trajectory path of a target vehicle; attribute information of the target trajectory point is determined, and the attribute information at least comprises expected curvature information and expected orientation information; acquiring the current driving speed of the target vehicle; determining an elliptical trajectory path of the target vehicle according to the attribute information and the current driving speed; and determining a preview point of the target vehicle based on the elliptical trajectory path and the expected trajectory path. According to the invention, the technical problems of low preview point selection accuracy and poor automatic driving performance of a vehicle preview point determination method in the prior art are solved.

An autonomous driving module mounting structure, includes a roof panel configuring an upper portion of a vehicle body, an open portion provided at the roof panel, and an autonomous driving module that includes a plurality of devices, that is mounted on an upper surface of an outer section of the roof panel, and that is disposed at a position that does not coincide with the open portion in top view.

A LIDAR data clustering apparatus, a system including the same, and a method thereof are provided. The apparatus includes a processor that is configured to generate a voxel map including at least one point data based on a LIDAR and remove a road surface point from the at least one point data. The processor also clusters the at least one point data. A storage of the apparatus is configured to store data obtained by the processor and an algorithm for driving the processor.

The embodiment of the invention provides a trajectory prediction method and device, electronic equipment and a computer storage medium. The method comprises steps that the map information and the perception information are acquired; the map information comprises the position information of various obstacles recognized by a target vehicle; the sensing information comprises information about various obstacles collected by sensors on the target vehicle; a dynamic obstacle is determined from a plurality of obstacles according to the map information and the perception information; when it is determined that the dynamic obstacle contains the signal lamp, the type of the signal lamp is determined; and a corresponding trajectory prediction model is determined according to the type of the signal lamp, and trajectory prediction is conducted on the dynamic obstacle.

The invention discloses a vehicle-mounted millimeter-wave radar antenna, a radar and a vehicle, and the vehicle-mounted millimeter-wave radar antenna comprises a 1-to-n power dividing feed network, and n is a natural number greater than 2; n single-column antenna arrays, each single-column antenna array is correspondingly connected with one output end of the power division feed network, each single-column antenna array is formed by connecting k antenna radiation units in series through a microstrip line, and k is a natural number greater than 3; and the upper bottom surface of the trapezoidal reflection matching section is connected with the antenna radiation unit at the tail end of the single-row antenna array, and the lower bottom surface of the trapezoidal reflection matching section is connected with the microstrip line. By adding an antenna design structure of a trapezoidal reflection matching section in front of the last antenna radiation unit, the sidelobe level of a directional diagram can be integrally reduced by about 2dB, the antenna directional diagram can be effectively optimized, the signal-to-noise ratio of a radar system can be improved, and the anti-interference capability of a millimeter wave radar system can be improved.

The invention discloses a data processing method, a data processing device and an automatic driving system, and the data processing method comprises the steps: carrying out the feature extraction of original point cloud data, and obtaining first point cloud feature data; performing feature extraction on the original image data to obtain first image feature data; based on preset target classification candidate information, screening the first point cloud feature data and the first image feature data to obtain second point cloud feature data and second image feature data; fusing the second point cloud feature data with the second image feature data to obtain fused data; and performing target identification on the fused data to obtain a target identification result. By adopting the above scheme, the processing efficiency and accuracy of data fusion can be guaranteed with low implementation cost, and thus a more accurate target recognition result can be obtained.