-
公开(公告)号：US20240037445A1
公开(公告)日：2024-02-01
申请号：US17969129
申请日：2022-10-19
Applicant: Denso International America, Inc. , DENSO CORPORATION
Inventor： Minglei Huang , Jinning Li , Chen Tang , Masayoshi Tomizuka , Wei Zhan
Abstract: Systems and methods for pessimistic offline reinforcement learning are described herein. In one example, a method for performing offline reinforcement learning determines when sampled states are out of distribution, assigns high probability weights to the sampled states that are out of distribution, generates a fitted Q-function by solving an optimization problem with a minimization term and a maximization term, estimates a Q-value using the fitted Q-function by estimating the overall expected reward assuming the agent is in the present state and performs a present action, and updates the policy according to an existing reinforcement learning algorithm. The minimization term penalizes an overall expected reward when a present state is out of distribution. The maximization term cancels the minimization term when the present state is an in-distribution state.
-
公开(公告)号：US11142196B2
公开(公告)日：2021-10-12
申请号：US16266095
申请日：2019-02-03
Applicant: DENSO International America, Inc. , DENSO CORPORATION
Inventor： Minglei Huang
Abstract: The present disclosure is directed toward a lane detection method. The method includes: acquiring road information for a section of a road upon which a subject vehicle is traveling, determining whether an adjacent vehicle is traveling in vicinity of the subject vehicle, defining, in response to determining the presence of the adjacent vehicle, a vehicle trajectory based on movement of the adjacent vehicle and the road information, detecting one or more lane markings along the road upon which the subject vehicle is traveling, defining a lane marking trajectory in response to detecting the one or more lane markings, calculating a lane accuracy for each estimated lane trajectory that includes the vehicle trajectory, the lane marking trajectory, or a combination thereof, and selecting a drive lane from among the estimated lane trajectories based on the lane accuracy.
-
公开(公告)号：US10223601B1
公开(公告)日：2019-03-05
申请号：US15782061
申请日：2017-10-12
Applicant: DENSO International America, Inc.
Inventor： Minglei Huang
Abstract: A method for generating images of traffic objects using deep learning is disclosed. The method includes obtaining image data from an image. The method includes assigning a set of default parameters to the image data, and the set of default parameters include values associated with at least one of a weather condition of the image and a defect condition of an object of the image. The method includes generating a set of predicted parameters based on the image data. The method includes determining an error for each parameter of the set of predicted parameters, and the error is based on a value of the parameter of the set of predicted parameters and a value of a corresponding default parameter of the set of default parameters. The method includes adjusting a weight of a corresponding connection of the deep learning module.
-
4.发明申请公开(公告)号：US20210304433A1
公开(公告)日：2021-09-30
申请号：US16835592
申请日：2020-03-31
Applicant: Denso International America, Inc.
Inventor： Minglei Huang , Hiroshi Inou
Abstract: A system for measuring accuracy of an electronic map or a localization of a vehicle may include one or more processors and a memory in communication with the one or more processors having an image capture module, a ground truth generating module, a curve function generating module, and an accuracy determining module. These modules contain instructions that when executed by the one or more processors causes the processors to obtain sensor data that includes images of one or more road lane markings of a road or trajectory data regarding the trajectory of the vehicle, determine ground truth curvature values, perform a curve fitting of map lane marking points or map trajectory points, and compare curvature values of a curve function generated by the curve fitting to the ground truth curvature values to determine the accuracy the electronic map or the localization of the vehicle.
-
Patent Agency Ranking
5.发明授权公开(公告)号：US11238607B2
公开(公告)日：2022-02-01
申请号：US16835592
申请日：2020-03-31
Applicant: Denso International America, Inc.
Inventor： Minglei Huang , Hiroshi Inou
Abstract: A system for measuring accuracy of an electronic map or a localization of a vehicle may include one or more processors and a memory in communication with the one or more processors having an image capture module, a ground truth generating module, a curve function generating module, and an accuracy determining module. These modules contain instructions that when executed by the one or more processors causes the processors to obtain sensor data that includes images of one or more road lane markings of a road or trajectory data regarding the trajectory of the vehicle, determine ground truth curvature values, perform a curve fitting of map lane marking points or map trajectory points, and compare curvature values of a curve function generated by the curve fitting to the ground truth curvature values to determine the accuracy the electronic map or the localization of the vehicle.
-
公开(公告)号：US09952058B2
公开(公告)日：2018-04-24
申请号：US15249858
申请日：2016-08-29
Applicant: DENSO International America, Inc.
Inventor： Minglei Huang
CPC classification number: G01C21/365 , B60Q1/346 , G01S7/4808 , G01S17/023 , G01S17/936 , G01S19/13 , G06K9/00805 , G06K9/00832
Abstract: A driver visibility detection system for a vehicle includes an indicator, an eye position obtainer, and a visible range calculator. The indicator indicates the driver to look at, through a windshield of the vehicle, a furthest ground position of a road on which the vehicle is traveling. The eye position obtainer recognizes an eye position of the driver when the driver looks at the furthest ground position. The visible range calculator calculates a visible range of the driver based on the eye position detected by the eye position obtainer.
-
公开(公告)号：US09896107B1
公开(公告)日：2018-02-20
申请号：US15281423
申请日：2016-09-30
Applicant: Denso International America, Inc.
Inventor： Minglei Huang
CPC classification number: B60W30/18163 , B60Q1/00 , B60W10/04 , B60W10/184 , B60W10/20 , B60W50/14 , B60W2420/42 , B60W2420/52 , B60W2710/00 , G05D1/0088 , G05D1/0246 , G08G1/166 , G08G1/167
Abstract: A system and method are provided and include a light source projector mounted on a subject vehicle, a controller that controls the light source projector to project a bounding box on a roadway upon which the subject vehicle is traveling in a destination lane of a lane change of the subject vehicle; and a sensor that detects a response image projected by a secondary vehicle on the roadway in the adjacent lane, the response image indicating either agreement or disagreement with the lane change of the subject vehicle. The controller generates an alert in the subject vehicle based on the detected response.
-
公开(公告)号：US11760412B2
公开(公告)日：2023-09-19
申请号：US17708751
申请日：2022-03-30
Applicant: DENSO CORPORATION
Inventor： Minglei Huang , Prasanna Sivakumar
CPC classification number: B62D6/02 , B62D15/021 , G01S7/4802
Abstract: A method includes obtaining steering data of the vehicle from one or more steering sensors, determining whether the vehicle is operating in one of a turning state and a non-turning state based on the steering data, performing a localization routine based on a first echo set of the 3D data points from among the plurality of 3D data points when the vehicle is operating in the turning state, and performing the localization routine based on a second echo set of the 3D data points from among the plurality of 3D data points when the vehicle is operating in the non-turning state, where a number of the plurality of 3D data points of the first echo set is greater than a number of the plurality of 3D data points of the second echo set.
-
公开(公告)号：US11117576B2
公开(公告)日：2021-09-14
申请号：US16266780
申请日：2019-02-04
Applicant: DENSO CORPORATION
Inventor： Minglei Huang , Yu Yan , Hiroshi Inou
Abstract: A lane trace control system for maintaining a vehicle in its intended lane of travel. The system includes a lane trace control module configured to: compare the road information obtained by a lane recognition camera with map data obtained from a map module; determine a confidence level of the lane recognition camera based on a magnitude of any differences identified between the road information obtained from the lane recognition camera and the obtained map data; generate a target wheel angle for keeping the vehicle in its intended lane based entirely on the road information obtained from the lane recognition camera when the confidence level is above a predetermined threshold; and generate the target wheel angle based on a combination of the road information obtained from the lane recognition camera and the obtained map data of the road when the confidence level is below the predetermined threshold.
-
-
-
-
-
-
-
-
Search Results
9
records
(took 0.045 seconds)
