-
公开(公告)号：US20180285689A1
公开(公告)日：2018-10-04
申请号：US15474027
申请日：2017-03-30
发明人： Xue Mei , Danil V. Prokhorov , Haibin Ling
CPC分类号： G06N3/08 , G06K9/00791 , G06K9/6273 , G06K9/6289 , G06N3/0445 , G06N3/0454 , H04N7/183
摘要： Described herein are systems and methods for multimodal recurrent network processing. In an embodiment, a system for evaluating multimodal data comprising a multimodal data input and a multimodal processing module is described. The multimodal data input may comprise the multimodal data, the multimodal data may comprise a first modality and a second modality. The multimodal processing module may be configured to receive the multimodal data comprising the first modality and the second modality; evaluate the first modality using a first recursive neural network comprising a first transformation matrix; evaluate the second modality using a second recursive neural network comprising the first transformation matrix; and determine an output based, at least in part, on evaluating the first modality and the second modality.
-
公开(公告)号：US10078335B2
公开(公告)日：2018-09-18
申请号：US15195744
申请日：2016-06-28
发明人： Xue Mei , Katsuhiro Sakai , Nobuhide Kamata
IPC分类号： G05D1/02
CPC分类号： G05D1/0248 , G01S7/003 , G01S7/4802 , G01S17/026 , G01S17/42 , G01S17/936 , G05D1/024 , G05D1/0274 , G05D2201/0213
摘要： Ray tracing can be used to detect hidden obstacles in an external environment of a vehicle. An outbound sensor signal can be transmitted into an external environment of the vehicle. The outbound sensor signal can be a LIDAR sensor signal. If a return sensor signal is not received for the outbound sensor signal, it can be determined whether an obstacle is located along a projected path of the outbound sensor signal. Such a determination can be made using one or more maps, such as a terrain map and/or a static obstacle map. Responsive to determining that an obstacle is located along the projected path of the outbound sensor signal, a driving maneuver for the vehicle relative to the obstacle can be determined. The vehicle can be caused to implement the determined driving maneuver.
-
公开(公告)号：US09915951B2
公开(公告)日：2018-03-13
申请号：US14979462
申请日：2015-12-27
发明人： Xue Mei , Katsuhiro Sakai , Nobuhide Kamata
CPC分类号： G05D1/0257 , B60W30/00 , G01S13/931 , G01S17/936 , G01S2013/9342 , G05D1/0088
摘要： An autonomous vehicle can encounter an external environment in which an object overhangs a current road of the autonomous vehicle. For example, the branch of a tree may overhang the road. Such an overhanging object can be detected and suitable driving maneuvers for the autonomous vehicle can be determined. Sensor data can be acquired from at least a forward portion of the external environment. One or more floating obstacle candidates can be identified based on the acquired sensor data. The identified one or more floating obstacle candidates can be filtered to remove any floating obstacle candidates that do not meet one or more predefined parameters. A driving maneuver for the autonomous vehicle can be determined at least partially based on a height clearance between the autonomous vehicle and floating obstacle candidates that remain after being filtered out. The autonomous vehicle can be caused to implement the determined driving maneuver.
-
公开(公告)号：US20170371338A1
公开(公告)日：2017-12-28
申请号：US15195770
申请日：2016-06-28
发明人： Nobuhide Kamata , Xue Mei , Katsuhiro Sakai
CPC分类号： G05D1/0214 , G01S13/931 , G01S15/93 , G01S15/931 , G01S17/026 , G01S17/58 , G01S17/93 , G01S17/936 , G01S2013/9364 , G01S2013/9367 , G05D1/0088 , G05D1/024 , G05D1/0255 , G05D1/0257 , G05D1/0274
摘要： Ray tracing and static obstacle maps can be used in the operation of a vehicle. Sensor data of at least a portion of an external environment of the vehicle can be acquired. A dynamic obstacle in the external environment of the vehicle can be detected based on the acquired sensor data. In response to detecting a dynamic obstacle, it can be determined whether a secondary occluded area is located behind the dynamic obstacle relative to a current location of the vehicle based on a static obstacle map. Responsive to determining that a secondary occluded area is located behind the dynamic obstacle relative to a current location of the vehicle based on a static obstacle map, a driving maneuver for the vehicle can be determined based on at least the dynamic obstacle and the secondary occluded area. The vehicle can be caused to implement the determined driving maneuver.
-
公开(公告)号：US20170185089A1
公开(公告)日：2017-06-29
申请号：US14979462
申请日：2015-12-27
发明人： Xue Mei , Katsuhiro Sakai , Nobuhide Kamata
CPC分类号： G05D1/0257 , B60W30/00 , G01S13/931 , G01S17/936 , G01S2013/9342 , G05D1/0088
摘要： An autonomous vehicle can encounter an external environment in which an object overhangs a current road of the autonomous vehicle. For example, the branch of a tree may overhang the road. Such an overhanging object can be detected and suitable driving maneuvers for the autonomous vehicle can be determined. Sensor data can be acquired from at least a forward portion of the external environment. One or more floating obstacle candidates can be identified based on the acquired sensor data. The identified one or more floating obstacle candidates can be filtered to remove any floating obstacle candidates that do not meet one or more predefined parameters. A driving maneuver for the autonomous vehicle can be determined at least partially based on a height clearance between the autonomous vehicle and floating obstacle candidates that remain after being filtered out. The autonomous vehicle can be caused to implement the determined driving maneuver.
-
公开(公告)号：US20160272200A1
公开(公告)日：2016-09-22
申请号：US14662927
申请日：2015-03-19
发明人： Danil V. Prokhorov , Xue Mei
CPC分类号： B60W30/09 , B60W10/18 , B60W10/20 , B60W40/04 , B60W2550/30 , B60W2710/18 , B60W2710/20 , G05D1/0088 , G05D1/0246 , G05D1/0293 , G05D2201/0213 , G06K9/00805
摘要： Arrangements related to the detection of objects in an external environment of a vehicle are presented. At least a portion of the external environment can be sensed to detect a first order object therein. It can be determined whether the first order object includes a translucent portion. Responsive to determining that the first order object includes a translucent portion, the translucent portion can be analyzed to determine whether a second order object is located within the translucent portion. Responsive to determining that a second order object is located within the translucent portion, the second order object can be analyzed. Information about the second order object can be presented to a vehicle occupant, an alert regarding the second order object can be presented to a vehicle occupant, and/or, when the vehicle is an autonomous vehicle, a driving maneuver for the autonomous vehicle can be determined.
摘要翻译： 介绍与车辆外部环境中物体检测相关的安排。 可以感测外部环境的至少一部分以检测其中的一阶物体。 可以确定一阶对象是否包括半透明部分。 响应于确定一阶对象包括半透明部分，可以分析半透明部分以确定二阶对象是否位于半透明部分内。 响应于确定二阶对象位于半透明部分内，可以分析二阶对象。 关于二阶对象的信息可以呈现给车辆乘客，关于二阶对象的警报可以呈现给车辆乘客，和/或当车辆是自主车辆时，自主车辆的驾驶操纵可以是 决心。
-
公开(公告)号：US20150235092A1
公开(公告)日：2015-08-20
申请号：US14180620
申请日：2014-02-14
发明人： Xue Mei , Danil Prokhorov
CPC分类号： G06K9/00805 , G06K9/3258 , G06K9/6203 , G06K2209/15 , G06T7/246 , G06T7/277 , G06T2207/10016 , G06T2207/20021 , G06T2207/20072 , G06T2207/20081 , G06T2207/30252
摘要： The method and apparatus segment parts from a main target image, represent each part as a vertex in a spanning tree, use a detector to generate a confidence map of a location of each part in a succeeding video frame and apply scale change to detector sliding windows centered about each pixel in the part location image. In the succeeding video frame, the target location is sampled and a tracking probability is generated for each part bounding box, with the tracking probability having the maximum value being selected as the location of the target in the succeeding video frame.
摘要翻译： 从主要目标图像的方法和装置分割部分将每个部分表示为生成树中的顶点，使用检测器生成后续视频帧中每个部分的位置的置信度图，并将尺度变化应用于检测器滑动窗口 以部件位置图像中的每个像素为中心。 在接下来的视频帧中，对目标位置进行采样，并且为每个部分边界框生成跟踪概率，其中跟踪概率具有最大值作为目标在后续视频帧中的位置。
-
公开(公告)号：US10137890B2
公开(公告)日：2018-11-27
申请号：US15195760
申请日：2016-06-28
发明人： Katsuhiro Sakai , Xue Mei , Nobuhide Kamata
摘要： Obstacles located in an external environment of a vehicle can be classified. At least a portion of the external environment can be sensed using one or more sensors to acquire sensor data. An obstacle candidate can be identified based on the acquired sensor data. An occlusion status for the identified obstacle candidate can be determined. The occlusion status can be a ratio of acquired sensor data for the obstacle candidate that is occluded to all acquired sensor data for the obstacle candidate. A classification for the obstacle candidate can be determined based on the determined occlusion status. A driving maneuver for the vehicle can be determined at least partially based on the determined classification for the obstacle candidate. The vehicle can be caused to implement the determined driving maneuver. The vehicle can be an autonomous vehicle.
-
公开(公告)号：US20170031015A1
公开(公告)日：2017-02-02
申请号：US14813429
申请日：2015-07-30
发明人： Xue Mei , Naoki Nagasaka , Bunyo Okumura
CPC分类号： G01S13/865 , B60W30/0956 , G01S13/42 , G01S13/58 , G01S13/931 , G01S17/42 , G01S17/936
摘要： Minimizing incorrect associations of sensor data for an autonomous vehicle are described. A driving environment of the autonomous vehicle includes a stationary object and a dynamic object. Such objects can be detected by radar sensors and/or lidar sensors. In one example, a history of radar observation can be used to minimize incorrect sensor data associations. In such case, the location of a stationary object in the driving environment can be determined. When a dynamic object passes by the stationary object, lidar data of the dynamic object is prevented from being associated with radar data obtained substantially at the determined location of the stationary object. In another example, identifiers assigned to radar data can be used to minimize incorrect sensor data associations. In such case, lidar data of an object can be associated with radar data having a particular identifier.
摘要翻译： 描述了对于自主车辆的传感器数据的不正确关联的最小化。 自主车辆的驾驶环境包括固定物体和动态物体。 雷达传感器和/或激光雷达传感器可以检测这些物体。 在一个示例中，可以使用雷达观测的历史来最小化不正确的传感器数据关联。 在这种情况下，可以确定静止物体在驾驶环境中的位置。 当动态对象通过固定物体时，动态对象的激光雷达数据被防止与基本上在静止物体的确定位置处获得的雷达数据相关联。 在另一个示例中，分配给雷达数据的标识符可以用于最小化不正确的传感器数据关联。 在这种情况下，物体的激光雷达数据可以与具有特定标识符的雷达数据相关联。
-
公开(公告)号：US20170015313A1
公开(公告)日：2017-01-19
申请号：US14801019
申请日：2015-07-16
发明人： Xue Mei , Naoki Nagasaka , Bunyo Okumura
CPC分类号： B60W30/09 , B60W10/04 , B60W10/20 , B60W30/12 , B60W30/16 , B60W2420/52 , B60W2550/10 , B60W2550/30 , B60W2710/20 , B60W2720/10 , G05D1/0088
摘要： Systems and methods relating to the operation of an autonomous vehicle relative to forward objects in an external environment are described. At least a forward portion of the external environment can be sensed to detect an object therein. A size adjustment factor can be determined to predict a laterally innermost point of the detected object relative to the autonomous vehicle. A driving maneuver for the autonomous vehicle can be determined based at least partially on the predicted laterally innermost point of the detected object relative to the autonomous vehicle.
摘要翻译： 描述与自主车辆相对于外部环境中的前进对象的操作相关的系统和方法。 可以感测外部环境的至少前部以检测其中的物体。 可以确定尺寸调节因子以相对于自主车辆预测检测对象的横向最内点。 至少部分地基于被检测物体相对于自主车辆的预测的横向最内点来确定用于自主车辆的驾驶操纵。
-
-
-
-
-
-
-
-
-
搜索结果
34 条记录 (耗时 0.03 秒)
