**对比文件名称：** US2019236955A1_Description_20251231_1937
**目标专利名称：** 用于车辆的消息广播
**本次调用模型名称：** DeepSeek
### 特征比对表格
| 技术特征描述及公开性判断结果 | 对比文件原文引用 | 公开性论述 |
| **技术特征A：一种控制车辆的方法**<br>**判断：直接公开** | [0008] “a method for accessing supplemental data from other vehicles includes, at an autonomous vehicle: ... selecting a first navigational action based on objects in the scene represented by the composite scan image in Block S 160 ; and autonomously executing the first navigational action in Block S 170 .”<br><br>[0025] “Accordingly, the autonomous vehicle can: elect a next navigational action... and then manipulate actuators within the vehicle (e.g., accelerator, brake, and steering actuators) in order to autonomously execute the elected navigational action.” | 对比文件明确描述了一种由自动驾驶车辆执行的方法，该方法包括基于场景信息选择导航动作并自主执行该动作以控制车辆。这与目标专利中“控制车辆的方法”这一上位概念相同，均涉及对车辆行驶行为的决策与控制。因此，该特征被直接公开。 |
| **技术特征B：接收包括用于发送车辆的运动规划的意图消息**<br>**判断：隐含公开** | [0040] “the autonomous vehicle can receive a query from a second vehicle for perception data representing objects within a particular ground area of interest at approximately the particular time.”<br><br>[0045] “Therefore, in this variation, the autonomous vehicle can transmit a query to the second vehicle directly via a local ad hoc wireless network and then receive supplemental perception data from this second vehicle via the local ad hoc wireless network.”<br><br>[0078] “The autonomous vehicle may then receive—from a second vehicle—supplemental perception data in the form of a second object map segment: generated by the second vehicle ... and specifying types, locations, velocities, and/or predicted trajectories of a second set of objects...” | 对比文件多次描述了车辆之间通过无线网络请求和接收“补充感知数据(supplemental perception data)”。这些数据包含了其他车辆检测到的对象的类型、位置、速度和/或**预测轨迹(predicted trajectories)**。预测轨迹反映了发送车辆对其周围对象未来运动路径的规划或预测，这与“运动规划”的核心内涵（对车辆未来位置变化的指示）一致。虽然对比文件未使用“意图消息(intention message)”这一术语，但其描述的包含预测轨迹等信息的感知数据在车辆间传输，本质上实现了与“意图消息”相同的功能，即传递关于车辆未来运动的信息。因此，本领域技术人员可以合理推断出对比文件隐含公开了“接收包括用于发送车辆的运动规划的意图消息”这一特征。 |
| **技术特征C：其中，所述运动规划包括所述发送车辆的轨迹以及与所述发送车辆相关联的一个或多个车辆描述符，并且其中，所述一个或多个车辆描述符包括车辆物理能力，其中，所述车辆物理能力包括转弯半径、车辆最高速度和车辆最大加速度中的至少一项**<br>**判断：隐含公开** | [0030] “Therefore, the autonomous vehicle can: isolate features in the current scan image that differ from immutable objects and surfaces represented in the localization map, since these differing features may (predominantly) represent mutable objects in the scene around the autonomous vehicle; and selectively transmit perception data representing these mutable objects to other vehicles nearby...”<br><br>[0031] “The autonomous vehicle can thus generate a 3D scan image that contains georeferenced representations of mutable and immutable objects, each labeled with its predicted type, state, 2D or 3D boundary, velocity, and/or predicted trajectory, etc.”<br><br>[0078] “...specifying types, locations, velocities, and/or predicted trajectories of a second set of objects...” | 对比文件公开了车辆间传输的感知数据包含对象的**预测轨迹(predicted trajectory)**，这对应于“运动规划”中的“轨迹”。同时，数据中还包含对象的**类型(type)**（如车辆、卡车、行人）。车辆的类型信息（如“卡车”与“轿车”）本身隐含了与该类型相关的典型物理能力差异，例如重量、尺寸、惯性和常见的性能参数范围（如卡车通常比轿车转弯半径大、最高速度低）。虽然对比文件未明确列举“转弯半径、车辆最高速度和车辆最大加速度”这些具体参数，但本领域技术人员知晓，在自动驾驶的感知和预测模型中，对象的类型是推断其物理能力和运动约束的关键先验信息。通过共享对象类型，接收方能够基于常识或数据库关联出该类型车辆的典型物理能力，从而用于行为预测。因此，可以合理推断对比文件隐含公开了运动规划包括轨迹和隐含车辆物理能力的车辆描述符（类型）。 |
| **技术特征D：对所述意图消息进行解析以识别用于所述发送车辆的所述运动规划**<br>**判断：直接公开** | [0045] “Upon receipt of this request, the second vehicle can execute methods and techniques described above to: isolate raw sensor data or derived perception data that temporally and spatially intersects this request...”<br><br>[0077] “Upon receipt of supplemental perception data, an autonomous vehicle can thus align these supplemental perception data to its own sensor data recorded locally at approximately the same time... based on geolocation tags in these supplemental perception data to form a composite sensor image.” | 对比文件明确描述了车辆在接收到来自其他车辆的感知数据（即隐含的意图消息）后，会对其进行处理，包括根据地理标签进行对齐、提取与特定时空请求相关的数据段。这个过程本质上就是对接收到的消息进行解析，以识别出其中包含的关于其他对象（包括车辆）的轨迹等信息。因此，该特征被直接公开。 |
| **技术特征E：以及至少部分地基于所述运动规划来控制所述车辆，包括：至少部分地基于由所述车辆物理能力限制的所述发送车辆的车辆行为模型，来设置针对所述发送车辆的行为预测**<br>**判断：隐含公开** | [0070] “After classifying the object, the autonomous vehicle can also access a motion model associated with the type of the object and predict a trajectory of this object based on the motion model and geospatial locations of this object... in preceding scan images.”<br><br>[0071] “The autonomous vehicle can then implement a path-planning model to elect a next navigational action that limits or reduces probability of collision with the object based on this predicted trajectory of the object.”<br><br>[0088] “The autonomous vehicle can also: detect and track an object in this sequence of past composite scan images; estimate a trajectory of the object based on changes in its geospatial locations during this sequence of previous composite scan images; ... The autonomous vehicle can then autonomously elect and execute a next navigational action based on the scene, objects in the scene, and predicted trajectories of these objects...” | 对比文件详细描述了自动驾驶车辆如何基于接收到的或自身感知的对象信息（包括类型和轨迹）来预测对象的轨迹，并基于此预测来选择和执行自身的导航动作以控制车辆。这对应于“至少部分地基于所述运动规划来控制所述车辆”。其中，预测对象轨迹时，会使用与对象**类型(type)**相关联的**运动模型(motion model)**。如前所述，对象类型隐含了其物理能力，而运动模型正是用于描述和预测特定类型对象行为的模型，其参数和约束自然受到该类型对象物理能力的限制。因此，基于对象类型（隐含物理能力）的运动模型来预测其行为（轨迹），实质上就是“至少部分地基于由所述车辆物理能力限制的所述发送车辆的车辆行为模型，来设置针对所述发送车辆的行为预测”。该技术特征被隐含公开。 |
| **技术特征F：所述一个或多个车辆描述符进一步包括传感器可感知属性**<br>**判断：直接公开** | [0030] “...represent mutable objects in the scene around the autonomous vehicle...”<br><br>[0031] “...each labeled with its predicted type, state, 2D or 3D boundary, velocity, and/or predicted trajectory, etc.”<br><br>[0079] “...specifying types, locations, velocities, and/or predicted trajectories of a second set of objects...” | 对比文件公开了共享的感知数据中包含对象的**类型(type)**，例如车辆、行人等。对象的类型（如轿车、卡车）是其通过传感器（如摄像头、激光雷达）可以感知和分类的属性。因此，“传感器可感知属性”这一特征被对比文件直接公开。 |
| **技术特征G：至少部分地基于所述运动规划来控制所述车辆包括：至少部分地基于所述运动规划来确定用于所述车辆的预期感兴趣区域；以及将检测算法应用于在所述预期感兴趣区域处的接收的传感器数据，以至少部分地基于所述传感器可感知属性来在所述接收的传感器数据中检测所述发送车辆**<br>**判断：隐含公开** | [0008] “...defining a ground area of interest containing the region of the scan image in Block S 130...”<br><br>[0010] “For example, the autonomous vehicle can: generate a query for raw sensor data, an object map, and/or other perception data representing objects within a particular ground area of interest...”<br><br>[0066] “...defines a ground area of interest that intersects this region and extends beyond the external object opposite the autonomous vehicle in Block S 130 ; and then broadcasts a request for supplemental perception data that represents objects and surfaces within this ground area of interest...” | 对比文件的核心内容之一就是车辆根据自身感知的不足（如视野被遮挡、数据不足），定义一个**地面感兴趣区域(ground area of interest)**，并向其他车辆请求该区域内的感知数据。虽然触发定义感兴趣区域的原因是自身感知不足，但请求到的数据中包含其他车辆的预测轨迹（即运动规划）。接收车辆在得到这些数据后，会将其与自身传感器数据融合。本领域技术人员可以合理推断，为了有效融合数据，接收车辆需要知道在何处（即哪个区域）寻找发送车辆。发送车辆的运动规划（预测轨迹）正好指明了其未来可能出现的位置区域，这自然可以用于确定一个“预期感兴趣区域”，以便接收车辆将感知资源（检测算法）聚焦于该区域，并利用已知的传感器可感知属性（如车辆类型）在该区域的传感器数据中更有效地检测发送车辆。这是一个合乎逻辑的技术延伸。因此，该特征被隐含公开。 |
| **技术特征H：至少部分地基于所接收的运动规划来选择所述检测算法**<br>**判断：隐含公开** | [0012] “As shown in FIG. 3, the autonomous vehicle can also request perception data of a particular type, such as: raw depth map data ... or object maps ... if high-bandwidth wireless communications are not supported by the local wireless network...”<br><br>[0080] “...wirelessly broadcast—via the wireless network—a query for derived perception data including types and locations of objects within a ground area of interest (e.g., an object map) at approximately the first time if wireless connectivity is poor and/or if bandwidth of the local wireless network falls below a threshold bandwidth at approximately the first time. However, if wireless connectivity is good and/or if bandwidth of the local wireless network exceeds the threshold bandwidth at approximately the first time, the autonomous vehicle can instead broadcast a query for raw sensor data...” | 对比文件描述了车辆根据无线网络带宽条件，选择请求不同类型的数据（原始数据或处理后的对象地图）。虽然这里直接描述的是“请求”阶段的选择，但接收到的数据类型（原始点云或结构化对象数据）直接影响接收方后续处理（检测、融合）所需采用的算法复杂度、输入格式和处理逻辑。例如，收到原始点云数据需要运行完整的对象检测和分类算法，而收到已标注类型的对象地图则可能只需进行数据关联和跟踪。本领域技术人员可以理解，基于接收到的信息（其形式和内容部分由运动规划等数据构成）来选择或适配处理算法，是优化系统性能的常规技术手段。因此，可以合理推断“至少部分地基于所接收的运动规划来选择所述检测算法”这一特征被隐含公开。 |
| **技术特征I：至少部分地基于所述运动规划来控制所述车辆包括：至少部分地基于所述传感器可感知属性来将车辆检测传感器数据与所述发送车辆进行关联**<br>**判断：直接公开** | [0077] “Upon receipt of supplemental perception data, an autonomous vehicle can thus align these supplemental perception data to its own sensor data recorded locally at approximately the same time (e.g., within two seconds of recordation of the supplemental perception data) based on geolocation tags in these supplemental perception data to form a composite sensor image.”<br><br>[0078] “The autonomous vehicle can then insert the second object map segment received from the second vehicle into the first object map generated by the autonomous vehicle to form a composite scan image depicting both objects detected by the second vehicle but not the autonomous vehicle and objects detected by the autonomous vehicle itself.” | 对比文件明确描述了车辆将接收到的补充感知数据与自身本地记录的传感器数据进行对齐(algin)和组合(incorporate, insert)，以形成复合扫描图像。补充感知数据中包含了对象的类型（传感器可感知属性）和位置等信息。这个对齐和组合的过程，本质上就是基于共享数据中的属性（如类型、位置）将来自不同源的、对同一对象的检测数据进行关联和融合。因此，“至少部分地基于所述传感器可感知属性来将车辆检测传感器数据与所述发送车辆进行关联”这一特征被直接公开。 |
| **技术特征J：确定所述发送车辆的行为是否符合所述行为预测；以及响应于确定所述发送车辆的所述行为不符合所述行为预测，至少部分地基于所述车辆物理能力来更新所述行为预测**<br>**判断：隐含公开** | [0088] “The autonomous vehicle can also: detect and track an object in this sequence of past composite scan images; estimate a trajectory of the object based on changes in its geospatial locations during this sequence of previous composite scan images; verify this estimated trajectory based on a difference between the predicted and actual geospatial location of the object during the current scan cycle; and then extrapolate this trajectory forward into the future if validated (e.g., if the difference between the predicted and actual geospatial location of the object is less than a threshold difference).” | 对比文件描述了自动驾驶车辆跟踪对象、估计其轨迹、并根据预测轨迹与实际位置的差异来验证(verify)该估计轨迹。如果验证通过（差异小于阈值），则继续外推该轨迹。这个过程隐含了“确定对象行为是否符合预测”的逻辑。虽然原文未明确描述当预测不符合时如何“更新”预测，但本领域技术人员知晓，在自动驾驶系统中，当对象行为偏离预测时，更新预测是维持系统可靠性的必要步骤。更新预测时，对象的类型（隐含物理能力）是约束其可能行为范围（如最大加速度、转弯能力）的关键因素。例如，如果一辆轿车突然偏离车道，基于其类型（轿车）所隐含的物理能力，系统会预测其可能的紧急变道或失控模式，而不会预测其像卡车一样进行急转弯。因此，可以合理推断该特征被隐含公开。 |
| **技术特征K：所述一个或多个车辆描述符进一步包括车辆位置属性**<br>**判断：直接公开** | [0031] “...georeferenced representations of mutable and immutable objects...”<br><br>[0078] “...specifying types, locations, velocities, and/or predicted trajectories of a second set of objects...”<br><br>[0077] “...based on geolocation tags in these supplemental perception data...” | 对比文件多次明确提及共享的感知数据是**地理参考的(georeferenced)** 或包含对象的**位置(locations)**，并提及数据中包含**地理位置标签(geolocation tags)**。这直接对应于“车辆位置属性”。因此，该特征被直接公开。 |
| **技术特征L：至少部分地基于所述运动规划来控制所述车辆包括：至少部分地基于所述车辆位置属性来确定所述发送车辆的位置；确定在所述车辆的位置与所述发送车辆的位置之间的比较是否指示误差；以及响应于确定在所述车辆的位置与所述发送车辆的位置之间的所述比较指示误差，来触发对所述车辆的位置的重新计算**<br>**判断：隐含公开** | [0025] “The autonomous vehicle can also compare features detected in this current scan image to a localization map to determine the position and orientation of the autonomous vehicle in real space.”<br><br>[0081] “In one variation, the autonomous vehicle requests supplemental perception data that extends beyond a region of a scan image for which the autonomous vehicle has access to insufficient perception data. Upon receipt, the autonomous vehicle: verifies presence and alignment of known objects in a high-data-density section of the scan image with like objects represented in a first segment of these supplemental perception data; and then incorporates a second segment of these perception data... responsive to verification of the first segment of the supplemental perception data.” | 对比文件描述了车辆通过将感知到的特征与定位地图进行比较来确定自身位置（定位）。同时，也描述了在收到其他车辆的补充感知数据后，会先验证这些数据与自身已知对象的一致性（验证和对齐）。本领域技术人员可以理解，这种验证过程本质上包含了比较自身感知到的对象位置与接收到的该对象位置信息。如果发现显著误差，这可能暗示自身定位存在偏差。在自动驾驶定位技术中，利用来自其他可靠源（如其他车辆提供的精确地理参考数据）的位置信息来校正自身定位，是一种常见的技术思路，例如协同定位。因此，可以合理推断，基于接收到的车辆位置属性（来自运动规划）与自身感知的车辆位置进行比较，并在指示误差时触发自身位置的重新计算，是该技术领域的合理应用和延伸。该特征被隐含公开。 |
| **技术特征M：至少部分地基于所述运动规划来控制所述车辆包括：确定所述运动规划是否是不安全的；以及响应于确定所述运动规划是不安全的，来向所述发送车辆发送安全警告**<br>**判断：隐含公开** | [0071] “The autonomous vehicle can then implement a path-planning model to elect a next navigational action that limits or reduces probability of collision with the object based on this predicted trajectory of the object.”<br><br>[0088] “...elect and execute a next navigational action based on the scene, objects in the scene, and predicted trajectories of these objects, such as to reduce probability of future collision with these objects.” | 对比文件的核心目的之一是帮助车辆基于更完整的场景信息（包括其他对象的预测轨迹）来规划路径，以降低碰撞概率。这意味着车辆能够评估场景中对象预测轨迹（即运动规划）的安全性。虽然对比文件未明确描述向发送轨迹的车辆“发送安全警告”，但其描述了车辆间可以互相请求和发送感知数据，建立了双向通信链路。本领域技术人员可以合理推断，当一辆车基于其更全面的感知信息（例如，看到了发送车辆视野外的危险）判断发送车辆的运动规划（预测轨迹）将导致碰撞时，通过已有的通信链路向该发送车辆发送一个警告或提示信息，是提高整体道路安全的直接且合理的技术应用。这是对已公开通信功能和安全性评估功能的合乎逻辑的扩展。因此，该特征被隐含公开。 |
<<<A>>><<<D>>><<<F>>><<<I>>><<<K>>>