Abstract: A high-definition map system receives sensor data from vehicles travelling along routes and combines the data to generate a high definition map for use in driving vehicles, for example, for guiding autonomous vehicles. A pose graph is built from the collected data, each pose representing location and orientation of a vehicle. The pose graph is optimized to minimize constraints between poses. Points associated with surface are assigned a confidence measure determined using a measure of hardness/softness of the surface. A machine-learning-based result filter detects bad alignment results and prevents them from being entered in the subsequent global pose optimization. The alignment framework is parallelizable for execution using a parallel/distributed architecture. Alignment hot spots are detected for further verification and improvement. The system supports incremental updates, thereby allowing refinements of subgraphs for incrementally improving the high-definition map for keeping it up to date.

Abstract: A system performs calibration of sensors mounted on a vehicle, for example, lidar and camera sensors mounted on a vehicle, for example, an autonomous vehicle. The system receives a lidar scan and camera image of a view and determines a lidar-to-camera transform based on the lidar scan and the camera image. The system may use a pattern, for example, a checkerboard pattern in the view for calibration. The pattern is placed close to the vehicle to determine an approximate lidar-to-camera transform and then placed at a distance from the vehicle to determine an accurate lidar-to-camera transform. Alternatively, the system determines edges in the lidar scan and the camera image and aligns features based on real-world objects in the scene by comparing edges.

Abstract: A vehicle, for example, an autonomous vehicle receives signals from a global navigation satellite system (GNSS) and determines accurate location of the vehicle using the GNSS signal. The vehicle performs localization to determine the location of the vehicle as it drives. The autonomous vehicle uses sensor data and a high definition map to determine an accurate location of the autonomous vehicle. The autonomous vehicle uses accurate location of the vehicle to determine RTK corrections that is used for improving GNSS location estimates at a future location. The RTK corrections may be transmitted to other vehicles.

Abstract: A system performs rolling shutter correction to transform data captured by sensors of a vehicle, for example, cameras or lidar mounted on a vehicle, for example, an autonomous vehicle. The images captured by a rolling shutter camera mounted on a moving vehicle show rolling shutter distortion. The rolling shutter correction transforms the data representing points of scenes to perform rolling shutter compensation, i.e., compensation for the rolling shutter distortion. The rolling shutter compensation ensures that data representing as three dimensional points, for example, data captured by lidar is consistent with data represented in images captured by a rolling shutter camera. The system performs rolling shutter compensation by estimating a distance travelled by the vehicle between the time that the camera captured a point and the time that the image scan was completed and translating the 3D points by the estimated distance.

Abstract: A system generates a high definition map for an autonomous vehicle to travel from a source location to a destination location. The system determines a low resolution route and receives high definition map data for a set of geographical regions overlaying the low resolution route. The system uses lane elements within the geographical regions to form a set of potential partial routes. The system calculates the error between the potential partial route and the low resolution route and removes potential partial routes with errors above the threshold. Once completed, the system selects a final route and sends signals to the controls of the autonomous vehicle to follow the final route. The system determines whether surface areas adjacent to a lane that are not part of the road are safe for the vehicle to drive in case of emergency. The system stores information describing navigable surface areas with representations of lanes.

Abstract: A method and system for synchronizing a lidar and a camera on an autonomous vehicle. The system selects a plurality of track samples for a route including a lidar scan and an image. For each track sample, the system calculates a time shift by iterating many time deltas. For each time delta, the system adjusts a camera timestamp by that time delta, projects a lidar scan onto the image as a lidar projection according to the adjusted camera timestamp, and calculates an alignment score of the lidar projection for that time delta. The system defines the time shift for each track sample as the time delta with the highest alignment score. The system then models time drift of the camera compared to the lidar based on the calculated time shifts for the track samples and synchronizes the lidar and the camera according to the modeled time drift.

Abstract: A system calibrates one or more sensors mounted to an autonomous vehicle. From the one or more sensors, the system identifies a primary sensor and a secondary sensor. The system determines a reference angle for the primary sensor, and based on that reference angle for the primary sensor, a scan-start time representing a start of a scan and a scan-end time representing an end of a scan. The system receives, from the primary sensor, a primary set of scan data recorded from the scan-start time to the scan-end time. The system receives, from the secondary sensor, a secondary set of sensor data recorded from the scan-start time to the scan-end time. The system calibrates the primary and secondary sensors by determining a relative transform for transforming points between the first set of scan data and the second set of scan data.

Abstract: An online system builds a high definition (HD) map for a geographical region based on sensor data captured by a plurality of autonomous vehicles driving through a geographical region. The autonomous vehicles detect map discrepancies based on differences in the surroundings observed using sensor data compared to the high definition map and send messages describing these map discrepancies to the online system. The online system ranks the autonomous vehicles based on factors including an upload rate indicating how often the vehicle was used providing data to the online system. The sensor data from vehicles is uploaded to the online system (e.g., in the cloud) to create the HD map while spreading the burden of uploading this data as evenly as possible across a fleet of vehicles. Data uploads are expensive and time consuming, so the system makes this negligible for each vehicle by balancing/managing the uploads carefully across the fleet.

Abstract: A vehicle computing system performs enhances relatively sparse data collected by a LiDAR sensor by increasing the density of points in certain portions of the scan. For instance, the system generates 3D triangles based on a point cloud collected by the LiDAR sensor and filters the 3D triangles to identify a subset of 3D triangles that are proximate to the ground. The system interpolates points within the subset of 3D triangles to identify additional points on the ground. As another example, the system uses data collected by the LiDAR sensor to identify vertical structures and interpolate additional points on those vertical structures. The enhanced data can be used for a variety of applications related to autonomous vehicle navigation and HD map generation, such as detecting lane markings on the road in front of the vehicle or determining a change in the vehicle's position and orientation.

Abstract: Embodiments relate to methods for efficiently encoding sensor data captured by an autonomous vehicle and building a high definition map using the encoded sensor data. The sensor data can be LiDAR data which is expressed as multiple image representations. Image representations that include important LiDAR data undergo a lossless compression while image representations that include LiDAR data that is more error-tolerant undergo a lossy compression. Therefore, the compressed sensor data can be transmitted to an online system for building a high definition map. When building a high definition map, entities, such as road signs and road lines, are constructed such that when encoded and compressed, the high definition map consumes less storage space. The positions of entities are expressed in relation to a reference centerline in the high definition map. Therefore, each position of an entity can be expressed in fewer numerical digits in comparison to conventional methods.

Abstract: Embodiments relate to methods for efficiently encoding sensor data captured by an autonomous vehicle and building a high definition map using the encoded sensor data. The sensor data can be LiDAR data which is expressed as multiple image representations. Image representations that include important LiDAR data undergo a lossless compression while image representations that include LiDAR data that is more error-tolerant undergo a lossy compression. Therefore, the compressed sensor data can be transmitted to an online system for building a high definition map. When building a high definition map, entities, such as road signs and road lines, are constructed such that when encoded and compressed, the high definition map consumes less storage space. The positions of entities are expressed in relation to a reference centerline in the high definition map. Therefore, each position of an entity can be expressed in fewer numerical digits in comparison to conventional methods.

Abstract: High definition maps for autonomous vehicles are very high resolution and detailed, and hence require storage of a great deal of data. A vehicle computing system provides multi-layered caching makes this data usable in a system that requires very low latency on every operation. The system determines which routes are most likely to be driven in the near future by the car, and ensures that the route is cached on the vehicle before beginning the route. The system provides efficient formats for moving map data from server to car and for managing the on-care disk. The system further provides real-time accessibility of nearby map data as the car moves, while providing data access at optimal speeds.

Abstract: As an autonomous vehicle moves through a local area, pairwise alignment may be performed to calculate changes in the pose of the vehicle between different points in time. The vehicle comprises an imaging system configured to capture image frames depicting a portion of the surrounding area. Features are identified from the captured image frames, and a 3-D location is determined for each identified feature. The features of different image frames corresponding to different points in time are analyzed to determine a transformation in the pose of the vehicle during the time period between the image frames. The determined poses of the vehicle are used to generate an HD map of the local area.

Abstract: As an autonomous vehicle moves through a local area, pairwise alignment may be performed to calculate changes in the pose of the vehicle between different points in time. The vehicle comprises an imaging system configured to capture image frames depicting a portion of the surrounding area. Features are identified from the captured image frames, and a 3-D location is determined for each identified feature. The features of different image frames corresponding to different points in time are analyzed to determine a transformation in the pose of the vehicle during the time period between the image frames. The determined poses of the vehicle are used to generate an HD map of the local area.

Abstract: A high-definition map system receives sensor data from vehicles travelling along routes and combines the data to generate a high definition map for use in driving vehicles, for example, for guiding autonomous vehicles. A pose graph is built from the collected data, each pose representing location and orientation of a vehicle. The pose graph is optimized to minimize constraints between poses. Points associated with surface are assigned a confidence measure determined using a measure of hardness/softness of the surface. A machine-learning-based result filter detects bad alignment results and prevents them from being entered in the subsequent global pose optimization. The alignment framework is parallelizable for execution using a parallel/distributed architecture. Alignment hot spots are detected for further verification and improvement. The system supports incremental updates, thereby allowing refinements of subgraphs for incrementally improving the high-definition map for keeping it up to date.

Abstract: A high-definition map system receives sensor data from vehicles travelling along routes and combines the data to generate a high definition map for use in driving vehicles, for example, for guiding autonomous vehicles. A pose graph is built from the collected data, each pose representing location and orientation of a vehicle. The pose graph is optimized to minimize constraints between poses. Points associated with surface are assigned a confidence measure determined using a measure of hardness/softness of the surface. A machine-learning-based result filter detects bad alignment results and prevents them from being entered in the subsequent global pose optimization. The alignment framework is parallelizable for execution using a parallel/distributed architecture. Alignment hot spots are detected for further verification and improvement. The system supports incremental updates, thereby allowing refinements of subgraphs for incrementally improving the high-definition map for keeping it up to date.

Abstract: A high-definition map system receives sensor data from vehicles travelling along routes and combines the data to generate a high definition map for use in driving vehicles, for example, for guiding autonomous vehicles. A pose graph is built from the collected data, each pose representing location and orientation of a vehicle. The pose graph is optimized to minimize constraints between poses. Points associated with surface are assigned a confidence measure determined using a measure of hardness/softness of the surface. A machine-learning-based result filter detects bad alignment results and prevents them from being entered in the subsequent global pose optimization. The alignment framework is parallelizable for execution using a parallel/distributed architecture. Alignment hot spots are detected for further verification and improvement. The system supports incremental updates, thereby allowing refinements of sub-graphs for incrementally improving the high-definition map for keeping it up to date.