JP2023554304A - Side safety area - Google Patents

Abstract

Techniques for determining a vehicle's safe area are discussed. In some cases, the first safety area can be based on the vehicle traveling in the environment and the second safety area can be based on the steering control or speed of the vehicle. The width of the safe area may be updated based on the position of the bounding box relative to the vehicle. The location of the bounding box can be based on the vehicle traveling along the trajectory. Sensor data can be filtered based on sensor data within the safe area.

Description

The present disclosure relates to techniques for determining a safe area of a vehicle.

Cross References to Related Applications This patent application is filed under U.S. Utility Patent Application No. 17/124,220 filed on December 16, 2020 and U.S. Utility Patent Application No. 17/124 filed on December 16, 2020. , No. 237 claims priority. The contents of application numbers 17/124,220 and 17/124,237 are fully incorporated herein by reference.

Vehicles can capture sensor data to detect objects in the environment. Although sensor data is generally available for object detection, system limitations associated with processing sensor data may result in objects not being detected in rare situations. For example, before or during the capture of sensor data, turns of the vehicle due to the approach of an object to a position relative to the trajectory of the vehicle may not be processed in a timely manner. Due to delays in processing sensor data, detection of a potential collision and deployment of safety measures in a timely manner may not be possible.

U.S. Patent Application No. 16/189,726

The detailed description will be explained with reference to the accompanying drawings. In the drawings, the left-most digit of a reference number identifies the drawing in which the reference number first appears. The use of the same reference numbers in different drawings indicates similar or identical elements or features.

FIG. 2 is a pictorial flow diagram illustrating an example process for determining a safe area based on a trajectory that includes vehicle turns, according to examples of the present disclosure. FIG. 3 illustrates an environment that includes a portion of a safety area with a width determined based on the orientation of a bounding box associated with a vehicle, according to an example of the present disclosure. FIG. 2 is a pictorial flow diagram illustrating an example process for determining a segment of a safe area based on a bounding box associated with a vehicle, according to examples of the present disclosure. FIG. 1 is a block diagram illustrating an example system for implementing the techniques described herein. FIG. 3 is a diagram illustrating an environment including a boundary determined based on a trajectory of a vehicle and a safety area determined based on a line associated with the boundary, according to an example of the present disclosure. FIG. 3 is a diagram illustrating an environment including a safe area determined based on a trajectory including a vehicle left turn, according to an example of the present disclosure. FIG. 2 is a pictorial flow diagram illustrating an example process for determining a safe area based on stationary vehicles, according to examples of the present disclosure. 3 is a flowchart illustrating an example process for determining a safe area based on a trajectory that includes vehicle turns, according to examples of the present disclosure. 3 is a flowchart illustrating an example process for determining the width of a portion of a safe area based on the orientation of a bounding box as related to a vehicle, according to examples of the present disclosure.

Techniques for determining safe areas for vehicles are described herein. For example, the technique may include determining one or more safe areas based on the trajectory of the vehicle. In some cases, a safe area can be determined based on one or more of the direction of the track (eg, left or right turn), location, operation being performed, or speed of the track. In some examples, the safety area can include a first safety area created as a constant width surrounding the track, and a second safety area, where the second safety area is located at a location (e.g. , perpendicular to the intersection when passing through the intersection) or based on the turns, maneuvers, or speeds associated with the vehicle. In some cases, the safe area may be expanded based on projecting the movement of the vehicle along the trajectory and determining whether the points associated with the bounding box are outside the safe area. , otherwise it is possible to update. The sensor data can be filtered to reduce the amount of processing while evaluating the sensor data about the likelihood of collision with objects in the environment in a more accurate and reliable manner. Such safety margins may be used by safety systems associated with the vehicle. In such instances, available sensor (or other) data related to areas outside the safety area may be filtered to provide increased resources available for processing data related to the safety area. It's fine.

As discussed above, the first safe area may be determined based on a width associated with the received trajectory, while the second safe area may be determined based on aspects of the trajectory. is possible. For example, the second safety area can be determined to be an area close to and/or adjacent to the vehicle (eg, a longitudinal end of the vehicle or a side of the vehicle). In some cases, the second safety area can be determined to be an area in front of the vehicle and orthogonal to the first safety area. In some cases, the second safety area may be determined to be an area parallel to the first safety area and adjacent to the vehicle. A vehicle may receive and filter sensor data related to the environment in which the vehicle is driving. The sensor data can be filtered to determine sensor data associated with the first safe area and/or the second safe area. The sensor data related to the filtered safety area is used to determine which objects are related to the first safety area and to identify any potential objects in order to cause any safety actions (e.g., update operations, initiate an emergency stop, etc.). can be used to determine safety issues (e.g., collision or otherwise).

The width of each portion of the safety area may be determined based on information related to the vehicle and/or the environment. The information can be used to determine a bounding box associated with the vehicle. In some examples, a width associated with a portion of the safety area (eg, the first safety area) can be determined based on a bounding box (eg, a bounding box associated with the vehicle). Furthermore, the received trajectory can be discretized. One or more segments associated with a safe area (eg, a first safe area, or a second safe area) can be determined based on the discretized trajectory. In some examples, the segments of the safe area may be determined based on discretized segments associated with the trajectory. The width of each segment can be determined based on a first width (eg, a fixed width) or a second width based on a bounding box and a point related to an edge of the segment. The extent of any portion of the safety area is determined based at least in part on speed limits, sensor ranges, objects previously observed in the area, etc., to limit the amount of processing performed in the safety system. good.

The techniques described herein can enhance computing functionality in a variety of additional ways. In some cases, determining safe areas can be used to reduce the amount of data processing required to avoid potential collisions in the environment. It is possible to determine a second safety area based on the first safety area and the trajectory associated with the vehicle, and it is possible to determine objects more efficiently and accurately. By utilizing the second safe area for filtering sensor data, resources can be saved and/or reallocated to different tasks. By analyzing the width of a portion of the safe area instead of the entire safe area, it is possible to determine the probability of a collision and reduce the amount of data required for analysis. Saving resources available for vehicle control by determining the likelihood of a collision at an earlier time and by simplifying the amount and/or complexity of processing required to determine a modified acceleration profile. Is possible.

The techniques described herein can be implemented in a variety of ways. An example of implementation is provided below with reference to subsequent figures. The methods, apparatus, and systems described herein are applicable to vehicles such as, but not limited to autonomous vehicles, and are applicable to a variety of systems. In another example, the present technology can be used in an aviation or navigational context, or in any system into which data is input and configured to determine movement relative to an object in an environment. Additionally, the techniques described herein can be applied to real data (e.g., captured using a sensor), simulated data (e.g., generated by a simulator), or a third data for the two. Can be used in combination with

FIG. 1 is a pictorial flow diagram 100 illustrating an example process for determining a safe area based on a trajectory that includes vehicle turns, according to examples of the present disclosure.

Act 102 may include determining a safe area (eg, a first safe area) based on the trajectory. The first safe area can be determined based on the trajectory of the vehicle traveling through the environment. In some examples, the first safety area can be determined based at least in part on the width and/or length of the vehicle, the vehicle's current speed and/or speed relative to the trajectory, etc. be. In some examples, the maximum width of the first safety area may be the width of the vehicle and/or the width of the lane in which the vehicle is currently located. In some examples, the width of each portion of the first safety area can be the same width. In other examples, the width of each of the one or more portions of the first safety area can be expanded based on a representation of the vehicle at a position along the first safety area at a future time. It is. In some examples, the first safety area can be determined based on a fixed distance orthogonal to the trajectory. The center point of each cross-section of the first safety area can be at the same position as the point of the trajectory. The trajectory may be determined in conjunction with a turn (eg, a right turn) of the vehicle. The first safe area may be determined to be associated with a right turn based on the trajectory being associated with a right turn.

Example 104 illustrates an environment that includes a first safety area (eg, safety area 106) based on trajectory 108. Safe area 106 may be determined based on a trajectory associated with vehicle 110 traveling through the environment. In some examples, safe area 106 can be determined based on the width and/or length of vehicle 110, the current speed and/or trajectory of vehicle 110, etc. The width of each portion of the safety area 106 may be the same width. The center point of each cross-section of the safety area 106 may be at the same location as the point of the trajectory 108. Trajectory 108 may be determined in conjunction with a turn (eg, a right turn) of vehicle 110. Safe area 106 may be determined to be associated with a right turn based on trajectory 108 being associated with a right turn. In some examples, the width of the safety area 106 can be the same distance as the width of the vehicle 110. In some examples, the width distance of the safety area 106 may be greater or less than the width distance of the vehicle 110 by a threshold distance.

Act 112 may include determining a safe area (eg, a second safe area) based on the trajectory. The second safe area can be determined based on a trajectory associated with a turn (eg, a right turn). In some examples, the second safety area can be determined based at least in part on a portion of the safety area associated with the vehicle being turn controlled. In some examples, the second safe area can be determined based on a turn in the environment (eg, a turn angle associated with the turn) matching or exceeding a threshold angle. The second safety area can be determined based on information related to the environment and/or the vehicle. In some examples, the second safe area is such that the distance between the vehicle and the intersection (e.g., a line that is related to, parallel to, and overlaps the boundary of the crosswalk closest to the vehicle) is less than a threshold distance. and/or based on the speed of the vehicle being below a threshold speed. In some examples, the maximum width of the second safety area may be the width of a cross lane orthogonal to the lane in which the vehicle is currently located. Further examples of second safe areas are described throughout this disclosure.

In some examples, the orientation of the second safe area can be based on the orientation of the first safe area. For example, the direction of the second safe area can be determined based on a line tangent to a portion of the first safe area. The second safety area can be substantially orthogonal to the first safety area. In some cases, the second safety area can be adjacent to (eg, in contact with) the vehicle. For example, the location of a portion of the boundary of the second safety area on the side of the vehicle may be the location of a portion of the longitudinal end of the vehicle (e.g., the longitudinal end closest to the location where vehicle movement is controlled). It is possible to relate. In some examples, a portion of the longitudinal end of the vehicle may be a component of the vehicle (eg, a bumper). The part of the boundary (e.g. the near side boundary (e.g. the boundary at the side of the second safety area closest to the part of the longitudinal end) may be the point of the boundary closest to the vehicle The portion of the longitudinal end of the vehicle may be the point closest to the longitudinal end boundary.

In some cases, the second safety area can be spaced apart from (eg, not adjacent to) a longitudinal end of the vehicle. For example, the distance between the second safety area and the longitudinal end can be determined to be greater than or equal to a threshold distance (eg, 1 meter, 10 meters, 30 meters, 100 meters, etc.). In some cases, the proximal boundary of the second safety area can be determined before the longitudinal end of the vehicle. In some cases, the proximal boundary of the second safety area can be determined inside the vehicle.

In some examples, the second safety area is fixed relative to the vehicle such that the position of the second safety area is updated as the vehicle travels through the environment. In some examples, the distance between the second safety area and the vehicle is variable (e.g., as the distance between the vehicle and the intersection falls below a threshold distance, or as the speed of the vehicle falls below a threshold speed, the distance increases). can be reduced).

In some examples, the first safe area may be determined based on the trajectory, so that the first safe area may be considered a drivable area. However, in some examples, the second safe area may include drivable areas as well, but may include non-drivable areas such as sidewalks, buildings, etc.

Example 114 shows a second safety area (eg, safety area 116) based on the trajectory. Safe area 116 may be determined based on a trajectory associated with a turn (eg, a right turn). In some examples, safe area 116 may be determined based on a turn in the environment (eg, a turn angle associated with the turn) meeting or exceeding a threshold turn angle. In some examples, the maximum width of safety area 116 may be the width of a cross lane orthogonal to the lane in which the vehicle is currently located.

In various examples described herein, the second safe area may be defined based at least in part on map data available to the vehicle. In such instances, widths, lengths, or other extents may be bounded based on geometric data associated with the map and/or parameters associated therewith (e.g., speed limits, crosswalks, etc.). It's fine.

Act 118 may include determining that the object is associated with the second safety area. In some cases, the object may be determined based on sensor data received from one or more sensors of the vehicle. For example, sensor data can include data from multiple types of sensors in a vehicle, such as light detection and ranging (LIDAR) sensors, RADAR sensors, image sensors, depth sensors (time-of-flight, structured light, etc.) It is. Sensor data utilized to determine objects in an environment can be filtered as filtered sensor data (eg, a subset of sensor data). The subset of sensor data may include data related to the first safety area and/or the second safety area. The object may be determined relative to the first safe area and/or the second safe area based on the subset of sensor data. The likelihood that the object will traverse the trajectory (eg, the likelihood of crossing) can be determined based on the subset of sensor data.

In some examples, the width of the safety area 116 can be the same distance as the width of the vehicle 110. In some examples, the safety area 116 width distance may be a threshold distance greater or less than the vehicle width distance. The width distance may be automatically set based on the width of the vehicle 110 and/or dynamically adjusted based on the size of the roadway overlapping and/or associated with the safety area 116. As explained in relation to Figure 2, the width of the safety area can be increased or otherwise updated based on the simulated or predicted movement of the vehicle along the trajectory. be.

Example 120 shows an object 122 associated with a second safe area (eg, safe area 116). In some cases, object 122 may be determined based on sensor data received from one or more sensors of the vehicle. The sensor data used to determine the object 122 can be filtered as filtered sensor data. Filtered sensor data may include data related to safe area 116. Object 122 may be determined to be associated with safe area 116 based on the filtered sensor data.

In some examples, sensor data (eg, LIDAR data points) associated with secure area 106 may be fused with sensor data (eg, LIDAR data points) associated with secure area 116. For example, sensor data (eg, overlapping sensor data) associated with each of safe area 106 and safe area 116 may be processed once and utilized for each of safe area 106 and safe area 116. Instead of processing duplicate sensor data twice, half of the processing can be eliminated by processing once for safe area 106 and safe area 116. By processing and/or analyzing duplicate sensor data only once for both safe area 106 and safe area 116, computational resources may be saved and/or optimized.

Acts 124 may include object-based control of the vehicle. For example, a vehicle can be controlled based on the likelihood that an object will cross the trajectory. In some examples, a vehicle can be controlled to slow down (eg, slow down) and/or stop. The vehicle can be controlled to reduce the likelihood of crossing by modifying one or more of the acceleration or steering commands. The vehicle may be controlled to stop the vehicle at a location associated with a distance between the vehicle and the object that exceeds a threshold distance. For example, the speed of the vehicle may be controlled to decrease (eg, decrease to 10 miles per hour (mph), 5 mph, 0 mph, etc.). Controlling the vehicle to slow down causes the object to move within the environment (e.g., within the second safety area) based on the distance between the vehicle and the object determined to be greater than a threshold distance. When moving, it is possible to avoid possible collisions between the vehicle and objects.

In some examples, a vehicle can be controlled to accelerate (eg, speed up) based on the object. For example, the speed of the vehicle may be controlled to increase (eg, increase to 10 miles per hour (mph), 15 mph, 20 mph, etc.). By controlling the speed of the vehicle to speed up, it is possible to avoid a possible collision between the vehicle and an object. By controlling the vehicle to slow down based on a determination that the distance between the vehicle and the object is greater than a threshold distance, the object is within the environment (e.g., within the second safety area). When turning and moving, it is possible to avoid possible collisions between the vehicle and objects. The vehicle can be controlled based on the option to accelerate or decelerate, based on the predicted likelihood of a collision associated with each option. An option for determining control of the vehicle may be determined based on a predicted likelihood associated with the option being less than a predicted likelihood associated with other options.

Example 126 shows that vehicle 110 is controlled based on object 122. In some examples, vehicle 110 can be controlled to one or more of slowing down, stopping, and/or accelerating. Vehicle 110 can be controlled to stop at position 128. In some examples, vehicle 110 may be controlled to return on trajectory 108 based on the determination that object 122 is in safe area 116. Analysis of sensor data related to safety area 116 allows vehicle 110 to react to object 122 based on changes in object 122 movement. Vehicle 110 may ignore the detection of object 122 prior to entering safe area 116 based on the objects present in safe area 116 .

Accordingly, multiple safe areas may be determined based on the trajectory associated with the vehicle as described herein. The safe area can be used to more accurately determine the movement of objects in the environment in which the vehicle is driving. One of the safety areas (eg, a side safety area) may be orthogonal to another safety area that is parallel or co-linear with the track. Side safety areas are available for filtering sensor data. Filtering sensor data allows objects to be determined quickly and accurately. The object can be determined as the vehicle approaches the intersection. The vehicle can be controlled to stop to avoid a possible collision with an obstacle. For example, the vehicle may be slowed to a lower speed, slowed to a stop, accelerated, and/or take another action (e.g., turn) to reduce the likelihood of a potential collision with an object. It is possible to control.

FIG. 2 is a diagram illustrating an environment 200 that includes a portion of a safety area having a width determined based on the orientation of a bounding box associated with a vehicle, according to an example of the present disclosure.

In some examples, the location of bounding box 202 relative to vehicle 204 may be determined based on trajectory 206 of vehicle 204. The location of the bounding box may be determined based on a simulation of vehicle 204 traveling along trajectory 206. The simulation may include bounding box 202 propagating along each portion of trajectory 206. The longitudinal orientation of bounding box 202 can match (eg, touch) trajectory 206 as bounding box 202 propagates. Bounding box 202 may include and/or be associated with information such as location, orientation, orientation, and/or size (e.g., length, width, height, etc.) associated with the vehicle. Is possible. Bounding box 202 may include one or more bounding box points (eg, bounding box points 208), each of which is associated with a corner of vehicle 204. The information may include, for each position of bounding box 202 along the trajectory, a position associated with each of the points of bounding box 202.

In some examples, trajectory 206 may be discretized into segments, and multiple segments of safety area 212 (eg, segment 210) are associated with segments of trajectory 206. The position of bounding box 202 along trajectory 206 may be one of the segments of the discretized trajectory. The number of segments included in the plurality of segments of the safe area 212 is not limited, and the number of segments is arbitrary. In some examples, the shape of each segment is polygonal (eg, rectangular).

In some examples, the position of bounding box 202 representing vehicle 204 at a future time may be determined along trajectory 206. The position associated with vehicle 204 along safe area 212 may be based on a simulation of vehicle 204 traveling along trajectory 206 at a future time.

In some examples, it is possible to determine a distance to a bounding box point for each bounding box position along the trajectory. For the location of bounding box 202, it is possible to determine the distance from each bounding box point to trajectory 206 and the distance from the trajectory to the boundary of safe area 212 (eg, boundary 214 or boundary 216). For example, for the location of bounding box 202, a first point (e.g., bounding box point (e.g., bounding box point 208)) and a second point (e.g., the closest point on the trajectory (e.g., closest point 220) )) It is possible to determine the distance between For each point of bounding box 202, the distance between the nearest point 220 and an edge (e.g., edge 224) and/or a boundary (e.g., nearest boundary (e.g., boundary 214) of a segment (e.g., segment 210)) It is possible to determine a distance 222.

In some examples, it is possible to determine the difference between distance 218 and distance 222. Distance 218 may be determined to match or exceed distance 222 based on bounding box 202 having a wider radius of turn at bounding box point 208 than trajectory 206 at closest point 220. The width of segment 210 is then determined such that distance 218 matches or exceeds distance 222 (e.g., distance 222 (e.g., a first distance) is less than distance 218 (e.g., a second distance), or can be updated to include the distance 218. In some examples, safe area 212 may be determined to have a portion associated with segment 210 that is subsequently updated to include distance 218.

In some examples, the width of segment 210 can be updated based on distances associated with different sides of segment 210 (eg, right side and/or inside for a turn). For example, it is possible to determine the distance between the bounding box points and the closest point 220 associated with different sides. It is possible to determine the distance between the closest point 220 and edges (eg, edges 226) associated with different sides of the segment and/or boundary 216. The width of the segment 210 is then based on the distance between the bounding box points associated with different sides and the closest point 220 matching or exceeding the distance between the closest point 220 and the edge 226. can be updated to include the distance between the bounding box points associated with different sides and the closest point 220.

The features described above with respect to FIG. 2 are not limited to them and can be implemented in combination with any of the features described above with respect to FIG. 1. For example, any of the features described above with respect to FIG. 2 may be implemented in combination with any of the features described above with respect to FIG. 1. By combining the features described above with respect to both FIGS. 1 and 2, the vehicle can be controlled with greater precision and safety. For example, the updated safety area width of FIG. 2 can be used in conjunction with the second safety area of FIG. 2 to control the vehicle to avoid a potential collision.

FIG. 3 is a pictorial flow diagram illustrating an example process for determining a segment of a safe area based on a bounding box associated with a vehicle, according to examples of the present disclosure.

Act 302 may include determining segments associated with the safe area. In some examples, a trajectory associated with a vehicle may be discretized into multiple segments associated with safety areas. The safety area can have an outer-turn boundary and an inner-turn boundary. A location of a bounding box associated with the vehicle may be determined based on each of the plurality of segments. The position of the bounding box can be determined along the trajectory. A safe area can include any number of segments. The number of segments can be greater than the threshold so that a safe area determined based on multiple segments can be treated as continuous. In some examples, the maximum width of the safe area may be the width of the vehicle and/or the width of the lane in which the vehicle is currently located.

Example 304 shows a trajectory 306 associated with a vehicle 308. Trajectory 306 can be discretized. A plurality of segments (eg, segment 310) associated with a safe area (eg, safe area 312) can be associated with segments of discretized trajectory 306. For example, each of the segments can be associated with a portion of the first safe area. The safety area 312 can have a boundary 314 on the outside of the turn and a boundary 316 on the inside of the turn. The location of a bounding box associated with vehicle 308 may be determined based on each segment of discretized trajectory 306. The location of the bounding box can be determined along trajectory 306.

In some examples, the width of each segment (e.g., segment 310) is based on a distance on a first side of the segment (e.g., distance 318) and a distance on a second side of the segment (e.g., distance 324). It is possible to decide. Distance 318 may be the distance between an edge of the segment (eg, edge 320) and the closest point in trajectory 306 (eg, point 322). For example, distance 318 may be the length of an imaginary line extending orthogonally from point 322 and ending at boundary 316. Distance 324 may be a distance between another edge of the segment (eg, edge 326) and a point in trajectory 306. For example, distance 324 may be an imaginary line extending orthogonally from point 322 and ending at edge 326 . A point in trajectory 306 associated with a segment may have a position at the center of the segment.

The width of each segment associated with an individual portion of the safety area may be the sum of the first side distance and the second side distance. For example, the width of segment 310 may be the sum of distance 318 and distance 324. In some cases, the width of each segment can be the same width. However, the width is not limited to the same width, but can be varied based on various parameters. For example, the width of any of the segments associated with individual portions of the safety area may depend on, but are not limited to, the speed of the vehicle, the type of terrain the vehicle is traveling on, the type of weather, the dimensions of the road and/or lanes, The determination can be based on a variety of information, including characteristics (eg, type and/or age of tires or brakes), speed limits of the road or intersection on which the vehicle is traveling, etc. The information utilized to determine the width may be real-time or historical data related to similar vehicles in similar environments. Trajectory 306 may be discretized into any number of segments related to safety area 312 (e.g., 1 or a number on the order of 10, 100, 1000, etc.), including numbers approaching infinity. It is possible. A combination of multiple segments of safe area 312 can be associated with multiple segments of discretized trajectory 306 and can approximate a continuous area.

Act 328 may include determining the width of a segment associated with an expanded safety area (eg, a modified safety area). In some examples, the width of the segment can be determined based on a distance associated with the segment on the first side and another distance associated with the second segment. Each of the distances associated with the first side and the second side may be determined to be a greater distance between the first distance and the second distance. The first distance may be associated with a boundary of the safe area and/or an edge of a segment associated with the safe area. For example, the first distance may be the distance between the closest point in the trajectory and an edge of a segment of the safe area, and/or the distance between the closest point in the trajectory and the boundary of a portion of the safe area. Is possible. The second distance can be associated with a point in the simulated bounding box. For example, the second distance may be the distance between the bounding box point and the closest point. In some examples, the maximum width of the expanded safety area may be the width of the lane in which the vehicle is currently located.

Example 330 shows the width of a segment (eg, segment 310) associated with an expanded safety area (eg, safety area 332). In some examples, the width of segment 310 is the same as a distance associated with a first side of segment 310 (e.g., distance 334) and another distance associated with a second side of segment 310 (e.g., distance 336). It is possible to decide based on. The width of each segment associated with a respective portion of the safety area may be the sum of a first side distance (e.g., distance 334) and a second side distance (e.g., distance 336). be.

In some examples, a location associated with vehicle 308 may be determined along safe area 332 at a future time. It is possible to determine the maximum distance from trajectory 306 at a future time, location, of a point associated with the representation of vehicle 308. The vehicle representation may include a bounding box. The width of a portion of safety area 332 (eg, safety area 332 associated with segment 310) may be determined as the maximum distance at that location and may be utilized for control of vehicle 308. For example, the maximum width may include a combination of distance 334 and distance 336. The width of the portion of safe area 332 may be determined as a combination of distance 334 and distance 336.

In some examples, each of the distances associated with the first side and the second side can be determined to be a greater distance between the first distance and the second distance. The first distance used to determine distance 334 may be associated with a current boundary of the current safe area (e.g., boundary 314) and/or with a current edge of a segment associated with a portion of the safe area. It is possible. For example, the first distance used to determine distance 334 may be the distance between the closest point (e.g., point 322) and the current edge of segment 310 (e.g., 320), and/or (eg, point 322) and some current boundary (eg, boundary 314) in current safe area 312. The second distance can be associated with a point in the simulated bounding box. In some examples, the second distance used to determine distance 334 can be the distance between a point in the bounding box and the closest point in the trajectory (eg, point 322). A distance (eg, distance 334) associated with a first side of a segment (eg, 310) can be determined to be a second distance based on a second distance that is greater than or equal to the first distance. The width of segment 310 can be determined based on a distance (eg, distance 334) associated with the first side determined to be the second distance. Additionally or alternatively, the width of the portion of safety area 332 can be determined based on a distance (e.g., distance 334) associated with the first side determined to be the second distance. It is.

In some examples, the first distance used to determine distance 336 is within the current boundary (e.g., boundary 316) of the safe area (e.g., safe area 312) and/or to a portion of the safe area. It may be associated with the current edge (eg, edge 326) of the associated segment. For example, the first distance used to determine distance 336 may be the distance between the closest point (e.g., point 322) and the current edge of segment 310 (e.g., edge 326), and/or the distance between the nearest It may be the distance between a point (eg, point 322) and some current boundary (eg, boundary 316) in current safe area 312. The second distance used to determine distance 336 may be the distance between a point in the bounding box and the closest point in the trajectory (eg, point 322). A distance (e.g., distance 336) associated with a first side of a segment (e.g., segment 310) can be determined to be a second distance based on a second distance that is greater than the first distance. be. The width of the segment 310 may be determined based on a distance associated with the second side (eg, distance 336) set as the second distance. Additionally or alternatively, the width of the portion of safety area 332 can be determined based on a distance associated with the first side (e.g., distance 336) determined as the second distance. .

In some examples, edge 338 may extend beyond boundary 314 on the outside of the turn. Edge 340 may extend beyond boundary 316 on the inside of the turn. Safe area 312 can have a variable width.

The subset of sensor data may be determined based on safe area 312 and/or safe area 332. It is possible to detect objects represented in a subset of sensor data. It is possible to determine the likelihood that the object will cross trajectory 306 (eg, the likelihood of crossing). Vehicle 308 can be controlled based on likelihood. The features described above with respect to FIG. 3 are not limited to them and can be implemented in combination with any of the features described above with respect to FIG. 1 or FIG. 2. For example, vehicle 308 may be controlled such that one or more of the acceleration commands or steering commands are modified to reduce the likelihood of crossing.

Thus, as discussed herein, a trajectory associated with a vehicle may be discretized into segments. Multiple segments of the safe area can be associated with segments of the discretized trajectory. A segment and/or portion of the safety area may have a width adjusted based on a simulated bounding box associated with the vehicle. The width can be adjusted based on the portion of the bounding box that matches or extends beyond the safe area. By adjusting the width, it is possible to more accurately predict and avoid the possibility of a collision between the vehicle and an object in the environment in which the vehicle is traveling. By adjusting the width of a portion of the safety area for vehicles with a longer than average length, it is possible to reduce the likelihood of a collision to a greater degree. For example, by adjusting the width of a portion of the safety area for large trucks (eg, semi-trailer trucks), the likelihood of a potential collision can be reduced to a greater degree.

FIG. 4 is a block diagram illustrating an example system 400 for implementing the techniques described herein. In at least one example, system 400 may include a vehicle 402. In the illustrated system 400, vehicle 402 is an automobile, although vehicle 402 may be any type of vehicle. Vehicle 402 can be implemented as any of the vehicles in FIGS. 1-3 and 5-7.

Vehicle 402 may be an autonomous vehicle, such as a motor vehicle configured according to a Level 5 classification published by the U.S. National Highway Traffic Safety Administration. The Level 5 class describes vehicles that are capable of performing all safety critical functions throughout the journey without the driver (or occupants) being expected to be in constant control of the vehicle. In such examples, the vehicle 402 may be configured to control all functions from initiation to completion of the trip, including all parking functions, such that the driver and/or steering wheel , accelerator pedals, and/or brake pedals may not be included. This is just one example, and the systems and methods described herein can be applied to vehicles that always require manual driver control, to vehicles that are partially or fully autonomously controlled, on the ground, in the air, or may be incorporated into any water vehicle.

Vehicle 402 includes one or more computing devices 404, one or more sensor systems 406, one or more emitters 408, one or more communication connections 410 (also referred to as communication devices and/or modems), It may include at least one direct connection 412 (eg, physically coupled to vehicle 402 for exchanging data and/or providing power) and one or more drive systems 414. One or more sensor systems 406 can be configured to capture data related to the environment.

Sensor system 406 includes time-of-flight sensors, position sensors (e.g., GPS, compass, etc.), inertial sensors (e.g., inertial measurement units (IMUs), accelerometers, magnetometers, gyroscopes, etc.), LIDAR sensors, radar sensors, sonar sensors. , infrared sensors, cameras (e.g. RGB, IR, intensity, depth, etc.), microphone sensors, environmental sensors (e.g. temperature sensors, humidity sensors, light sensors, pressure sensors, etc.), ultrasonic transceivers, wheel encoders, etc. Is possible. Sensor system 406 may include multiple examples of each of these or other types of sensors. For example, time-of-flight sensors may include time-of-flight sensors located individually at the corners, front, rear, sides, and/or top of vehicle 402. As another example, camera sensors may include multiple cameras located at various locations on the exterior and/or interior of vehicle 402. Sensor system 406 can provide input to first computer 404 .

Vehicle 402 may also include emitters 408 that emit light and/or sound. In this example, emitter 408 includes an internal audiovisual emitter for communicating with the occupants of vehicle 402. By way of example, and not limitation, internal emitters include speakers, lights, signs, display screens, touch screens, tactile emitters (e.g., vibration and/or force feedback), mechanical actuators (e.g., seatbelt tensioners, seat positioners, headrest positioners). etc.). In this example, emitter 408 may also include an external emitter. By way of example and not limitation, external emitters in this example include lights (e.g., indicator lights, signs, light arrays, etc.) to signal heading or other indicators of vehicle action, and pedestrians or other nearby vehicles. may include one or more acoustic emitters (e.g., speakers, speaker arrays, horns, etc.) for audibly communicating with the acoustic beam steering technology, one or more of which may constitute an acoustic beam steering technique. It is.

Vehicle 402 may also include a communications connection 410 that enables communication between vehicle 402 and one or more local or remote computers (eg, remote operating computers) or remote services. For example, communication connection 410 may facilitate communication with other local computers in vehicle 402 and/or drive system 414. Communication connection 410 may also allow vehicle 402 to communicate with other nearby computers (eg, other nearby vehicles, traffic lights, etc.).

Communication connection 410 may include a physical and/or logical interface for connecting first computer 404 to another computer or one or more external networks 416 (eg, the Internet). For example, communication connection 410 may include Wi-Fi®-based communications, such as through frequencies defined by the IEEE 802.11 standard, short-range radio frequencies, such as Bluetooth, cellular communications (e.g., 2G , 3G, 4G, LTE, 5G, etc.), satellite communications, discrete short range communication (DSRC), or any suitable wired or wireless communication protocol that allows individual computers to interface with other computers. can.

In at least one example, vehicle 402 can include a drive system 414. In some examples, vehicle 402 may have a single drive system 414. In at least one example, when vehicle 402 has multiple drive systems 414, individual drive systems 414 can be located at opposite ends (eg, front and rear, etc.) of vehicle 402. In at least one example, drive system 414 can include a sensor system 406 that detects the condition of drive system 414 and/or the surroundings of vehicle 402. By way of example and not limitation, sensor system 406 may include one or more wheel encoders (e.g., rotary encoders) that sense the rotation of wheels of the drive system, inertial sensors (e.g., inertial metrology) that measure direction and acceleration of the drive system. devices, accelerometers, gyroscopes, magnetometers, etc.), cameras or other image sensors, ultrasonic sensors that acoustically detect objects in the vicinity of the drive system, LIDAR sensors, radar sensors, etc. Some sensors, such as wheel encoders, may be unique to drive system 414. In some cases, sensor system 406 in drive system 414 can be redundant or complementary to a corresponding system in vehicle 402 (eg, sensor system 406).

Drive system 414 includes a high voltage battery, a motor that propels the vehicle, an inverter that converts direct current from the battery to alternating current for use by other vehicle systems, a steering motor, and a steering rack (which may be electronic). A steering system including a steering system, a braking system including hydraulic or electronic actuators, a suspension system including hydraulic and/or pneumatic components, a stability control system to distribute braking force to reduce loss of traction and maintain control, an HVAC system, lighting devices (e.g., head/tail lighting that illuminates the exterior surroundings of the vehicle) and one or more other systems (e.g., cooling systems, safety systems, on-board charging systems, DC/DC converters); Many vehicle systems can be included, including other electronic components such as high voltage battery junctions, high voltage cables, charging systems, charging ports, etc. Additionally, drive system 414 may include a drive system controller that receives and preprocesses data from sensor system 406 to control operation of various vehicle systems. In some examples, a drive system controller may include one or more processors and memory communicatively coupled to the one or more processors. Memory may store one or more components for implementing various functions of drive system 414. Additionally, drive system 414 may also include one or more communication connections that enable an individual drive system to communicate with one or more other local or remote computers.

Vehicle 402 may include one or more second calculators 418 that provide redundancy, error checking, and/or validation of decisions and/or instructions determined by first calculator 404. .

As an example, first computer 404 may be considered a primary system, while second computer 418 may be considered a secondary system. The primary system may generally perform processing that controls how the vehicle is operated within the environment. The primary system may perform various artificial intelligence (AI) techniques, such as machine learning, to understand the vehicle's surrounding environment and/or direct the vehicle to move within the environment. For example, the primary system may run AI techniques to localize the vehicle, detect objects around the vehicle, segment sensor data, determine object classification, and predict object trajectories. , and generate vehicle trajectories. In examples, the primary system processes data from multiple types of sensors in the vehicle, such as light detection and ranging (LIDAR) sensors, radar sensors, image sensors, depth sensors (time-of-flight, structured light, etc.).

The secondary system may verify the operation of the primary system and may take over control of the vehicle from the primary system when there is a problem with the primary system. The secondary system may implement probabilistic techniques based on the position, velocity, acceleration, etc. of the vehicle and/or objects around the vehicle. For example, the secondary system may perform one or more probabilistic techniques to independently localize the vehicle, detect objects around the vehicle, segment sensor data, identify object classifications, The trajectory of an object may be predicted or the trajectory of a vehicle may be generated. In examples, the secondary system processes data from fewer sensors, such as a subset of the sensor data processed by the primary system. For example, a primary system may process LIDAR data, radar data, image data, depth data, etc., while a secondary system processes only LIDAR data and/or radar data (and/or time-of-flight data). You may do so. However, in other examples, the secondary system may process sensor data from any number of sensors, such as data from each sensor, data from the same number of sensors as the primary system.

The secondary system may perform any of the techniques of FIGS. 1-3 and 5-7 as described herein. For example, the secondary system may determine one or more safety areas that are utilized to control the vehicle. The secondary system may be a redundant backup system.

Additional examples of vehicle architectures with primary and secondary systems are found, for example, in U.S. Pat. It can be seen in .

First computer 404 may include one or more processors 420 and memory 422 communicatively coupled to one or more processors 420 . In the illustrated example, memory 422 of first computer 404 stores a localization component 424 , a perception component 426 , a prediction component 428 , a planning component 430 , a map component 432 , and one or more system controllers 434 . Although shown as residing in memory 422 for illustrative purposes, localization component 424, perception component 426, prediction component 428, planning component 430, map component 432, and one or more system controllers 434 may additionally include or alternatively, the first computer 404 (e.g., stored on a different component of the vehicle 402) may be accessed and/or the vehicle 402 (e.g., stored remotely) may be accessed. It is considered possible to do so.

In memory 422 of first computer 404 , localization component 424 may include functionality to receive data from sensor system 406 to determine the location of vehicle 402 . For example, localization component 424 may include and/or request/receive a three-dimensional map of the environment and may continually determine the autonomous vehicle's position in the map. . In some cases, localization component 424 uses time-of-flight data, image data, LIDAR data, radar data, sonar data, IMU data, GPS using SLAM (simultaneous localization and mapping) or CLAMS (calibration, localization and mapping, simultaneously). data, wheel encoder data, or any combination thereof, and accurately determine the position of the autonomous vehicle. In some cases, localization component 424 can provide data to various components of vehicle 402 to determine an initial position of the autonomous vehicle for generating a trajectory, as described herein. .

Perception component 426 may include functionality to perform object detection, segmentation, and/or classification. In some examples, the perception component 426 detects the presence of objects closest to the vehicle 402 and/or the entity type (e.g., car, pedestrian, cyclist, building, tree, road surface, curve, sidewalk, etc.). It is possible to provide processed sensor data that indicates a classification of an entity as an unknown entity, etc.). In additional or alternative examples, the perception component 426 may provide processed sensor data that is indicative of one or more characteristics related to the detected entity and/or the environment in which the entity is located. In some examples, characteristics associated with an entity include, but are not limited to, x location (global positioning), y location (global positioning), z location (global positioning), orientation, entity type (e.g., classification), entity may include the speed of the entity, the range (size) of the entity, etc. Features related to the environment can include, but are not limited to, the presence of another entity in the environment, the state of another entity in the environment, time of day, day of the week, season, weather conditions, darkness/light indicators, etc. .

As discussed above, the perception component 426 may use a perception algorithm to determine perception-based bounding boxes associated with objects in the environment based on the sensor data. For example, perception component 426 can receive and classify image data to determine objects represented in the image data. Then, by using the detection algorithm, the perception component 426 can generate a two-dimensional bounding box and/or a perception-based three-dimensional bounding box associated with the object. Perception component 426 can further generate a three-dimensional bounding box associated with the object. As mentioned above, a three-dimensional bounding box can provide additional information such as position, orientation, pose, and/or size (e.g., length, width, height, etc.) relative to an object. .

Perceptual component 426 may include functionality for storing perceptual data generated by perceptual component 426 . In some cases, the perception component 426 can determine trajectories that correspond to objects that have been classified into object types. For illustrative purposes only, the perception component 426 may capture one or more images of the environment including objects, such as pedestrians, through the use of the sensor system 406. The pedestrian may be in a first position at time T and in a second position at time T+t (eg, moving during a period from time T to time t). In other words, the pedestrian can move from the first position to the second position during this time interval. Such movement can be recorded, for example, as stored sensory data associated with the object.

Stored sensory data may include fused sensory data captured by vehicle 402 in some examples. The fused sensory data may be a fusion of sensor data from sensor systems 406 such as image sensors, LIDAR sensors, radar sensors, time-of-flight sensors, sonar sensors, Global Positioning System sensors, internal sensors, and/or any combination thereof. Other combinations are possible. The stored sensory data may additionally or alternatively include classification data that includes a semantic classification of objects (eg, pedestrians, vehicles, buildings, road surfaces, etc.) represented in the sensor data. The stored sensory data may additionally or alternatively include trajectory data (position, orientation, sensor characteristics, etc.) corresponding to the movement of an object classified as a dynamic object through the environment. The trajectory data can include multiple trajectories of multiple different objects over time. This trajectory data provides images of certain types of objects (e.g. pedestrians, animals, etc.) when the object is not moving (e.g. stationary) or in motion (walking, running, etc.). It can be used for identification. In this example, the calculator determines that the trajectory corresponds to a pedestrian.

Prediction component 428 can generate one or more probability maps that represent predicted probabilities of where one or more objects may be located in the environment. For example, prediction component 428 may generate one or more probability maps for vehicles, pedestrians, animals, etc. within a threshold distance from vehicle 402. In some cases, the prediction component 428 is capable of measuring the trajectory of the object and includes discretized predictive probability maps, temperature maps, probability distributions, discretized probability distributions, and/or observations and predictions. It is possible to generate a trajectory for an object based on its behavior. In some cases, one or more probability maps can represent the intent of one or more objects in the environment.

Planning component 430 may determine a path to be followed by vehicle 402 as it travels through the environment. For example, planning component 430 may determine different routes and paths and different levels of detail. In some cases, planning component 430 can determine a route to travel from a first location (eg, a current location) to a second location (eg, a destination). For purposes of this explanation, a route may be a sequence of waypoints traveling between two locations. As non-limiting examples, waypoints may include streets, intersections, Global Positioning System (GPS) coordinates, and the like. Further, the planning component 430 can generate instructions for guiding the autonomous vehicle along at least a portion of the route from the first location to the second location. In at least one example, planning component 430 can determine how to navigate from a first waypoint in the sequence of waypoints to a second waypoint in the sequence of waypoints. In some examples, the instructions can be a path or part of a path. In some examples, multiple passes may be generated substantially simultaneously (i.e., within technical tolerances) according to the receding horizon technique. A single path among the multiple paths in the receding horizon data with the highest confidence level may be selected for driving the vehicle.

In other examples, planning component 430 may alternatively or additionally use data from perception component 426 and/or prediction component 428 to determine a path to be followed by vehicle 402 as it travels through the environment. . For example, planning component 430 may receive data about objects related to the environment from perception component 426 and/or prediction component 428. Using this data, the planning component 430 can determine a route to travel from the first location to the second location while avoiding objects in the environment. In at least some examples, such a planning component 430 may determine that there is no collision-free path and then proceed to a safe stop that avoids all collisions and/or reduces damage. may be provided to bring the vehicle 402 to the vehicle 402.

Memory 422 may further include one or more maps 432 that may be used to navigate the vehicle within the environment. For the purposes of this description, any 2D, 3D, or N-D dimension that can provide information about the environment, such as, but not limited to, topology (such as intersections), streets, mountains, roads, terrain, and the environment in general. It can be any number of data structures modeled in . In some cases, the map includes, but is not limited to, texture information (e.g., color information (e.g., RGB color information, Lab color information, HSV/HSL color information), etc.), intensity information (e.g., LIDAR information, RADAR information, etc.) , spatial information (e.g., image data projected onto a mesh, individual "surface elements" (e.g., polygons associated with individual colors and/or intensities), reflectivity information (e.g., specularity information, retroreflectivity information) , BRDF information, BSSRDF information, etc.). In one example, the map may include a three-dimensional mesh of the environment, as described herein. In some cases, the map may include: The map can be stored in an environment tile format such that each tile represents a discretized portion of the environment, and can be loaded into working memory as needed. The one or more maps 432 can include at least one map (e.g., an image and/or a mesh). In some examples, the vehicle 402 controls based at least in part on the map 432. That is, map 432 may be used in conjunction with localization component 424, perception component 426, prediction component 428, and/or planning component 430 to determine the location of vehicle 402 and identify objects in the environment. and can be used to generate predicted probabilities associated with objects and/or vehicles 402 and/or to generate paths and/or trajectories to travel within an environment.

In some examples, one or more maps 432 may be stored on a remote computer (such as computer 448) that is accessible via network 416. In some examples, multiple maps 432 can be saved based on characteristics (eg, type of entity, time of day, day of the week, season of the year, etc.), for example. Saving multiple maps 432 may have similar memory requirements, but may increase the speed at which data within the maps can be accessed.

In at least one example, the first computer 404 can include one or more system controllers 434 that control steering, drive power, brakes, safety, emitters, communications, and other functions of the vehicle. It is possible to configure the system to control. These system controllers 434 communicate with and/or control corresponding systems of the drive system 414 and/or other components of the vehicle 402 that may be configured to operate according to paths provided by the planning component 430. It is possible.

The second computer 418 may include one or more processors 436 and a memory 438 that includes components that verify and/or control aspects of the vehicle 402, as described herein. In at least one example, one or more processors 436 may be similar to processor 420 and memory 438 may be similar to memory 422. However, in some examples, processor 436 and memory 438 may include different hardware than processor 420 and memory 422 for added redundancy.

In some examples, memory 438 may include a localization component 440, a perception/prediction component 442, a planning component 444, and one or more system controllers 446.

In some examples, localization component 440 may receive sensor data from sensor 406 to determine one or more positions and/or orientations (together, attitude) of autonomous vehicle 402. Here, position and/or direction may relate to points and/or objects in the environment in which autonomous vehicle 402 is located. In examples, the direction may include an indication of yaw, roll, and/or pitch of the autonomous vehicle 402 relative to a reference plane and/or relative to a point and/or object. In examples, localization component 440 may perform less processing than localization component 424 of first computer 404 (eg, high-level localization). For example, localization component 440 may not determine the pose of autonomous vehicle 402 with respect to a map, but simply determine the pose of autonomous vehicle 402 with respect to detected objects and/or surfaces in the vicinity of autonomous vehicle 402. (e.g., a local location, not a global location). Such position and/or orientation is illustratively determined using probabilistic filtering techniques such as, for example, Bayesian filters (Kalman filters, Extended Kalman filters, Unscented Kalman filters, etc.) using some or all of the sensor data. It's fine.

In some examples, perception/prediction component 442 can include functionality to detect, identify, classify, and/or track objects represented in sensor data. For example, perception/prediction component 442 may perform clustering operations and operations that estimate or determine heights associated with objects, as described herein.

In some examples, perception/prediction component 442 may include an M estimator, but may not include an object classifier, such as a neural network, decision tree, etc., for classifying objects. In additional or alternative examples, perception/prediction component 442 may include any type of M-model configured to disambiguate object classification. In contrast, perception component 426 may include a pipeline of hardware and/or software components, including one or more machine learning models, a Bayesian filter (e.g., a Kalman filter), a graphics processing device, (GPU) etc. may be provided. In some examples, the perceptual data determined by the perception/prediction component 442 (and/or 426) may include object detection (e.g., identification of sensor data related to objects in the autonomous vehicle's surrounding environment), object classification ( object tracking (e.g., historical, current, and/or predicted object position, velocity, acceleration, and/or heading); etc. may be provided.

Perception/prediction component 442 may process the input data to determine one or more predicted object trajectories. For example, based on the object's current location and the object's velocity over the next few seconds, the perception/prediction component 442 may predict the path that the object will travel over the next few seconds. In some examples, the predicted path may include using linear assumptions of a given position, direction, velocity, and/or orientation. In other examples, the predicted path may include more complex analysis.

In some examples, planning component 444 can include functionality to receive a trajectory from planning component 430 and certify that the trajectory is collision-free and/or within a safety margin. In some examples, the planning component 444 can generate a safe stop trajectory (e.g., a trajectory that stops the vehicle 402 at a "comfortable" deceleration (e.g., less than a maximum deceleration)) and In such examples, the planning component 444 may generate an emergency stop trajectory (eg, maximum deceleration with or without steering input).

In some examples, system controller 446 may include functionality to control safety critical components of the vehicle (eg, steering, brakes, motors, etc.). In this way, the second computer 418 can provide redundancy and/or additional hardware and software layers for vehicle safety.

Vehicle 402 can be connected to computer 448 via network 416 and can include one or more processors 450 and memory 452 communicatively coupled to one or more processors 450. It is possible. In at least one example, one or more processors 450 may be similar to processor 420 and memory 452 may be similar to memory 422. In the illustrated example, memory 452 of calculator 448 stores component 454, which may correspond to any of the components described herein.

Processors 420, 436, and/or 450 may be any suitable processor capable of executing instructions to process data and perform operations as described herein. It is. By way of example and not limitation, processors 420, 436 and/or 450 include one or more central processing units (CPUs) that convert electrical data into electrical data that can be stored in registers and/or memory; It may include graphics processing units (GPUs) or any other device or portion of a device. In some examples, integrated circuits (e.g., ASICs, etc.), gate arrays (e.g., FPGAs), and other hardware devices are also processors insofar as they are configured to execute encoded instructions. It is possible to think.

Memory 422, 438, and/or 452 are examples of non-transitory computer-readable media. Memories 422, 438, and/or 452 may contain an operating system and one or more software applications, instructions, programs, and/or software for performing the functions of the methods and various systems described herein. It is possible to store data. In various implementations, memory 422, 438, and/or 452 may be static random access memory (SRAM), synchronous dynamic RAM (SDRAM), non-volatile/flash memory, or other memory capable of storing information. It may be implemented using any suitable memory technology, such as any type of memory. The architectures, systems, and individual elements described herein may include many other logical, programmatic, and physical components, some of which are illustrated in the accompanying figures. This is merely an example in relation to the description in the specification.

In some cases, some or all of the components described herein may include any models, algorithms, and/or machine learning algorithms. For example, memory 422, 438, and/or 452 may optionally be implemented as a neural network. In some examples, components in memory 422, 438, and/or 452 may not include machine learning algorithms to reduce complexity and to be verified and/or certified from a security perspective. (or may include simplified or verifiable machine learning algorithms).

As described herein, an exemplary neural network is a biologically derived algorithm that passes input data through a series of connected layers to produce an output. Each layer of the neural network may also constitute another neural network or any number of layers (convolutional or not). As understood in the context of this disclosure, neural networks may utilize machine learning, which refers to a broad class of these algorithms in which output is generated based on learned parameters. There is.

Although discussed in the context of neural networks, any type of machine learning that is compatible with this disclosure can be used. For example, machine learning or machine learning algorithms may include, but are not limited to, regression algorithms (e.g., ordinary least squares regression (OLSR), linear regression, logistic regression, stepwise regression, multivariate adaptive regression splines (MARS), locally weighted scatterplot smoothing (LOESS)), instance-based algorithms (e.g. ridge regression, least absolute shrinkage selection operator (LASSO), elastic net, least angle regression (LARS)), decision tree algorithms (e.g. classification regression trees (CART), iterative binary trees 3 (ID3), chi-square automatic interaction detection (CHAID), decision stamps, conditional decision trees), Bayesian algorithms (e.g., Naive Bayes, Gaussian Naive Bayes, multinomial Naive Bayes, averaging One Dependency Classifier (AODE), Bayesian Belief Network (BNN), Bayesian Network), Clustering Algorithm (e.g. k-means, k-medians, Expectation Maximization (EM), Hierarchical Clustering), Association Rule Learning Algorithm (e.g., perceptron, backpropagation, Hopfield network, Radial Basis Function Network (RBFN), etc.), deep learning algorithms (Deep Boltzmann Machine (DBM), Deep Belief Networks (DBN), Convolutional Neural Network (C NN), Stacked Auto -Encoders, etc.), dimensionality reduction algorithms (e.g., Principal Component Analysis (PCA), Principal Component Regression (PCR), Partial Least Squares Regression (PLSR), Summon Mapping, Multidimensional Scaling (MDS), Projection Pursuit, Linear Discriminant Analysis) (LDA), mixture discriminant analysis (MDA), quadratic discriminant analysis (QDA), flexible discriminant analysis (FDA), etc.), ensemble algorithms (e.g. Boosting, Bootstrapped Aggregation (Bagging), AdaBoost, Stacked Generalization (B rendering), Gradient May include Boosting Machines (GBM), Gradient Boosted Regression Trees (GBRT), Random Forest), SVM (Support Vector Machine), supervised learning, unsupervised learning, semi-supervised learning, etc. It is possible.

Additional examples of architectures include neural networks such as ResNet50, ResNet101, VGG, DenseNet, PointNet, etc.

FIG. 5 is a diagram illustrating an environment 500 that includes boundaries determined based on vehicle trajectories and safety areas determined based on lines associated with the boundaries, according to examples of the present disclosure.

In some examples, environment 500 can include a first boundary (e.g., boundary 502) and a second boundary (e.g., boundary 504) associated with a first safe area (e.g., safe area 506). It is possible. Safe area 506 may be determined based on a trajectory associated with vehicle 508. A second safe area (eg, safe area 510) may be determined based on line 512 associated with boundary 504. An intersection 514 between boundary 502 and line 512 associated with boundary 504 associated with safe area 506 may be determined. Safe area 506 may be determined based on intersection points 514. In some examples, the intersection point 514 may be associated with a turn in the environment that meets or exceeds a threshold turn angle. Safe area 510 can be utilized to filter sensor data related to safe area 510 as filtered sensor data. Safe area 510 and filtered sensor data may be utilized to determine objects 516 associated with safe area 510. The width of the safety area 510 includes all cross lanes and/or other areas from which one or more objects may enter the lane associated with the trajectory of the vehicle 508. It is possible to decide to include one or more of the following. In some examples, the maximum width of safe area 510 may be related to the width of a cross lane orthogonal to the lane in which vehicle 508 is currently located. However, the maximum width is not limited to such a width and can be associated with a width that is smaller or larger than the width of the intersecting lane.

Thus, as described herein, the second safety area makes it easier to determine whether an object is approaching the vehicle when the vehicle enters or is performing a turn. It can be determined that In order to avoid possible collisions due to lack of information related to the environment, it is possible to determine the object based on the second safe area. By determining the second safety area based on the vehicle's approach or turn preparation, it is possible to increase the size of the environment considered in determining the target. The width of the second safety area may include cross lanes and/or other areas from which one or more objects may enter the lane associated with the trajectory of the vehicle 508. It is possible to decide to include one or more of all of the above. In some examples, the maximum width of the second safe area can be related to the width of a cross lane orthogonal to the lane in which the vehicle is currently located. However, the maximum width is not limited to such a width and can be associated with a width that is smaller or larger than the width of the intersecting lane.

FIG. 6 is a diagram illustrating an environment 600 that includes a safe area determined based on a trajectory that includes a vehicle left turn, according to an example of the present disclosure.

In some examples, environment 600 can include a first safe area (eg, safe area 602). Safe area 602 may be determined based on a trajectory associated with vehicle 604 that includes turns (eg, left turns). In some examples, the safe area 602 can be determined based at least in part on the width and/or length of the vehicle 110, the current speed of the vehicle 604, and/or the speed associated with the trajectory. It is possible. Each part of the safety area 602 can have the same width. A safe area 602 can be determined to be associated with a left turn based on the trajectory associated with the left turn.

In some examples, the second safety area (e.g., safety area 606 can be determined based on a trajectory. The safety area 606 can be determined based on a trajectory associated with a turn (e.g., a left turn). In some examples, the safe area 606 can be determined based at least in part on the safe area 602 associated with the vehicle 604 that is being controlled to turn. In some examples, the safe area 606 can be determined based on turns in the environment (e.g., turn angles associated with turns) that meet or exceed a threshold turn angle. , can be associated with a first boundary (e.g., boundary 608) and a second boundary (e.g., boundary 610).Safety area 606 can be determined based on line 612 associated with boundary 610. A point of intersection 614 between the boundary 608 and a line 612 associated with the safe area 602 can be determined. A safe area 606 can be determined based on the point of intersection 614. In an example, intersection point 614 can be associated with a turn in the environment that meets or exceeds a threshold turn angle. Safe area 606 filters sensor data associated with safe area 606 as filtered sensor data. The safe area 606 and the filtered sensor data can be used to determine objects 616 associated with the safe area 606. The width of the safe area 606 is lane and/or oncoming lane and/or any other area from which one or more objects may enter the lane associated with the trajectory of the vehicle. In some examples, the maximum width of safety area 606 may be related to the width of an oncoming lane adjacent to the lane in which vehicle 508 is currently located. However, the maximum width is not limited to such a width and can be associated with a width that is smaller or larger than the width of the oncoming lane.

In some examples, safety area 606 may be adjacent to and substantially parallel to safety area 602 based on safety area 602 associated with a left turn. In some examples, safety area 606 may be spaced apart from vehicle 604. For example, the distance between safety area 606 and vehicle 604 can be greater than or equal to a threshold distance (eg, 10 cm, 1 meter, 3 meters, etc.). In other examples, safety area 606 may be bounded by (eg, capable of contacting) vehicle 604. In other examples, safety area 606 can be spaced apart from (eg, not bordering) vehicle 604.

In some examples, sensor data received by the vehicle 604 determines an object 616 in the nearest lane adjacent to the vehicle based on a safety area 606 associated with (e.g., overlaps with) the nearest lane. It can be used for In some examples, sensor data received by the vehicle 604 is configured to determine an object 616 in a lane adjacent to the nearest lane or to determine a lane one or more lanes away from the nearest lane. It is possible to use it. In some examples, the sensor data received by vehicle 604 may be detected in one or more lanes, including the nearest lane, a lane adjacent to the nearest lane, and/or a lane one or more lanes away from the nearest lane. It can be used to determine objects 616 in the lane.

Therefore, not only is it possible to determine a first safe area as described herein, but also a second safe area can be determined based on the vehicle making a left turn. By making the second safety area parallel to the first safety area, it is possible to more easily determine whether an oncoming vehicle object is approaching. Although the features described above with respect to FIG. 6 relate to left turns of the vehicle, they are not limited to such features. The feature can also be used for other types of turns, such as entering oncoming traffic to bypass double-parked vehicles. The second safety area can also be used to determine if an object is approaching from behind on a one-way street. The vehicle is controlled to slow to a slower speed, slow to a stop, accelerate, and/or take other actions (e.g., turn) to reduce the likelihood of collision with an object. It is possible to do so. The width of the second safety area may include one or more adjacent and/or oncoming traffic lanes, and/or one or more objects from that area into one or more adjacent and/or oncoming traffic lanes and/or another area associated with the vehicle's trajectory. It is possible to decide to include one or more of all possible areas of entry. In some examples, the maximum width of the second safety area can be related to the width of the oncoming lane closest to the lane in which the vehicle is currently located. However, the maximum width is not limited to such a width and can also be associated with a width smaller or larger than the width of the oncoming lane. In some examples, the length of the second safety area can be a fixed distance (eg, 10 m, 50 m, 150 m, 200 m, etc.). In some examples, the length can be based on the speed of the vehicle 604, the shape of the first safety area, etc.

FIG. 7 is a pictorial flow diagram illustrating an example process 700 for determining a safe area based on stationary vehicles, according to examples of the present disclosure.

Act 702 may include determining that the vehicle is approaching an intersection. In some examples, information related to the environment and/or the vehicle in which the vehicle is traveling may be determined. The information can be utilized to determine whether a vehicle is approaching an intersection. For example, a vehicle can receive and analyze sensor data to identify an approaching intersection. The first safe area may be determined based on a trajectory associated with the vehicle. The first safe area may be determined to be associated with a trajectory that does not include turns associated with the vehicle.

Example 704 depicts an environment that includes a vehicle 706 approaching an intersection. In some examples, vehicle 706 may be associated with track 708. A first safe area (eg, safe area 710) may be determined based on a trajectory 708 associated with vehicle 706 traveling through the environment. Safe area 710 may be associated with trajectory 708 that does not include turns associated with the vehicle.

Act 712 may include determining a second safe area based on parked vehicles. In some examples, the second safe area is determined based on a distance between the vehicle and the intersection that is less than a threshold distance and/or based on a speed of the vehicle that is less than a threshold speed. Is possible.

Example 714 shows a vehicle 706 slowing down and stopping based on a stop sign. In some examples, a second safe area (e.g., safe area 716 can be determined based on the distance between vehicle 706 and the intersection and/or based on the speed of the vehicle). Safe area 716 may be perpendicular to safety area 710, which is parallel to trajectory 708, as described above.

Act 718 may include removing the safe area based on the speed meeting or exceeding a threshold. The second safety area can be removed when the speed of the vehicle increases as the vehicle is controlled to proceed through the intersection. The second safe area can be deleted based on the speed meeting or exceeding the threshold speed.

Example 720 illustrates a second safe area (eg, safe area 716) that is removed based on speed meeting or exceeding a threshold. The second safe area 716 may be removed as the speed of the vehicle increases as the vehicle 706 is controlled to proceed through the intersection. Vehicle 706 can be controlled based on trajectory 708. Safe areas 716 can be removed based on the speed meeting or exceeding a threshold speed. The features described above with respect to FIG. 7 are not limited to these and can be implemented with any of the features described with respect to any of FIGS. 1-3, FIG. 5, and FIG. 6.

FIG. 8 is a flowchart illustrating an example process 800 for determining a safe area based on a trajectory that includes turns of a vehicle, according to examples of the present disclosure.

At operation 802, processing can include receiving a trajectory associated with a vehicle. A trajectory can be associated with a vehicle moving through an environment.

At operation 804, processing can include determining a first safe area. The first safe area can be determined based on the trajectory. The first safe area may have a constant width.

At operation 806, processing can include determining a second safe area that is different from the first safe area based on the trajectory. The second safe area may be orthogonal to the first safe area. The direction of the second safety area can be determined based on vehicle operation, vehicle position, and the like. The second safe area can be determined based on the trajectory associated with the turn (eg, right turn or left turn).

At operation 808, processing can include determining a subset of sensor data associated with one or more of the first or second safety areas. The object may be determined based on a subset of sensor data received from one or more sensors of the vehicle. For example, sensor data can include data from multiple types of sensors, such as LIDAR sensors, image sensors, depth sensors (time-of-flight, structured light, etc.), and the like. The subset of sensor data can be output by one or more machine learning models, Bayesian filters (e.g., Kalman filters), and/or graphics processing units (GPUs), etc. based on the sensor data input. be.

At act 810, processing may include determining whether the likelihood matches or exceeds a threshold likelihood. If no, processing can return to operation 802. If yes, processing may continue to operation 812.

At act 812, processing may include controlling the vehicle based on the likelihood. In some examples, a vehicle can be controlled to slow down (eg, slow down) and/or stop. In some examples, the vehicle can be controlled to accelerate. The option that is determined to be controlled to accelerate or decelerate may be based on the predicted crash likelihood associated with that option being smaller than for other options.

FIG. 9 is a flowchart illustrating an example process 900 for determining the width of a portion of a safe area based on the orientation of a bounding box as relative to a vehicle, according to examples of the present disclosure.

At operation 902, processing can include receiving a trajectory associated with a vehicle. Trajectories can be associated with vehicles and can be discretized.

At operation 904, processing can include determining a safe area based on the trajectory. It is possible to determine multiple segments related to the safe area. A segment of the plurality of segments is capable of extending a first distance substantially orthogonally from the trajectory.

At operation 906, processing may include determining a position associated with the vehicle along the safe area at a future time. The location can be associated with a bounding box representing the vehicle at a future time.

At operation 908, processing may include determining a maximum distance of a point associated with a representation of the vehicle at a future time from the trajectory at the location. Points in the vehicle representation can be associated with corners of the bounding box.

In operation 910, processing may include defining the width of the portion of the safe area as the maximum distance at the location. The width of the portion of the safe area can be set to the first distance based on the first distance matching or exceeding the second distance. By setting the first distance, the safe area can be expanded to include the first point. Otherwise, the width of the portion of the safe area may be set to the second distance based on the first distance not matching or exceeding the second distance.

At operation 912, processing may include determining whether portions of the safe area associated with all segments of the trajectory have been considered. If yes, processing may continue to operation 914. If no, the process may return to operation 906 to determine a first safe area and a second safe area associated with a portion of the safe area associated with the next segment of the trajectory. .

At operation 914, processing may include controlling the vehicle based on the safe area. The vehicle can be controlled based on the width of the portion of the safety area, which is set as the first distance.

Illustrative Section A: A system comprising one or more processors and one or more computer-readable media storing instructions executable by the one or more processors, the instructions being a medium that, when executed, causes the system to perform an operation, the operation comprising: receiving a trajectory associated with a vehicle traversing an environment; determining a first safety for the vehicle based on the trajectory; determining a second safety area for the vehicle based on the trajectory; receiving sensor data from a sensor related to an environment in which the vehicle is traveling; determining a subset of sensor data associated with a safe area and the second safe area; detecting an object represented in the subset of sensor data; based at least in part on the subset of sensor data; , determining a likelihood that the object intersects the trajectory; determining a modified acceleration profile associated with the vehicle based on the likelihood; and controlling the vehicle based on the modified acceleration profile. A system that includes:

B: The system of paragraph A, wherein the trajectory includes crossing an intersection, and determining the second safe area is based at least in part on a portion of the environment that is substantially orthogonal to the trajectory. system.

C: The system of paragraph A or B, wherein the trajectory includes a left turn, and the second safety area includes a region of the environment closest to the lane in which the vehicle is currently located.

D: The system of any of paragraphs AC, wherein the second safety area is substantially orthogonal to the first safety area.

E: The system in any of paragraphs AD, wherein determining the orientation of the second safe area is based at least in part on a line tangent to a point associated with the first safe area.

F: A method comprising: receiving a trajectory associated with a vehicle; determining a first safety area based on at least a portion of the trajectory; determining a second safe area different from a safe area of the first or second safe areas; determining a subset of sensor data associated with one or more of the first or second safe areas; A method comprising: determining a likelihood that a trajectory of an object intersects a trajectory of the vehicle; and controlling the vehicle based at least in part on the likelihood.

G: The method of paragraph F, wherein controlling the vehicle includes determining a modified acceleration profile associated with the vehicle based at least in part on the trajectory of the object; and at least a portion of the modified acceleration profile. controlling the vehicle based on the method.

H: The method of paragraph F or G, wherein controlling the vehicle further includes controlling the vehicle to decelerate to a stop based on the likelihood matching or exceeding a threshold likelihood. .

I: In the method of any of paragraphs F-H, the width of the portion of the first safety area is such that the width of the portion of the first safety area corresponds to at least a portion of the representation of the vehicle's position along the first safety area at a future time. The method is based on and expanded upon.

J: In the method of any of paragraphs F to I, determining the second safety area includes determining a first boundary associated with the first safety area, determining an intersection of the first boundary and a line associated with a second boundary associated with the first safe area based on the intersection; and determining the second safe area based on the intersection. The method further comprising determining.

K: The method of any of paragraphs F-J, wherein the second safety area is substantially orthogonal to the first safety area.

L: In the method of any of paragraphs F-K, the first width of the first safety area is related to the width of the lane in which the vehicle is currently located, and the second width of the second safety area is , the width of the oncoming lane closest to the lane in which the vehicle is currently located.

M: In any of the methods of paragraphs F to L, determining the second safety area includes determining the trajectory associated with a left turn, and determining the second safety area when the vehicle is currently The method further comprises: determining a portion of the environment associated with an oncoming lane closest to the located lane.

N: In the method of any of paragraphs F to M, determining the second safe area includes determining the second safe area based on at least a portion of the speed of the vehicle that is below a threshold speed. The method further includes:

O: The method of any of paragraphs F to N, wherein determining the second safe area is further based on at least a portion of points related to the vehicle.

P: one or more non-transitory computer-readable media storing instructions that, when executed, cause one or more processors to perform operations, the operations causing the vehicle to receiving an associated trajectory; determining a first safe area based on at least a portion of the trajectory; a second safe area different from the first safe area based on at least a portion of the trajectory; determining a subset of sensor data associated with one or more of the first or second safety areas, wherein a trajectory of an object represented in the subset of sensor data is a trajectory of the vehicle; One or more non-transitory computer-readable media comprising: determining an intersecting likelihood; and controlling the vehicle based at least in part on the likelihood.

Q: In the one or more non-transitory computer-readable media of paragraph P, controlling the vehicle comprises modifying an acceleration profile associated with the vehicle based at least in part on the trajectory of the object. and controlling the vehicle based at least in part on the modified acceleration profile.

R: In one or more non-transitory computer-readable media in paragraphs P or Q, controlling the vehicle is based on the likelihood meeting or exceeding a threshold likelihood; One or more non-transitory computer-readable media further comprising a control to slow the vehicle to a stop.

S: In one or more non-transitory computer-readable media in any of paragraphs PR, the width of the portion of the first secure area is the width of the first secure area at a future time. one or more non-transitory computer-readable media expanded based on at least a portion of the representation of the vehicle's position along.

T: In any of paragraphs P-S, in one or more non-transitory computer-readable media, determining the second secure area includes a first secure area related to the first secure area. determining, based on at least a portion of the trajectory, an intersection of the first boundary and a line associated with a second boundary associated with the first safety area; , one or more non-transitory computer-readable media, further comprising determining the second safe area based on the intersection point.

U: A system having one or more processors and one or more computer-readable media storing instructions executable by the one or more processors, the instructions being executed. a medium for causing the system to perform operations when the system is configured to receive a trajectory associated with a vehicle; determining a safe area based on at least a portion of the trajectory; determining a plurality of segments associated with an area, one segment of the plurality of segments extending a first distance substantially orthogonally from the trajectory; , determining the position of a bounding box along the trajectory representing the vehicle at a future time; determining a second distance between a point of the bounding box and a point closest to the trajectory; determining that the second distance matches or exceeds the first distance; and determining a modified safety area based on the second distance that exceeds the first distance. , controlling the vehicle based on the modified safety area.

V: The system of paragraph U, wherein the position of the bounding box is based at least in part on a simulation of the vehicle traveling along the trajectory at the future time, and the points of the bounding box are A system associated with the corners of a bounding box.

W: The system of paragraph U or V, wherein the operations include: receiving sensor data from a sensor associated with the vehicle; determining a subset of sensor data based at least in part on the modified safety area; The controlling the vehicle further includes detecting an object represented in the subset of sensor data and determining a likelihood that the object will traverse the trajectory, and controlling the vehicle further comprises detecting an object represented in the subset of sensor data and determining a likelihood that the object will traverse the trajectory, and controlling the vehicle further comprises detecting an object represented in the subset of sensor data; Further based on the part, the system.

X: The system of any of paragraphs U-W, wherein the width of the safety area is related to the width of the vehicle.

Y: The system of any of paragraphs U-X, wherein the width of the modified safety area is related to the width of the lane in which the vehicle is currently located.

Z: A method, comprising: receiving a trajectory associated with a vehicle; determining a safe area associated with the vehicle based at least in part on the trajectory; and determining a safe area along the safe area at a future time. determining a maximum distance from the trajectory at the location of a point associated with the representation of the vehicle at a future time; and controlling the vehicle based at least in part on the safe area.

AA: The method of paragraph Z further comprises determining a plurality of segments associated with the trajectory, and wherein the position is associated with a segment of the plurality of segments.

AB: The method of paragraph Z or AA, wherein the position associated with the vehicle along the safety area is based at least in part on a simulation of the vehicle traveling along the trajectory at the future time; The method wherein the points are associated with corners of a bounding box representing the vehicle.

AC: The method of any of paragraphs Z-AB comprises: receiving sensor data from a sensor associated with the vehicle; determining a subset of the sensor data based on a modified safety area that includes the portion of the safety area; and detecting an object represented in the subset of sensor data and determining a likelihood that the object will traverse the trajectory, and controlling the vehicle further comprises: A method further based on at least a portion of said likelihood.

AD: The method of any of paragraphs Z-AC, wherein the maximum width of the safety area is related to the width of the vehicle.

AE: The method of any of paragraphs Z-AD, wherein the maximum width of the modified safety area that includes the portion of the safety area is related to the width of the lane in which the vehicle is currently located.

AF: The method of any of paragraphs Z-AE, wherein the representation of the vehicle includes a bounding box.

AG: The method of any of paragraphs Z-AF comprises: receiving sensor data; determining a subset of the sensor data based at least in part on the safe area; and determining a likelihood that the object will cross the trajectory, wherein controlling the vehicle is further based at least in part on the likelihood.

AH: The method of paragraph AG, wherein controlling the vehicle includes modifying one or more acceleration or steering commands to reduce the likelihood of crossing.

AI: The method of any of paragraphs Z-AH further comprises receiving sensor data from a sensor associated with the vehicle, wherein the sensor data is LIDAR data, camera data, radar data, ultrasound data, or A method that includes depth data.

AJ: one or more non-transitory computer-readable media storing instructions that, when executed, cause one or more processors to perform operations, the operations causing the vehicle to receiving an associated trajectory; determining a safety area associated with the vehicle based at least in part on the trajectory; determining a position associated with the vehicle along the safety area at a future time; , determining a maximum distance from the trajectory at the location of a point associated with a representation of the vehicle at a future time, defining a width of the portion of the safe area as the maximum distance at the location; controlling the vehicle based at least in part on the safety area.

AK: In paragraph AJ, on one or more non-transitory computer-readable media, the act further includes determining a plurality of segments associated with the trajectory, and the position is determined by determining a plurality of segments associated with the trajectory. one or more non-transitory computer-readable media associated with a segment of the computer-readable medium;

AL: in one or more non-transitory computer-readable media in paragraphs AJ or AK, the position associated with the vehicle along the safety area is determined by traveling along the trajectory at the future time. one or more non-transitory computer-readable media based at least in part on a simulation of the vehicle, wherein the points are associated with corners of the bounding box.

AM: In one or more non-transitory computer-readable media according to any of paragraphs AJ-AL, the operations include receiving sensor data from a sensor associated with the vehicle; determining the subset of sensor data based on at least a portion of a modified safety area comprising a portion; detecting the object represented in the subset of sensor data; and detecting the object represented in the subset of sensor data; determining a likelihood of traversing one or more non-transitory computer-readable media, wherein controlling the vehicle is further based at least in part on the likelihood.

AN: In one or more non-transitory computer-readable media in any of paragraphs AJ-AM, the maximum width of the safe area is one or more non-transitory computer-readable media that are related to the width of the vehicle. computer-readable medium.

Although the example sections described above are described with respect to one particular implementation, in the context of this document, the subject matter of the example sections may be implemented via a method, device, system, computer-readable medium, and/or another implementation. It should be understood that this is also possible. Additionally, any of Examples A-AN can be practiced alone or in combination with any one or more of the other Examples A-AN.

Summary Although one or more examples of the technology described herein have been described, various modifications, additions, substitutions, and equivalents thereof are included within the scope of the technology described herein. It will be done.

The description of the examples refers to the accompanying drawings, which constitute a part of the specification, and which illustrate particular embodiments of the claimed subject matter. It is to be understood that other examples may be used and changes or modifications, such as structural changes, may be made. Such examples, changes or modifications do not necessarily depart from the scope of the intended claimed subject matter. Although the procedures herein may be presented in a certain order, in some cases the order may be arranged such that certain inputs are provided at different times or in a different order without changing the functionality of the systems and methods described. subject to change. Also, the disclosed procedures may be performed in a different order. Furthermore, the various calculations described herein need not be performed in the order disclosed, and other examples using alternative orders of calculations may be readily implemented. In addition to reordering, calculations may also be decomposed into sub-calculations with the same result.

Claims (15)

receiving a trajectory associated with the vehicle;
determining a first safe area based on at least a portion of the trajectory;
determining a second safe area different from the first safe area based on at least a portion of the trajectory;
determining a subset of sensor data associated with one or more of the first safety area or the second safety area;
determining the likelihood that a trajectory of an object represented in the subset of sensor data intersects the trajectory of the vehicle;
controlling the vehicle based at least in part on the likelihood;
A method of providing. The method includes:
receiving from a sensor the sensor data related to the environment in which the vehicle is traveling; and
further comprising detecting the object represented in the subset of the sensor data;
2. The method of claim 1, wherein the subset of sensor data is related to the first safety area and the second safety area. Controlling the vehicle includes:
determining a modified acceleration profile associated with the vehicle based at least in part on the trajectory of the object; and
controlling the vehicle based at least in part on the modified acceleration profile;
The method according to claim 1 or 2, comprising: Determining the second safe area comprises:
determining a first boundary associated with the first safe area;
determining an intersection between the first boundary and a line associated with a second boundary associated with the first safety area based on at least a portion of the trajectory; and
determining the second safe area based at least in part on the intersection points;
4. A method according to any one of claims 1 to 3, further comprising: The method includes:
expanding a width of a portion of the first safety area based at least in part on a representation of the vehicle at a position along the first safety area at a future time;
determining the second safe area based at least in part on a speed of the vehicle below a threshold speed; or
determining the second safety area based at least in part on points associated with the vehicle;
5. A method according to any one of claims 1 to 4, comprising at least one of: The method includes:
determining a plurality of segments extending a first distance substantially perpendicular to the trajectory and associated with the first safe area;
determining a position along the trajectory of a bounding box representing the vehicle at a future time;
determining a second distance between a point of the bounding box and a point closest to the trajectory;
determining that the second distance matches or exceeds the first distance; and
determining a modified safety area based on at least a portion of the second distance that exceeds the first distance;
6. The method according to any one of claims 1 to 5, further comprising: The method includes:
i) determining a position along the trajectory of a bounding box representing the vehicle at a future time, the position being based at least in part on a simulation of the vehicle traveling along the trajectory at the future time; thing,
ii) defining a width of a portion of said first safety area as being related to a first width of said vehicle, or defining a second width of a lane in which said vehicle is currently located;
iii) determining a position at a future time of a bounding box representing the vehicle relative to one of a plurality of segments associated with the trajectory; or
iv) determining a position associated with the vehicle along the first safety area based at least in part on a simulation of the vehicle traveling along the trajectory at a future time; is represented by a bounding box at said future time, and at least a portion of said bounding box is based on the distance between a point associated with a corner of said bounding box and the nearest point in said trajectory. used to determine a modified safe area that includes a part of the safe area of 1;
7. The method according to any one of claims 1 to 6, further comprising at least one of: The method includes:
further comprising determining a modified safe area that includes a portion of the first safe area;
the maximum width of the modified safety area is related to the width of the lane in which the vehicle is currently located, and
8. A method according to any one of claims 1 to 7, wherein the subset of sensor data relates to the modified safety area. 9. A computer program product comprising coded instructions which, when executed on a computer, implement the method according to any of claims 1 to 8. A system,
one or more processors;
one or more non-transitory computer-readable media storing instructions executable by the one or more processors, the instructions, when executed, causing the system to:
receiving a trajectory associated with the vehicle;
determining a first safe area based on at least a portion of the trajectory;
determining a second safe area different from the first safe area based on at least a portion of the trajectory;
determining a subset of sensor data associated with one or more of the first safety area or the second safety area;
determining the likelihood that a trajectory of an object represented in the subset of sensor data intersects the trajectory of the vehicle;
controlling the vehicle based at least in part on the likelihood;
A system that executes processes including the second safe area includes an area of the environment proximate to the lane in which the vehicle is currently located; or
determining the direction of the second safety area is based at least in part on a tangent of a point associated with the first safety area;
11. The system of claim 10, wherein the system is at least one of: determining a position associated with the vehicle along the first safety area at the future time based at least in part on a simulation of the vehicle traveling along the trajectory at the future time;
determining a maximum distance from the trajectory at the location of a point associated with the representation of the vehicle at the future time; and
12. The system of claim 10 or 11, further comprising: defining a width of a portion of the first safe area as the maximum distance at the location. i) determining the second safe area is based at least in part on a portion of the environment that is substantially orthogonal to the trajectory, and the trajectory includes a trajectory passing through an intersection; or
ii) determining the second safe area includes, as the second safe area and based at least in part on the determination that the trajectory is associated with a left turn, a lane in which the vehicle is currently located; further comprising: determining a portion of the environment associated with oncoming traffic;
13. A system according to any one of claims 10 to 12, wherein the system is at least one of: controlling the vehicle includes modifying one or more acceleration or steering commands to reduce the likelihood of crossing; or
The system according to any one of claims 10 to 13, wherein controlling the vehicle includes controlling the vehicle to stop based at least in part on a likelihood that matches or exceeds a threshold likelihood. . 15. A system according to any one of claims 10 to 14, wherein the sensor data includes one or more of LIDAR data, camera data, ultrasound data, or depth data.

Images