Planning system for autonomous navigation around lane-sharing road factors

Abstract

A system for estimating a spacing profile of a road factor includes a first module and a second module. The first module includes instructions that cause one or more processors to receive data related to road factor behavior detected in an environment of an autonomous vehicle and characteristics of the road factor, initiate an analysis of the road factor behavior, and estimate a spacing profile of the road factor as part of the analysis. The spacing profile includes a lateral gap preference and one or more predicted behaviors of the road factor related to a change in the lateral gap. The second module includes instructions that cause one or more processors to determine one or more components of a maneuver of the autonomous vehicle based on the estimated spacing profile and send control instructions for performing the maneuver of the autonomous vehicle.

Description

[0001] Cross-reference to related applications

[0002] This application claims the benefit of U.S. Provisional Application No. 63 / 236,541, filed on August 24, 2021, the entire disclosure of which is incorporated herein by reference. Technical Field

[0003] This disclosure relates to the field of autonomous driving, and more specifically, to systems and methods for estimating spacing profiles of road factors. Background Technology

[0004] Autonomous vehicles, such as those that do not require a human driver, can be used to help transport passengers or goods from one location to another. Such vehicles can operate in a fully autonomous mode, where passengers can provide some initial input, such as their destination, and the vehicle maneuvers itself to that destination. Therefore, such vehicles may depend heavily on systems capable of determining the autonomous vehicle's location at any given time and detecting and recognizing objects outside the vehicle, such as other vehicles, stop lights, pedestrians, etc.

[0005] Data from one or more of these systems can be used to detect objects and their corresponding characteristics (positioning, shape, heading, speed, etc.). These characteristics can be used to predict the trajectories of other objects. These trajectories can define what the objects might do in a short period of time in the future. These trajectories can then be used to control the vehicle to avoid these objects. Therefore, detection, identification, and prediction are key functions for the safe operation of autonomous vehicles. Summary of the Invention

[0006] This disclosure provides a system for estimating a spacing profile of road factors, the system comprising: a first module including instructions that cause one or more processors to: receive data relating to road factor behavior and characteristics detected in the environment of an autonomous vehicle; initiate analysis of the road factor behavior; and estimate a spacing profile of the road factors as part of the analysis, the spacing profile including lateral clearance preferences and one or more predicted behaviors of the road factors relating to changes in lateral clearance; and a second module including instructions that cause one or more processors to determine one or more components of autonomous vehicle maneuvering based on the estimated spacing profile; and send control instructions for performing autonomous vehicle maneuvering.

[0007] In one example, the analysis is initiated based on when the road factor is predicted to interact with the autonomous vehicle. In another example, the analysis is initiated based on the autonomous vehicle's lateral clearance requirement with respect to the road factor. In yet another example, the analysis is initiated based on whether the road factor is performing an overtake maneuver by the autonomous vehicle. In yet another example, the analysis is initiated based on the existing lateral clearance between the autonomous vehicle and the road factor. In yet another example, the analysis is initiated based on a machine learning model.

[0008] In another example, the predicted behavior of one or more road factors includes overtaking maneuvers or cut-off maneuvers associated with the autonomous vehicle. In yet another example, the spacing profile includes a predictability score for the road factors. In this example, the spacing profile includes a predictability score for the autonomous vehicle. In yet another example, the system also includes a third module comprising instructions that cause one or more processors to: receive control instructions; and an operating system that controls the autonomous vehicle based on the control instructions. In this example, the system also includes an operating system. Furthermore, in this example, the system also includes an autonomous vehicle.

[0009] Other aspects of this disclosure provide a method for estimating a spacing profile of road factors. The method includes: receiving data associated with road factor behavior and characteristics detected in the environment of an autonomous vehicle by one or more computing devices; initiating an analysis of the road factor behavior by one or more computing devices; estimating a spacing profile of the road factors by one or more computing devices as part of the analysis, the spacing profile including lateral clearance preferences and one or more predicted behaviors of the road factors associated with changes in lateral clearance; determining one or more components of autonomous vehicle maneuvering by one or more computing devices based on the estimated spacing profile; and sending control commands by one or more computing devices to perform autonomous vehicle maneuvering.

[0010] In one example, the analysis is initiated based on when the road factor is predicted to interact with the autonomous vehicle. In another example, the analysis is initiated based on the autonomous vehicle's lateral clearance requirement with respect to the road factor. In yet another example, the analysis is initiated based on whether the road factor is performing an overtaking maneuver by the autonomous vehicle. In yet another example, the analysis is initiated based on the existing lateral clearance between the autonomous vehicle and the road factor. In yet another example, the analysis is initiated based on a machine learning model.

[0011] In another example, one or more predicted behaviors for road factors include overtaking maneuvers or lane-changing maneuvers associated with the autonomous vehicle. In yet another example, the spacing profile includes a predictability score for either road factors or the autonomous vehicle. Attached Figure Description

[0012] Figure 1 These are functional diagrams of example vehicles based on various aspects of this disclosure.

[0013] Figure 2 It is based on example map information from various aspects of this disclosure.

[0014] Figure 3 These are example representative views of vehicles according to various aspects of this disclosure.

[0015] Figure 4 This is a flowchart of example methods based on various aspects of this disclosure.

[0016] Figure 5 Example scenarios based on various aspects of this disclosure are shown. Detailed Implementation

[0017] Overview

[0018] This technology relates to planning systems for autonomous vehicles that take into account road factors such as lane-sharing with cyclists. For scenarios where autonomous vehicles share lateral space with road factors, a one-size-fits-all solution may cause the autonomous vehicle to slow down more than needed and travel less along the route, or may fail to create the optimal or safest outcome for every possible scenario. Instead, in addition to safety parameters / requirements, the behavior of detected road factors can be used to adapt the amount of lateral clearance required to provide comfortable forward travel for the autonomous vehicle in a specific scenario.

[0019] The vehicle's computing devices can detect road factors and their behavior over time. The vehicle's computing devices can initiate an analysis of the road factor behavior based on the detected characteristics and behavior. For example, the analysis can be initiated when predicting the interaction between road factors and the autonomous vehicle. The analysis of road factor behavior is used to estimate a spacing profile of the road factors. The spacing profile includes lateral clearance preferences and one or more predicted behaviors of the road factors associated with changes in lateral clearance. In some embodiments, the spacing profile also includes a predictability score for the road factors, which can be based on the degree to which the behavior has matched previously predicted behavior and / or the predictability of environmental factors.

[0020] Based on the estimated spacing profile, the vehicle's computing equipment can determine autonomous vehicle handling. To perform this determination, the vehicle's computing equipment can update one or more constraints based on the spacing profile. These constraints can then be used to determine one or more components of the vehicle handling, such as speed, path, or route. The vehicle's computing equipment can then execute the determined vehicle handling by correspondingly controlling one or more operating systems of the autonomous vehicle.

[0021] The technology described in this article allows passengers to make smoother and safer journeys in autonomous vehicles. In particular, more forward travel is possible in autonomous vehicles compared to situations without the aforementioned navigation system. The technology also minimizes unnecessary overreaction and underreaction by autonomous vehicles to small lateral clearances, maintaining a level of safety and compliance requirements while ensuring comfort for cyclists, scooter riders, motorcyclists, pedestrians, runners, and other road users.

[0022] Example System

[0023] like Figure 1 As shown, a vehicle 100 according to one aspect of this disclosure includes various components. While certain aspects of this disclosure are particularly useful in conjunction with specific types of vehicles, the vehicle can be any type of vehicle, including but not limited to cars, trucks, motorcycles, buses, recreational vehicles, etc. The vehicle may have one or more computing devices, such as a computing device 110 including one or more processors 120, memory 130, and other components typically found in general-purpose computing devices.

[0024] One or more processors 120 can be any conventional processor, such as a commercially available CPU. Alternatively, one or more processors can be special-purpose devices, such as ASICs or other hardware-based processors.

[0025] Memory 130 stores information accessible by one or more processors 120, including instructions 132 and data 134 that can be executed or otherwise used by the processors 120. Memory 130 can be any type of memory capable of storing processor-accessible information, including computing device-readable media, or other media that store data readable by electronic devices, such as hard disk drives, memory cards, ROM, RAM, DVDs or other optical discs, and other writable and read-only memories. Systems and methods can include different combinations of the foregoing, whereby different portions of instructions and data are stored on different types of media.

[0026] The memory 130 may store various models used by the computing device 110 to make determinations about how to control the vehicle 100. For example, the memory 130 may store one or more object recognition models for identifying objects and road users detected from sensor data. As another example, the memory 130 may store one or more behavioral models for providing the probability of detected objects taking one or more actions. As yet another example, the memory 130 may store one or more speed planning models for determining the speed profile of the vehicle 100 based on map information and predicted trajectories of other road users detected by sensor data.

[0027] Instructions 132 can be any set of instructions that can be executed directly (such as machine code) or indirectly (such as scripts) by a processor. For example, instructions can be stored as computing device code on a computing device readable medium. In this regard, the terms "instruction" and "program" are used interchangeably herein. Instructions can be stored in an object code format for direct processing by a processor, or in any other computing device language, including a collection of independent source code modules or scripts that are interpreted on demand or pre-compiled. In some implementations, instructions 132 may include a plurality of modules 180, wherein each module may include one or more routines that operate independently of other modules. The functionality, methods, and routines of the instructions are explained in more detail below.

[0028] Processor 120 can retrieve, store, or modify data 134 according to instruction 132. For example, while the claimed subject matter is not limited to any particular data structure, the data can be stored in a computing device register, in a relational database as a table with multiple different fields and records, an XML document, or a flat file. The data can also be formatted in any computing device-readable format.

[0029] although Figure 1 While the processor, memory, and other components of computing device 110 are shown functionally within the same frame, those skilled in the art will understand that a processor, computing device, or memory may actually include multiple processors, computing devices, or memories that may or may not be housed in the same physical enclosure. For example, memory may be a hard disk drive or other storage medium located in a different enclosure than that of computing device 110. Therefore, references to processors or computing devices will be understood to include references to a collection of processors or computing devices or memories that may or may not operate in parallel.

[0030] The computing device 110 may include all components typically used in conjunction with a computing device, such as the processor and memory described above, as well as user input 150 (e.g., mouse, keyboard, touchscreen, and / or microphone) and various electronic displays (e.g., a monitor with a screen or any other electrical device operable to display information). In this example, the vehicle includes an interior electronic display 152 and one or more speakers 154 to provide information or an audiovisual experience. In this regard, the interior electronic display 152 may be located within the cabin of the vehicle 100 and may be used by the computing device 110 to provide information to passengers within the vehicle 100.

[0031] The computing device 110 may also include one or more wireless network connections 156 to facilitate communication with other computing devices, such as client computing devices and server computing devices described in detail below. The wireless network connection may include short-range communication protocols such as Bluetooth, Bluetooth Low Energy (LE), cellular connections, and various configurations and protocols including the Internet, the World Wide Web, intranets, virtual private networks, wide area networks, local area networks, private networks using one or more company-specific communication protocols, Ethernet, WiFi, and HTTP, as well as various combinations of the foregoing.

[0032] In one example, computing device 110 may be an autonomous driving computing system integrated into vehicle 100. The autonomous driving computing system may be able to communicate with various components of the vehicle to operate vehicle 100 in fully autonomous driving mode and / or semi-autonomous driving mode. For example, returning... Figure 1 The computing device 110 can communicate with various operating systems of the vehicle 100 to control the movement, speed, etc. of the vehicle 100 according to instructions 132 from the memory 130. These operating systems include a deceleration system 160, an acceleration system 162, a steering system 164, a signaling system 166, a navigation system 168, a positioning system 170, a sensing system 172, and a power system 174 (e.g., a gasoline or diesel-powered motor or an electric motor). Similarly, although these systems are shown external to the computing device 110, they can actually be integrated into the computing device 110, also serving as an autonomous driving computing system for controlling the vehicle 100.

[0033] As an example, computing device 110 may interact with deceleration system 160 and / or acceleration system 162 to control the speed of the vehicle. Similarly, computing device 110 may use steering system 164 to control the direction of vehicle 100. For example, if vehicle 100 is configured for use on a road, such as a car or truck, the steering system may include components that control the angle of the wheels to turn the vehicle. Computing device 110 may use signaling system 166 to signal the vehicle's intentions to other drivers or vehicles (e.g., by illuminating turn signals or brake lights when necessary).

[0034] The computing device 110 can use the navigation system 168 to determine and follow a route to a location. For example, the navigation system can be used to generate routes between locations and plan trajectories for vehicles to follow. Although described as a single system, the navigation system can actually include multiple systems to achieve the routing and planning functions described above. In this regard, the navigation system 168 and / or data 134 can store detailed map information, such as highly detailed maps identifying roads, lane lines, intersections, pedestrian crossings, speed limits, traffic signals, buildings, signs, real-time traffic information, the shape and height of vegetation, or other such objects and information.

[0035] In other words, the detailed map information can define the geometry of the vehicle's expected environment, including roads and their speed limits (legal speed limits). Specifically, the map information can include a road map defining the geometry of road features such as lanes, medians, curbs, crosswalks, etc. As an example, the road map can include multiple points and / or line segments connected to each other, thereby defining the geometry (e.g., size, shape, dimension, and location) of the aforementioned road features. The road map can also include information identifying how the vehicle is expected to travel on a given road, including direction (i.e., the legal traffic direction in each lane), lane location, speed, etc. For example, the map information can include information about traffic controls (such as traffic lights, stop signs, yield signs, etc.). This information, combined with real-time information received from the perception system 172, can be used by the computing device 110 to determine which traffic directions are oncoming lanes and / or have right-of-way at a given location.

[0036] Figure 2This is an example of map information 200 that includes a segment of road at intersection 230. In this example, map information 200 depicts a portion of map information including information identifying the shape, location, and other characteristics of various features. For example, map information 200 includes roads 210 and 220 intersecting at intersection 230. Map information 200 includes lane markings or lane lines 241A and 243A of road 210 on the first side of intersection 230, and lane lines 241B and 243B of road 210 on the second side of intersection 230 opposite to the first side. Furthermore, map information includes lane lines 242 and 246 of road 220 passing through intersection 230 from the third side to the fourth side opposite to the third side, lane line 244A of road 220 on the third side of intersection 230, and lane line 244B of road 220 on the fourth side of intersection 230. Lane markings can be of different types, such as two-lane lane markings 241A, 241B, 244A, and 244B, and zigzag lane markings 242, 243A, 243B, and 246. Lane markings can also define various lanes, such as lanes 251, 252, 253, 254, 255, 256, and 258.

[0037] Lane portions 251A, 253A, and 255A of road 210 are on the first side of intersection 230, and lane portions 251B, 253B, and 255B of road 210 are on the second side of intersection 230 opposite to the first side. Lane portions 252A, 254A, 256A, and 258A of road 220 are on the third side of intersection 230, and lane portions 252B, 254B, 256B, and 258B of road 220 are on the fourth side of intersection 230 opposite to the third side. Lanes can be explicitly marked in map information 200 as shown, or can be implied by road width. Map information 200 can also mark bicycle lanes. As shown, map information 200 can also include stop lines 261 and 263 of road 210. Stop line 261 can be associated with stop sign 265, and stop line 263 can be associated with stop sign 267.

[0038] In addition to these features, map information 200 may also include information identifying speed limits and traffic directions for each lane, as well as information that allows computing device 110 to determine whether a vehicle has the right-of-way to complete a specific maneuver (e.g., to complete a turn or cross a traffic lane or intersection). Map information 200 may also include information about traffic signs (such as traffic lights, stop signs, one-way signs, no-turn signs, etc.). Map information 200 may include information about other environmental features (such as curbs, buildings, parking lots, driveways, waterways, vegetation, etc.).

[0039] Although detailed map information is described in this paper as image-based maps, map information does not need to be entirely image-based (e.g., raster). For example, detailed map information can include a graphical network or one or more road maps containing information such as roads, lanes, intersections, and the connections between these features. Each feature can be stored as graphical data and can be associated with information such as geographic location and whether it is linked to other related features (e.g., stop signs can be linked to roads and intersections). In some examples, the associated data can include a grid-based index of the road map to allow for efficient lookup of certain road map features.

[0040] The perception system 172 also includes one or more components for detecting objects outside the vehicle, such as other road factors, obstacles in the road, traffic signals, signs, trees, etc. Other road factors may include cyclists, scooters, motorcycles, pedestrians, or runners. For example, the perception system 172 may include one or more imaging sensors, including visible light cameras, thermal imaging systems, laser and radio frequency detection systems (e.g., LIDAR, RADAR, etc.), sonar equipment, microphones, and / or any other detection devices that can record data that can be processed by the computing device 110.

[0041] When detecting objects, one or more imaging sensors of perception system 172 can detect their characteristics and behaviors, such as position (longitudinal and lateral distances relative to the vehicle), orientation, size, shape, type, direction / following, trajectory, lateral movement, speed of movement, acceleration, etc. The raw data from the sensors and / or the aforementioned characteristics can be quantized or arranged into descriptive functions or vectors and sent to computing device 110 for further processing. For example, computing device 110 can use positioning system 170 to determine the vehicle's position, and use perception system 172 to detect objects and respond to them as needed to follow a route or safely reach a destination.

[0042] Figure 3This is an example external view of vehicle 100. In this example, the roof-top 310 and dome-shaped housing 312 may include LIDAR sensors as well as various camera and radar units. Furthermore, housing 320 located at the front of vehicle 100 and housings 330, 332 on the driver's and passenger's sides of the vehicle may each house LIDAR sensors or systems. For example, housing 330 is located in front of the driver's door 360. Vehicle 100 also includes housings 340, 342 for radar units and / or cameras, also located on the roof of vehicle 100. Additional radar units and cameras (not shown) may be located at the front and rear of vehicle 100 and / or at other locations along the roof or roof-top housing 310. Vehicle 100 also includes many features typical of passenger vehicles, such as doors, wheels, windows, etc.

[0043] Example Method

[0044] In addition to the operations described above and in the figures, various other operations will now be described. The computing device 110 can detect road factors in the vehicle environment, such as a cyclist, and adjust one or more systems of the autonomous vehicle 100 based on the detected road factors. For example, in Figure 4 The diagram below shows a flowchart 400 for performing the adjustment. It should be understood that the following blocks or operations do not necessarily need to be performed in the exact order described below. Instead, the steps can be processed in different orders or simultaneously, and steps can be added or omitted.

[0045] In box 402, the vehicle's computing device 110 can use the perception system 172 to detect road factors and road factor behavior. The proximity of the vehicle 100 can be defined by the range of sensors and other detection systems of the vehicle 100's perception system 172. Sensor data obtained from the perception system 172 can include object data defining the cyclist 510. The vehicle's computing device 110 can use the object data along with characteristics of the road factors to identify the road factors. For example, road factors with a given location, pose, orientation, dimension / size, shape, speed, direction / following, trajectory, lateral movement, acceleration, or other positioning characteristics can be detected. Characteristics can also include physical characteristics such as estimated age, size, hand signals, light signals, type of clothing, type of bicycle, etc. For example, certain age groups such as children or the elderly may be associated with a greater preference for lateral clearance and a slower reaction time. Certain vehicles such as road bikes may be associated with a smaller preference for lateral clearance and a higher speed.

[0046] In addition to detecting rail factors, the vehicle's computing device 110 can also detect multiple objects in the vicinity of the vehicle. For example, sensor data from the perception system 172 may also include characteristics of each object, such as its size, shape, speed, orientation, and direction. Multiple objects may include moving and / or stationary objects. In particular, multiple objects may include other road users (such as vehicles, bicycles, or pedestrians), other types of obstacles (such as buildings, pillars, trees, or construction tools), or traffic features such as lights, signs, lane markings, curbs, or rail tracks.

[0047] exist Figure 5 In the scenario 500 depicted, vehicle 100 may be approaching stop line 263 in lane portion 251B of road 210 adjacent to intersection 230. Vehicle 100 may have a planned maneuver 502 for turning right at intersection 230. As shown by the dashed line, the planned maneuver 502 of the vehicle includes turning right from lane portion 251B through intersection 230 into lane portion 258A. The vehicle's computing device 110 can use perception system 172 to detect cyclists 510, vehicles 512, 514, buses 516, lane lines 241A, 241B, 243A, 243B, 242, 244A, 244B, 246, stop lines 261, 263, stop signs 265, 267, and the characteristics of the detected objects and features based on the location of vehicle 100. The characteristics of the detected objects and features may include the type of lane lines, the geometry of the lane lines and rail tracks, the position and pose of vehicles (vehicle 512 in lane section 252A, vehicle 514 in lane section 254A, and bus 516 in lane section 253A near the stop line 261), the trajectory of the vehicles (towards the intersection 230), the shape of the sign (octagonal), and the position and orientation of the sign. The vehicle's computing device 110 can also detect the flashing lights on the bus 516, particularly the left-turn flasher 511. The characteristics of the cyclist 510 may include at least their position (relative to vehicles (such as behind and to the right of vehicle 100) and relative to the road geometry (such as in lane section 251B near the curb)) and their direction of travel (from lane section 251B towards the intersection 230).

[0048] In box 404, the vehicle's computing device 110 can initiate an analysis of the behavior of road factors based on detected characteristics and behaviors. For example, the analysis can be initiated when a road factor is predicted to interact with the autonomous vehicle. Interacting with the autonomous vehicle means taking the autonomous vehicle's planning process into account as a factor or otherwise influencing the autonomous vehicle's operation. Predictions may include multiple characteristics of a road factor being projected to laterally overlap with or overtake the autonomous vehicle, multiple characteristics of scenarios associated with potential interactions with the autonomous vehicle detected by the vehicle's computing device, and / or the detected characteristics of scenarios satisfying a set of heuristics. Predictions may also be based on fundamental requirements, rules, or conventions of the road factor; for example, predictions may consider at least satisfying lateral clearances in local regulatory rules and operational design domains. When a road factor is predicted to interact with the autonomous vehicle, an analysis of the road factor's behavior (described below) can be initiated. Predictions may also include threshold probabilities (such as percentages or scores) for the interaction of the road factor with the autonomous vehicle to trigger the initiation of the analysis.

[0049] In some examples, initiating the analysis may include determining the future trajectory of a road factor based on its position relative to the autonomous vehicle, its past or current direction, or its past or current speed. For example, the analysis may be initiated when a road factor's future trajectory overtakes an autonomous vehicle. Overtaking an autonomous vehicle can be characterized by a road factor traveling in the same direction / orientation as the autonomous vehicle over a time period, having a higher speed than the autonomous vehicle, and / or having an increased amount of lateral overlap with the autonomous vehicle. Figure 5 In the illustrated scenario, the characteristics of the cyclist 510 detected by the perception system 172 include being currently behind vehicle 100, having the same direction / direction as the autonomous vehicle 100 for a threshold time period, and having a higher speed than the autonomous vehicle for a threshold time period. Based on these characteristics, the vehicle's computing device 110 can determine that the cyclist 510 may be performing an overtaking maneuver to pass vehicle 100, and that the cyclist 510 has a threshold probability of interacting with vehicle 100. As a result, the vehicle's computing device 110 can initiate an analysis of the cyclist 510.

[0050] In other examples, the initiation analysis may include using a machine learning model to predict how road factors will interact with the autonomous vehicle.

[0051] Once started, the analysis may continue to iterate until there are no longer any potential interactions between the road factors and the autonomous vehicle. For example, when the road factors are outside the maximum distance from the autonomous vehicle, turn onto a different street than the autonomous vehicle, or have stopped, there are no longer any potential interactions.

[0052] Analysis of road factor behavior is used to estimate the spacing profile of road factors. The spacing profile includes lateral clearance preferences and one or more predicted behaviors of road factors associated with changes in lateral clearance. In some implementations, the spacing profile also includes a predictability score for the road factors, which may be based on the degree to which the behavior has matched previously predicted behavior and / or the predictability of environmental factors.

[0053] In box 406, the vehicle's computing device 110 can estimate the spacing profile of road factors as part of the analysis. The estimation of the spacing profile can be based on existing lateral clearance between the autonomous vehicle and the road factors, or changes in lateral clearance over time. Existing lateral clearance and changes in lateral clearance can be determined using the detected locations of the road factors. For example, Figure 5 Cyclist 510 is detected as having maintained approximately a first lateral clearance over the past few seconds or iterations. Cyclist 510's lateral clearance preference can be initially set to at least the first lateral clearance, and the cyclist's predicted behavior could be that the cyclist will maintain the first lateral clearance in most situations. On the other hand, if the cyclist has been decreasing the lateral clearance from the first lateral clearance over the past few seconds or iterations, it can be determined that the lateral clearance preference is less than the first lateral clearance, and the cyclist's predicted behavior could be to continue decreasing the lateral clearance until the preferred lateral clearance is reached.

[0054] Other factors used to estimate spacing profiles include context, road factor gaze / awareness, and road factor characteristics. Context can be extracted from map information and may include lane width (because narrower lanes may make road factors more comfortable with smaller lateral clearances), adjacent lane type or boundary type (because cyclists in or next to bike lanes may be more able to react to autonomous vehicles making lateral nudges), or speed limits (the basis for determining the likely speed of road factors). Other contextual factors can be detected using perception systems; for example, traffic density and traffic speed. Figure 5 In the scenario, cyclist 510 is traveling along a trajectory between vehicle 100 and the curb. Based on the location of the curb, the lateral clearance preference in the spacing profile can be less than the distance between vehicle 100 and the curb, and the predicted behavior in the spacing profile can include the probability that cyclist 510 will respond to a reduction in the distance between vehicle 100 and the curb, as well as one or more possible responses.

[0055] Perception systems can be used to detect road factor gaze to track the direction and focus of human eyes associated with road factors, and road factor gaze can be used to determine road factor awareness of autonomous vehicles. Road factor awareness can be defined by the time or frequency with which road factors have been viewed by the autonomous vehicle in the last few seconds or iterations. Figure 5 In the scenario, cyclist 510's gaze is detected as directed toward autonomous vehicle 100 within a first threshold time period, indicating that cyclist 510 is aware of the autonomous vehicle. Therefore, cyclist 510's spacing profile may include a higher probability of reducing lateral clearance and / or a lateral clearance preference, and cyclist 510's predicted behavior may include a faster reaction time to autonomous vehicle 100. When the cyclist's gaze toward the autonomous vehicle is less than the first threshold time period, it can be determined that the cyclist is not fully aware of the autonomous vehicle, and the cyclist's predicted behavior may be a slower reaction and more overreaction to the movement of the autonomous vehicle. When the cyclist's gaze toward the autonomous vehicle exceeds a second threshold time period greater than the first threshold time period, it can be determined that the cyclist is highly aware of the autonomous vehicle, and the cyclist's predicted behavior may be more sensitive to the movement of the autonomous vehicle.

[0056] A generalization of preferences based on one or more physical characteristics of road factors can also be used to determine lateral clearance preferences and predict behavior.

[0057] Specifically regarding predicted behavior, these might include whether road factors are attempting to overtake or cut in front of the autonomous vehicle, which could be based on the speed difference between the road factors and the autonomous vehicle. That is, if the autonomous vehicle is slower than the road factors, the predicted behavior might include overtaking maneuvers that could be associated with a larger reaction time and / or a smaller lateral clearance preference. As mentioned above, Figure 5 The cyclist 510 can determine whether to overtake the autonomous vehicle 100 based on the cyclist 510's detection of the cyclist 510's direction and speed. Therefore, when the cyclist 510 passes between the autonomous vehicle 100 and the curb, the cyclist 510's predicted behavior can include a future preference for lateral clearance that may be reduced.

[0058] Predicted behavior can be predicted from detected behavior. Furthermore, predicted behavior can be based on predicted autonomous vehicle maneuvers and possible responses to those maneuvers. For example, if the autonomous vehicle might need to slightly maneuver into an oncoming lane with traffic in order to maintain the current trajectory preference or default lateral clearance of the road factors, it can be inferred that the autonomous vehicle might need to slightly maneuver into an oncoming lane with traffic. The system can assume that the road factors are aware that the vehicle is unlikely to slightly maneuver into an oncoming lane with traffic, and therefore determine that the road factors will be comfortable with a smaller lateral clearance to the autonomous vehicle. Figure 5 In the scenario described, the predicted behavior of cyclist 510 can also be determined based on the fact that autonomous vehicle 100 is unlikely to micro-move into lane 253 in response to overtaking maneuvers. In another example, it can be inferred that the autonomous vehicle may have to micro-move around double-parked vehicles, and the system can determine that the cyclist may not be aware of the double-parked vehicles. Therefore, the system can determine that a larger lateral clearance is required. In other cases, a smaller lateral clearance can be used when the behavior of the autonomous vehicle is predictable and likely clear to the cyclist. The predicted autonomous vehicle maneuvers can take into account the lateral clearance between the autonomous vehicle and other vehicles in the environment.

[0059] In box 408, the vehicle's computing device 110 can determine autonomous vehicle maneuvering based on an estimated spacing profile. To perform the determination, the vehicle's computing device can update one or more constraints based on the spacing profile. For example, constraints such as those related to the vehicle's lateral clearance preferences can be added, moved, changed / adjusted, or removed. Some constraints may include permeability features that can be updated. Constraints can be updated so that the effective vehicle's lateral clearance preferences can be more closely matched to the lateral clearance preferences of road factors, which are determined from real-world data in the same or similar manner as described above and can be updated in real time in the same or similar manner. The permeability of speed constraints can be based on the predictability score of road factors or specific road factor behavior. In particular, a higher permeability can be initially set for merging constraints based on road factor behavior because there is lower predictability about when a road factor will overtake at the start of overtaking maneuvers. Merging constraints can be the predicted position where the autonomous vehicle yields to a road factor. This can allow the autonomous vehicle to slow down less initially and travel more. Furthermore, based on a higher predictability score, there may be fewer or more custom constraints. Based on a lower predictability score, there may be more or more generalized constraints. In some implementations, one or more constraints may also be updated based on the autonomous vehicle's predictability score. The vehicle's predictability score may be based on confidence levels regarding the autonomous vehicle's responses to other road objects or road factors, the autonomous vehicle's detection system, or other factors used to manipulate the autonomous vehicle. Constraints may also change when road factor trajectories and behaviors indicate that an overtaking maneuver is imminent.

[0060] One or more constraints can then be used to determine one or more components of vehicle maneuvering, such as speed, path, or route. In cases of low predictability scores for road factors or autonomous vehicles, the vehicle's computing equipment can determine a slower speed or more cautious maneuvering based on one or more constraints. In cases of high predictability scores for road factors or autonomous vehicles, the vehicle's computing equipment can determine possible routes taking road factors into account and effectively navigate specific paths through those routes. Particularly in cases of high predictability of impending overtaking maneuvers due to road factors, the vehicle's computing equipment can determine a slower speed as needed to prepare to yield to road factors. In some implementations, the determination of one or more components involves selecting component settings that better satisfy one or more constraints.

[0061] In some cases, one or more of the obtained components can also be used in other calculations in the next iteration of the analysis, such as overlap calculations or lateral clearance calculations. For example, road factors and / or vehicle lateral clearance preferences can be used to update the projected trajectory of road factors, and then update vehicle handling in response to the updated projected trajectory.

[0062] In terms of the system, a first module can be configured for estimating the spacing profile, and a separate second module can be configured for determining vehicle handling. This separation of functions allows components of one module to be designed and transported separately from the other. Thus, the second module can be designed to determine the combined constraints with or without the spacing profile, and, in the absence of a spacing profile received from the first module, to determine more conservative constraints.

[0063] In block 410, the vehicle's computing device 110 can perform defined vehicle maneuvers by correspondingly controlling one or more operating systems of the autonomous vehicle. The vehicle's computing device 110 can send instructions to one or more operating systems of the vehicle 100, including a deceleration system 160, an acceleration system 162, and a steering system 164. In some embodiments, a third module, separate from the first and second modules, can be configured to receive control instructions and / or execute control instructions from the second module.

[0064] In some alternative implementations, determining vehicle handling may include determining the degree of conservatism the autonomous vehicle should exhibit when sharing lateral space with road factors, based on estimated spacing profiles. Greater conservatism may be associated with lower predictability, higher determinism of more aggressive road factor behaviors (such as overtaking maneuvers), or higher complexity of world-context-based vehicle handling. This conservatism may be a separate determination (such as a level or score of conservatism) or may be an integral part of the determination of autonomous vehicle handling.

[0065] In another embodiment, the same or similar methods can be applied to other types of road factors besides those described above. In some cases, spacing profiles for more than one road factor can be determined in parallel and used to determine vehicle handling.

[0066] The technology described in this paper allows passengers to make smoother and safer journeys in autonomous vehicles. Specifically, more forward travel is possible in autonomous vehicles compared to situations without the aforementioned navigation system. This technology also minimizes unnecessary overreaction and underreaction by autonomous vehicles to small lateral clearances, maintaining a level of safety and compliance requirements while ensuring comfort in terms of road factors and other road-related factors.

[0067] Unless otherwise stated, the foregoing alternative examples are not mutually exclusive, but can be implemented in various combinations to achieve unique advantages. Because these and other variations and combinations of the above features can be utilized without departing from the subject matter defined by the claims, the foregoing description of the embodiments should be understood in an illustrative rather than restrictive manner. Furthermore, the provision of examples described herein, as well as phrases such as “e.g.,” “comprising,” etc., should not be construed as limiting the subject matter of the claims to the specific examples; rather, the examples are intended to illustrate only one of many possible embodiments. Additionally, the same reference numerals in different figures may identify the same or similar elements.

Claims

1. A system for estimating a spacing profile of a road factor, the system comprising: a perception system; one or more processors configured to: receive sensor data from the perception system, the sensor data identifying one or more behaviors of a road factor in an environment of an autonomous vehicle; determine an autonomous vehicle preference for a lateral gap between the autonomous vehicle and the road factor; estimate a spacing profile of the road factor based on a change in the lateral gap over time, the estimated spacing profile including a road factor preference for the lateral gap, wherein the autonomous vehicle preference for the lateral gap decreases or increases based on the one or more behaviors of the road factor; and based on the estimated spacing profile, send control instructions to control one or more operating systems of the autonomous vehicle; wherein the estimated spacing profile indicates a predictability of the road factor based on an observed behavior of the road factor and a previously predicted behavior of the road factor.

2. The system of claim 1, wherein, the one or more processors are further configured to: determine one or more predicted behaviors of the road factor, wherein the determination of the one or more predicted behaviors of the road factor is based on an initiation when the road factor is predicted to interact with an autonomous vehicle.

3. The system of claim 1, wherein, the one or more processors are further configured to: determine one or more predicted behaviors of the road factor, wherein the determination of the one or more predicted behaviors of the road factor is based on an initiation of an autonomous vehicle requirement for a lateral gap with the road factor.

4. The system of claim 1, wherein, wherein, the one or more processors are further configured to: determine one or more predicted behaviors of the road factor, predict, based on the one or more predicted behaviors of the road factor, that the road factor will perform a pass maneuver; and determine a maneuver of the autonomous vehicle, wherein the determined maneuver includes decreasing a current speed of the autonomous vehicle in preparation for yielding to the road factor when the road factor performs the pass maneuver.

5. The system of claim 1, wherein, the one or more processors are further configured to: determine one or more predicted behaviors of the road factor, wherein the determination of the one or more predicted behaviors of the road factor is based on an initiation of an existing lateral gap between an autonomous vehicle and the road factor.

6. The system of claim 1, wherein, the one or more processors are further configured to: determine one or more predicted behaviors of the road factor, wherein the determination of the one or more predicted behaviors of the road factor is based on an initiation of a machine learning model.

7. The system of claim 1, wherein, the one or more processors are further configured to: determine one or more predicted behaviors of the road factor, wherein the one or more predicted behaviors of the road factor include a pass maneuver related to the autonomous vehicle.

8. The system of claim 1, wherein, the predictability of the road factor includes a predictability score of the road factor.

9. The system of claim 8, wherein, the predictability score of the road factor is further based on a degree of matching of the observed behavior of the road factor to the previously predicted behavior of the road factor.

10. The system of claim 8, wherein, the estimated spacing profile further includes a predictability score of the autonomous vehicle.

11. The system of claim 1, wherein, the one or more operating systems include at least one of a deceleration system or an acceleration system.

12. The system of claim 11, wherein, the one or more operating systems further include a steering system.

13. A method for estimating a spacing profile of a road factor, the method comprising: receiving, by one or more computing devices, sensor data from a perception system, the sensor data identifying one or more behaviors of a road factor in an environment of an autonomous vehicle; determining, by one or more computing devices, an autonomous vehicle preference for a lateral gap between the autonomous vehicle and the road factor; estimating, by one or more computing devices, a spacing profile of the road factor based on a change in the lateral gap over time, the estimated spacing profile including a road factor preference for the lateral gap, wherein the autonomous vehicle preference for the lateral gap decreases or increases based on the one or more behaviors of the road factor; sending, by one or more computing devices, control instructions to control one or more operating systems of the autonomous vehicle based on the estimated spacing profile; wherein the estimated spacing profile indicates a predictability of the road factor based on observed behaviors of the road factor and previously predicted behaviors of the road factor.

14. The method of claim 13, further comprising: determining one or more predicted behaviors of the road factor, wherein the determination of the one or more predicted behaviors of the road factor is based on initiation when the road factor is predicted to interact with the autonomous vehicle.

15. The method of claim 13, further comprising: determining one or more predicted behaviors of the road factor, wherein the determination of the one or more predicted behaviors of the road factor is based on initiation of a lateral gap requirement of the autonomous vehicle with respect to the road factor.

16. The method of claim 13, further comprising: determining one or more predicted behaviors of the road factor, predicting, based on the one or more predicted behaviors of the road factor, that the road factor will perform a pass maneuver; and determining a maneuver of the autonomous vehicle, wherein the determined maneuver includes reducing a current speed of the autonomous vehicle in preparation for yielding to the road factor when the road factor performs the pass maneuver.

17. The method of claim 13, further comprising: determining one or more predicted behaviors of the road factor, wherein the determination of the one or more predicted behaviors of the road factor is based on initiation of an existing lateral gap between the autonomous vehicle and the road factor.

18. The method of claim 13, further comprising: determining one or more predicted behaviors of the road factor, wherein the determination of the one or more predicted behaviors of the road factor is based on initiation of a machine learning model.

19. The method of claim 13, further comprising: determining one or more predicted behaviors of the road factor, wherein the one or more predicted behaviors of the road factor include a pass maneuver with respect to the autonomous vehicle.

20. The method of claim 13, wherein, the estimated spacing profile includes a predictability score of the autonomous vehicle.

21. The method of claim 13, wherein, the predictability of the road factor includes a predictability score of the road factor, and wherein the predictability score of the road factor is based on a degree of matching of observed behaviors of the road factor and previously predicted behaviors of the road factor.

22. A non-transitory computer-readable medium storing instructions that, when executed by one or more processors, cause the one or more processors to perform operations comprising: receiving sensor data from a perception system, the sensor data identifying one or more behaviors of a road factor in an environment of an autonomous vehicle; determining an autonomous vehicle preference for a lateral gap between the autonomous vehicle and the road factor; estimating a spacing profile of the road factor based on a change in the lateral gap over time, the estimated spacing profile including a road factor preference for the lateral gap, wherein the autonomous vehicle preference for the lateral gap decreases or increases based on the one or more behaviors of the road factor; and based on the estimated spacing profile, sending control instructions to control one or more operating systems of the autonomous vehicle; wherein the estimated spacing profile indicates a predictability of the road factor based on observed behaviors of the road factor and previously predicted behaviors of the road factor.

Images