HYBRID DATA-DRIVEN TRAFFIC PARTICIPANT PREDICTION USING ACHIEVABLE VOLUMES AND SUPERVISED LEARNING

Abstract

A method for predicting a road user occupancy corridor includes receiving input data, predicting road user occupancy levels using reachability analysis and the input data, determining road user occupancy corridor constraints using a machine learning model and the input data, and determining a road user occupancy corridor based on these constraints using the machine learning model and the road user occupancy levels using reachability analysis. The method also includes controlling the movement of a host vehicle based on the road user occupancy corridor.

Description

INTRODUCTION

[0001] The present disclosure relates to methods and systems for predicting road users and in particular hybrid data-driven systems and methods for predicting road users using achievable sets and supervised learning.

[0002] This introduction presents the general context of the disclosure. The work of the inventors mentioned herein, insofar as it is described in this introduction, as well as aspects of the description that may not otherwise be considered prior art at the time of filing, are neither expressly nor implicitly admitted as prior art against this disclosure.

[0003] Autonomous vehicles use models to predict the position of other road users or actors. State-of-the-art models are data-driven predictive models. These data-driven predictive models are not necessarily robust and output a finite number of hypotheses that do not adequately consider possible future positions of the road user. Other approaches include reachability-based road user predictions, which tend to be too conservative. In particular, reachability-based road user prediction models are primarily useful in the short term and, due to their cautious approach, are not suitable for complex, chaotic, and congested traffic environments. Therefore, it is desirable to develop a method and system for predicting road user positions or occupancy on a map that are conservative in the short term. SUMMARY OF THE INVENTION

[0004] The present disclosure describes a method for predicting a road user occupancy corridor. The method includes receiving input data, such as map data and initial trajectory data. The map data comprises a map of the road user's surroundings. The initial trajectory data comprises the road user's initial trajectory. The method further includes predicting a multitude of occupancy sets for the road user, utilizing, among other things, exact or approximate reachability analyses, such as Hamilton-Jacobi reachability, sample-based optimal control methods, polytopic or polyhedral set-based methods, zonotopic reachability, ellipsoidal reachability, and the input data.Furthermore, the method includes determining a road user's occupancy corridor using a machine learning model and input data, determining a road user's occupancy corridor using the road user's occupancy corridor constraints using the machine learning model and the road user's occupancy quantities based on reachability theory, and controlling the movement of a host vehicle based on the road user's occupancy corridor. The host vehicle is an autonomous vehicle. The method described in this section improves autonomous vehicle technology by accurately and realistically predicting a road user's position, thus enabling the host vehicle to maneuver appropriately to avoid contact with the road user.

[0005] According to certain aspects of the present disclosure, the method can include receiving perception data from a multitude of sensors on the host vehicle. The perception data contains information about the road user's surroundings. Furthermore, the method includes merging the perception data with map data to generate scene data. The scene data includes a nominal route and a road centerline. The road user is traveling on the road.The initial trajectory data includes the longitudinal and lateral acceleration of the road user. The reachability analysis utilizes the longitudinal and lateral acceleration of the road user, along with an assumed model, including a kinematic, dynamic, or machine learning model. The road user's occupancy sets are a collection of forward reachable position sets over a predetermined period. The procedure further includes determining that the road user's forward reachable set does not cross paths with other road users over a predetermined period and that the forward reachable set lies within permissible road positions. The occupancy corridor is determined based on the road user's downforce accelerations.Occupancy corridor restrictions for the road user include the road user's maximum and minimum acceleration. The procedure may involve determining whether a host vehicle's trajectory crosses the road user's occupancy corridor. The procedure may involve removing the host vehicle's trajectory that crosses the road user's occupancy corridor. The procedure may involve determining whether a host vehicle's trajectory does not cross the road user's occupancy corridor. The procedure may involve instructing the host vehicle to move along a trajectory that does not cross the road user's occupancy corridor. Implementations of the described techniques may include hardware, a method or process, or computer software on a computer-available medium.

[0006] The present disclosure also describes a system for predicting a traffic occupancy corridor. The system comprises sensors and a control unit that communicates with the sensors. The control unit includes a processor and a non-transient, computer-readable medium and is programmed to execute the method described above.

[0007] The present disclosure further describes a tangible, non-transitory, machine-readable medium comprising machine-readable instructions which, when executed by a processor, cause the processor to perform the procedure described above.

[0008] Further applications of this disclosure will become apparent from the detailed description below. It is understood that the detailed description and specific examples serve only for illustration and are not intended to limit the scope of the disclosure.

[0009] The above features and advantages, as well as other features and advantages of the system and method disclosed herein, will readily become apparent from the detailed description including the claims and from the exemplary embodiments in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The present disclosure will be better understood from the detailed description and the accompanying drawings, in which: Fig. 1 a block diagram representing a host vehicle with a hybrid data-driven system for predicting road users using achievable sets and supervised learning. Fig. 2 a schematic representation of the host vehicle and a road user moving on a road. Fig. 3 a flowchart of a hybrid data-driven approach to predicting road users using achievable sets and supervised learning. Fig. 4 a flowchart of a procedure for predicting road users using calculations of the achievable quantity. Fig. 5. A flowchart of a procedure for training a reasoner. DETAILED DESCRIPTION

[0011] Several examples from the revelation, illustrated in the accompanying drawings, will now be discussed in detail. Wherever possible, the same or similar reference symbols will be used in the drawings and the description to indicate the same or similar parts or steps.

[0012] With reference to Fig. A host vehicle 10 generally comprises a chassis 12, a vehicle body 14, and front and rear wheels 17 and can be referred to as a vehicle system. In the illustrated embodiment, the host vehicle 10 comprises two front wheels 17a and two rear wheels 17b. The body 14 is arranged on the chassis 12 and essentially encloses components of the host vehicle 10. The body 14 and the chassis 12 can together form a frame. The wheels 17 are each rotatably connected to the chassis 12 near a corresponding corner of the body 14. The host vehicle 10 has a front axle 19 connected to the front wheels 17a and a rear axle 25 connected to the rear wheels 17b.

[0013] In various embodiments, the host vehicle 10 can be an autonomous vehicle, and a system 98 is integrated into the host vehicle 10. The system 98 can be described as a hybrid data-driven traffic predictor system that utilizes achievable quantities and supervised learning. The host vehicle 10 is, for example, a vehicle that is automatically controlled to transport passengers (e.g., vehicle occupant 15) from one location to another. The host vehicle 10 can be configured as a truck, sedan, coupe, SUV, motorhome, etc. In one embodiment, the host vehicle 10 can be a so-called Level Two, Level Three, Level Four, or Level Five automation system.A Level Four system signifies a "high degree of automation" and refers to the driving-mode-specific execution of aspects of the dynamic driving task by an automated driving system, even if a human driver does not respond appropriately to a request for intervention. A Level Five system signifies "full automation," meaning that an automated driving system fully executes aspects of the dynamic driving task under a range of road and environmental conditions that can be handled by a human driver. In Level 3 vehicles, the vehicle system performs the entire dynamic driving task (DDT) within its designed scope.The driver is only expected to engage the DDT function if the host vehicle 10 actually "requests" the driver to take over, specifically if something goes wrong or the vehicle is about to leave its operating area. For Level 2 vehicles, the systems assist with steering, braking / acceleration, lane keeping, and adaptive cruise control.

[0014] As shown, the host vehicle 10 generally comprises a drive system 20, a transmission system 22, a steering system 24, a braking system 26, a sensor system 28, an actuator system 30, at least one data storage device 32, at least one control unit 34, and a communication system 36. In various embodiments, the drive system 20 may include an electric machine, such as a traction motor, and / or a fuel cell drive system. The host vehicle 10 may also include a battery (or battery pack) 21, which is electrically connected to the drive system 20. Accordingly, the battery 21 is configured to store electrical energy and supply the drive system 20 with electrical energy. In certain embodiments, the drive system 20 may include an internal combustion engine.The transmission system 22 is configured to transmit the power of the drive system 20 to the vehicle wheels 17 according to selectable gear ratios. According to various embodiments, the transmission system 22 can be a stepped automatic transmission, a continuously variable transmission, or another suitable transmission. The braking system 26 is configured to provide a braking torque for the vehicle wheels 17. In various embodiments, the braking system 26 can include friction brakes, an electric motor, a regenerative braking system such as an electric machine, and / or other suitable braking systems. The steering system 24 influences the position of the vehicle wheels 17 and can include a steering wheel 33. Although a steering wheel 33 is shown for illustrative purposes, in some embodiments conceivable within the scope of this disclosure, the steering system 24 may not include a steering wheel 33.

[0015] The sensor system 28 comprises one or more sensors 40 (i.e., sensor units) that detect perceptible conditions of the external environment and / or the internal environment of the host vehicle 10. The sensors 40 communicate with the control unit 34 and may, in particular, include one or more radars, one or more lidar sensors, one or more proximity sensors, one or more odometers, one or more GPR sensors, one or more steering angle sensors, GNSS transceivers (e.g., one or more GPS transceivers), one or more tire pressure sensors, one or more cameras, one or more gyroscopes, one or more accelerometers, one or more inclinometers, one or more speed sensors, one or more three-dimensional (3D) depth sensors, one or more ultrasonic sensors, one or more inertial navigation systems (IMUs), thermal imaging sensors, one or more microphones 31, and / or other sensors.Each sensor 40 is designed to generate a signal indicating the detected perceptible conditions (i.e., sensor data) of the external environment and / or the internal environment of the host vehicle 10. Since the sensor system 28 delivers sensor data to the control unit 34, the sensor system 28 and its sensors 40 are considered information sources (or simply sources).

[0016] The actuator system 30 comprises one or more actuator devices 42 that control one or more vehicle functions, in particular the drive system 20, the transmission system 22, the steering system 24 and the braking system 26. In various embodiments, the vehicle functions can also include interior and / or exterior functions of the vehicle, in particular doors, a trunk and cabin functions, such as ventilation, music, lighting, etc.

[0017] The data storage device 32 stores data used for the automatic control of the host vehicle 10. In various embodiments, the data storage device 32 stores defined maps of the drivable environment. In various embodiments, the defined maps can be predefined and retrieved from a remote system. For example, the defined maps can be compiled by the remote system, transmitted (wirelessly and / or via cable) to the host vehicle 10, and stored in the data storage device 32. The data storage device 32 can be part of the control unit 34 or be redirected from it, or it can be part of the control unit 34 and part of a redirection system.

[0018] The control unit 34 comprises at least one processor 44 and a non-transient, computer-readable storage device or medium 46. The processor 44 may be a custom-made or commercially available processor, a central processing unit (CPU), a graphics processing unit (GPU), an auxiliary processor among several processors connected to the control unit 34, a semiconductor-based microprocessor (in the form of a microchip or chipset), a macroprocessor, a combination thereof, or generally, a device for executing instructions. The computer-readable storage device or medium 46 may include volatile and non-volatile storage, such as read-only memory (ROM), random-access memory (RAM), and keep-alive memory (KAM). A KAM is persistent or non-volatile memory that can be used to store various operating variables while the processor 44 is powered off.The computer-readable storage device or medium 46 can be implemented using a whole range of storage devices, such as PROMs (programmable read-only memory), EPROMs (erasable PROMs), EEPROMs (electrically erasable PROMs), flash memory, or other electrical, magnetic, optical, or combined storage devices capable of storing data, some of which represent executable instructions used by the control unit 34 to control the host vehicle 10. The control unit 34 of the host vehicle 10 can be referred to as the vehicle control unit and programmed to execute a method 100 (. Fig. 3) and / or a procedure 200 ( Fig. 4), as described in detail below.

[0019] The instructions can comprise one or more separate redirection programs, each containing an ordered list of executable instructions for implementing logical functions. When executed by the processor 44, the instructions receive and process signals from the sensor system 28, perform logic, calculations, procedures, and / or algorithms for the automatic control of the host vehicle 10's components, and generate control signals for the actuator system 30 to automatically control the host vehicle 10's components based on the logic, calculations, procedures, and / or algorithms. Although in Fig. While only a single control unit 34 is shown in Figure 1, embodiments of the host vehicle 10 can comprise a plurality of control units 34 that communicate and cooperate via a suitable communication medium or a combination of communication media to process the sensor signals, perform logic, calculations, procedures and / or algorithms, and generate control signals to automatically control functions of the host vehicle 10. In various embodiments, one or more instructions from the control unit 34 are included in the system 98.

[0020] The host vehicle 10 includes a user interface 23, which is a touchscreen 43 ( Fig. 3) can act in the dashboard. The user interface 23 can, in particular, include an alarm, such as one or more loudspeakers 27 for emitting an audible tone, haptic feedback in a vehicle seat or other object, one or more displays 29, one or more microphones 31 (e.g., a microphone array), and / or other devices suitable for conveying a notification to the vehicle user of the host vehicle 10. The user interface 23 is electronically connected to the control unit 34 and is configured to receive input from a user (e.g., a driver or a passenger). The user interface 23 can, for example, include a touchscreen and / or buttons configured to receive input from a person (e.g., a vehicle occupant 15). Accordingly, the control unit 34 is configured to receive user input via the user interface 23.Although the microphone is 31 in . Fig. While microphone 1 is shown as part of the user interface 23, other microphones 31 can also be part of the sensor system 28. The microphones 31 are designed to capture voice commands from vehicle occupants 15 in the passenger compartment 39 of the host vehicle 10. However, it is also conceivable to connect the microphones 31 to other parts of the host vehicle 10. Regardless of its exact position, the microphone 31 is connected to the control unit 34.

[0021] The host vehicle 10 can include one or more displays 29 configured to show one or more images as described below. For example, the display 29 can be configured to show information to one or more occupants 15 in the passenger compartment 39 of the host vehicle 10. The display 29 is connected to the control unit 34. Accordingly, the control unit 34 is configured to control the operation of the display 29.

[0022] The communication system 36 is connected to the control unit 34 and is configured to wirelessly transmit information to and from other units, such as the remote units 48, for example, using V2X (Vehicle-to-Everything) technology. In particular, the communication system 36 can transmit and / or receive information from other vehicles (V2V communication), the infrastructure (V2I communication), remote systems in a remote call center (e.g., GENERAL MOTORS' ON-STAR), and / or personal electronic devices, such as a mobile phone. Some of the remote units 48 may also be configured to wirelessly communicate information to and from other units, such as the remote vehicles 48, for example, using V2X technology.In this disclosure, remote vehicles 48 capable of wirelessly transmitting information to and from other units, such as the remote vehicles 48, are referred to as V2X-enabled remote vehicles. In certain embodiments, the communication system 36 is a wireless communication system configured to communicate via a wireless local area network (WLAN) using IEEE 802.11 standards or radio data communication. However, additional or alternative communication methods, such as Dedicated Short Range Communication (DSRC), are also considered in this disclosure. DSRC channels refer to one-way or two-way short- to medium-range wireless communication channels specifically designed for use in vehicles, as well as to corresponding protocols and standards.Accordingly, the communication system 36 can include one or more antennas and / or communication transceivers 37 for receiving and / or transmitting signals, such as Cooperative Sensing Messages (CSMs). The communication system 36 is configured to wirelessly transmit information between the host vehicle 10 and another vehicle (i.e., a remote unit 48). Furthermore, the communication system 36 is configured to wirelessly transmit information between the host vehicle 10 and the infrastructure.

[0023] With reference to Fig. 2 The host vehicle 10 can drive on a road 54 in various driving scenarios while other road users 56 (e.g., other vehicles, pedestrians, cyclists, etc.) are also moving on the same road 54. In such driving scenarios, it is desirable to predict the occupancy corridor 58 of the road user 56 in the short term (e.g., within zero to five seconds) so that the host vehicle 10 can avoid contact with the road user 56. In particular, the host vehicle 10 can consider the center line 60 of the road 54 and the predicted occupancy corridor 58 of the road user 56 to select a route 62. The hybrid, data-driven system for road user prediction 98 disclosed here includes both analyses of the achievable quantity and recorded real or simulation data.In this way, the system predicts quantities that closely constrain the recorded trajectories of road users 56 who exhibit similar characteristics in varying driving scenarios and conditions. In other words, road users 56 with similar characteristics have similar predicted occupancy corridors 58. In the system, the occupancy corridor prediction 58 depends on various complex characteristics, including, but not limited to, the road user's aggressiveness, the uncertainty of the scene, the road geometry, the weather conditions, the vehicle class, and so on. The system calculates controlled occupancy corridor predictions faster than other systems by using a machine learning model, thus saving computing power and energy.System 98 is also less conservative than other approaches because its occupancy corridor predictions depend on the aggressiveness of road users, the uncertainty of the scene, the vehicle class / type, the map geometry, and other complex features, and are only used in the short term. System 98 is more conservative than purely data-driven approaches because it outputs quantities, boundary lines, and / or regions instead of individual trajectories.

[0024] Fig. Figure 3 is a flowchart of a hybrid data-driven procedure 100 for predicting road users using achievable sets and supervised learning. The procedure 100 begins in blocks 102 and 104. In block 102, the controller 34 receives map data. The map data includes, in particular, a map of the road user's surroundings 56, the road centerlines 60, and the nominal route of the road user 56. In block 104, the controller 34 receives initial trajectory data of the road user 56. The initial trajectory data includes an initial trajectory of the road user 56. The map data and the initial trajectory data can be collectively referred to as input data. Accordingly, the controller 34 first receives the input data. The procedure 100 then proceeds to block 106.

[0025] In block 106, scene fusion takes place. Specifically, in block 106, the control unit 34 receives perception data from the sensors 40 (e.g., cameras, lidar, etc.) of the host vehicle 10 and / or the road user 56. The perception data contains information about the road user 56's environment. This perception data is then merged with the map data to generate scene data. The scene data includes a nominal route of the road user 56, a centerline 60 of road 54, the road geometry of road 54, etc. As described above, the host vehicle 10 and / or the road user 56 can move along road 54. Subsequently, the process 100 proceeds to block 108. In block 108, the control unit 34 uses reachability analysis and input data to predict a variety of occupancy levels for the road user 56 over a predetermined period in the future.In the present disclosure, the term "occupancy" means a position of the road user 56 on the road. The term "occupancy quantity" therefore refers to a multitude of likely locations of the road user 56 on the road 54 at a given time.

[0026] Fig. Figure 4 is a flowchart of a procedure 200 for predicting the occupancy quantities of road user 56 over a given period in the future using reachability analysis and input data. The input data includes the state of road user 56 (i.e., the road user state) and the associated uncertainty. Considering the road user state and the associated uncertainty, the controller 34 defines the initial occupancy quantity of road user 56 at an initial time as the area that road user 56 occupies on road 54 at the initial time (i.e., in the first time step). Once the initial occupancy quantity is determined, the procedure 200 proceeds to block 202. In block 202, the controller 34 uses reachability theory to determine (i.e., calculate) the next forward reachable quantity (i.e., the forward occupancy quantity of road user 56).The procedure then proceeds to block 204. In block 204, the control unit 34 determines whether the occupancy quantity (i.e., the quantity reachable forward) of road user 56 determined in block 202 crosses other road users within the specified period and / or is within a permissible road position (e.g., within the edge lines 64 of road 54). Fig. 2)) If the occupancy quantity of road user 56 does not intersect any other road users and / or lies within the permissible road positions, then the control unit 34 marks this occupancy quantity as valid and proceeds to block 206. In block 206, the control unit 34 marks this new occupancy quantity as the current occupancy quantity in order to determine the next occupancy quantity of road user 56. Subsequently, the process 100 returns to block 202. In process 200, the control unit 34 thus performs successive predictions of the forward occupancy quantity over the specified period in the future. These quantities are the occupancy quantities that represent the position that road user 56 could reach within the specified time.

[0027] Referring back to Fig. 3. The occupancy volume predictions determined in block 108 are forwarded to a reasoner in block 110. The reasoner also receives the initial trajectory data of road user 56 determined in block 104 and the scene data generated in block 106. In addition to the scene data, the reasoner can receive (as input data) the trajectory data of road user 56, the aggressiveness of the road user, the road user type (e.g., sedan, SUV, truck, bicycle, pedestrian, etc.), the road conditions, the weather conditions, nominal predictions and the confidence (i.e., the uncertainty) of these nominal predictions, map data, and the uncertainty of the scene (e.g., perception and / or location uncertainty). The reasoner selects the hyperparameters that determine the occupancy volume predictions based on the scene data. These hyperparameters include, in particular, the length of the prediction horizon (e.g.,The time horizon for occupancy predictions, the time step size, the time to brake, the time to collision, and / or the time to the nearest road user are all parameters that can be used. The length of the prediction horizon for occupancy predictions (i.e., how far into the future) can be variable. For example, the length of the prediction horizon for occupancy prediction can depend on the estimated time to collision with the nearest road user and the nominal uncertainty of the prediction hypothesis. As output, the Reasoner uses a machine learning model and the input data to determine the road user occupancy corridor constraints. In other words, the Reasoner outputs the hyperparameters that constrain the occupancy corridor prediction.The occupancy corridor restrictions of road user 56 include, in particular, the maximum acceleration of road user 56, the minimum acceleration of road user 56, the length of the prediction horizon for the occupancy quantity prediction, and / or the (e.g., coarse or fine) time step. Supervised learning is used to train a function approximator, e.g., a neural network. Therefore, the reasoner can be a machine learning model, e.g., a neural network.

[0028] Fig. Figure 5 is a flowchart of Procedure 300 for training the reasoner described above. Procedure 300 is performed offline and uses supervised learning with simulated and / or recorded driving data to learn the mapping of inputs to outputs. Procedure 300 begins in Block 302. In Block 302, the control unit 34 receives the input data. As described above, the input data can include the trajectory data of the road user 56, the aggressiveness of the road user, the road user type (e.g., sedan, SUV, truck, bicycle, pedestrian, etc.), the road conditions, the weather conditions, nominal predictions and the confidence (i.e., the uncertainty) of these nominal predictions, map data, scene data, and the uncertainty of the scene (e.g., perception and / or location uncertainty). Procedure 300 then proceeds to Block 304.In block 304, the reasoner receives the input data (i.e., simulated and / or recorded driving data) and outputs the accelerations of road user 56. The process then transitions to block 306. In block 306, an occupancy predictor receives the accelerations of road user 56 and outputs occupancy corridors 58 for road user 56. The process then transitions to blocks 308 and 310. In block 308, the actual ground trajectories of road user 56 are received. These actual ground trajectories can be based on simulated and / or recorded driving data. The actual ground trajectories of road user 56 are fed into a loss function. The loss function is a mathematical function that measures how well a machine learning model performs in a given dataset.The machine learning algorithm (the Reasoner) minimizes the value of the loss function by adjusting the model parameters during the training process. The loss function is used to update the Reasoner's machine learning parameters (e.g., the neural network parameters). Therefore, the updated machine learning parameters are fed back to the Reasoner in block 304. The training process (300) is repeated until the value of the loss function reaches a predefined minimum.

[0029] Referring back to Fig.In step 3, the process 300 transitions from block 110 to block 112. As described above, the reasoner outputs the hyperparameters that determine the occupancy predictions based on the scene data. In block 112, the occupancy predictor receives the hyperparameters that govern the occupancy predictions and determines (e.g., calculates) the occupancy levels for the relevant road users 56 for the future. The process then transitions to block 114.

[0030] In block 114, the control unit 34 receives the occupancy levels and uncertainty corridors determined in block 108, as well as the occupancy corridors determined in block 112. Using this information, the predictor resolver (in block 114) combines the nominal (i.e., basic) machine learning predictions with the occupancy levels. The predictor resolver performs two tasks. First, it ensures that the data-driven occupancy corridor determination (i.e., the determination by machine learning) has the same data structure as the occupancy level predictions determined in block 108. Second, the predictor resolver also corrects the nominal predictions that are unreachable using projection and convex optimization. For example, a nominal prediction that lies outside the occupancy corridor before the specified time period is unreachable or inconsistent with the training data.The correction mechanism finds the trajectory closest to the occupancy prediction that lies within the occupancy ranges in the first time steps and minimizes the errors for the remaining times. For this purpose, the predictor resolver can use an optimization cost function according to the following equations: minψ J=12(X−Ψ)T(X−Ψ) ψ(i)∈Oi, i=1,…, N X=[x(0), x(1), …, x(N)]T Ψ=[ψ(0),ψ(1), …,ψ(N)]T where: J is the optimization cost function, X are the waypoints of the corrected trajectory prediction, Ψ are the waypoints of the unfeasible trajectory prediction, i is the time index, N is the number of time steps, and O i The occupancy set for time index i.

[0031] In block 114, the predictor resolver outputs an occupancy corridor 58 for road user 56. The process 100 then proceeds to block 116. In block 116, the occupancy corridor 58 of road user 56 is fed to the behavior planning algorithm of the host vehicle 10. The control unit 34 of the host vehicle 10 controls the movement of the host vehicle 10 based on the occupancy corridor 58 of road user 56. For example, the control unit 34 can determine possible trajectories for the host vehicle 10 and compare them with the occupancy corridor 58 of road user 56. If a possible trajectory of the host vehicle 10 intersects the occupancy corridor 58 of road user 56, the control unit 34 removes this possible trajectory.If a possible trajectory of the host vehicle 10 overlaps with the long-term data-driven occupancy prediction of the road user 56, the control unit 34 allows this possible trajectory in the short term. If a possible trajectory of the host vehicle 10 does not overlap with the occupancy corridor 58 or the long-term data-driven occupancy prediction of the road user 56, the possible trajectory is considered acceptable, and the control unit 34 instructs the host vehicle 10 to move autonomously along this possible trajectory.

[0032] Although exemplary embodiments are described above, these embodiments are not intended to describe all possible forms encompassed by the claims. The terms used in the patent specification are to be understood as descriptive rather than limiting, and it is understood that various modifications may be made without departing from the spirit or scope of the present disclosure. As already described, the features of different embodiments can be combined to form further embodiments of the system and method disclosed herein, which may not be expressly described or illustrated. Although different embodiments may be possible with regard to one or more desired properties,When a design is described as advantageous or preferred over other embodiments or implementations of the prior art, the person skilled in the art knows that compromises may be made with regard to one or more features or properties in order to achieve the desired overall system properties, which depend on the specific application and implementation. These properties may include, in particular, cost, robustness, durability, life-cycle costs, marketability, appearance, packaging, size, usability, weight, manufacturability, ease of assembly, etc. As such, embodiments that are described as less desirable than other embodiments or implementations of the prior art with respect to one or more properties are not outside the scope of the disclosure and may be desirable for certain applications.

[0033] The drawings are simplified and not to scale. For the sake of simplicity and clarity only, directional terms such as above, below, left, right, upwards, over, over, below, under, back, and front may be used in relation to the drawings. These and similar directional terms are not to be interpreted as limiting the scope of the disclosure in any way.

[0034] Here, embodiments of the present disclosure are described. It is understood, however, that the disclosed embodiments are merely examples and that other embodiments may take different and alternative forms. The figures are not necessarily to scale. Some features may be larger or smaller to show details of certain components. Therefore, the specific constructive and functional details disclosed here are not to be understood as limiting, but merely as a representative basis for teaching a person skilled in the art to apply the system and method disclosed herein in different ways. As is known to the person skilled in the art, various features illustrated and described with reference to one of the figures can be combined with features illustrated in one or more other figures to form embodiments that are not expressly illustrated or described.The combinations of features shown provide representative embodiments for typical applications. However, for certain applications or embodiments, different combinations and modifications of the features that comply with the teachings of this disclosure may be desirable.

[0035] Embodiments of the present disclosure may be described herein in the form of functional and / or logical block components and various processing steps. It should be noted that such block components may be implemented by a number of hardware, software, and / or firmware components configured to perform the specified functions. For example, in one embodiment of the present disclosure, various integrated circuit components may be used, such as memory elements, digital signal processing elements, logic elements, lookup tables, or the like, which can perform a variety of functions under the control of one or more microprocessors or other control devices.Furthermore, the person skilled in the art knows that embodiments of the present disclosure can be used in conjunction with a number of systems and that the systems described here are merely exemplary embodiments of the present disclosure.

[0036] For the sake of brevity, techniques of signal processing, data fusion, signaling, control, and other functional aspects of the systems (and the individual operating components of the systems) are not described in detail here. Furthermore, the connecting lines shown in the various figures are intended to represent exemplary functional relationships and / or physical connections between the various elements. It should be noted that in an embodiment of the present disclosure, alternative or additional functional relationships or physical connections may exist.

[0037] This description serves only for illustration and is in no way intended to limit the revelation, its application, or its benefits. The comprehensive teachings of revelation can be implemented in a multitude of forms. Although this revelation includes certain examples, its true scope should not be limited to these, as other variations will become apparent upon reviewing the drawings, the description, and the claims that follow.

Claims

[1] A method for predicting a traffic congestion corridor, comprising: Receiving input data, wherein the input data comprises map data and initial trajectory data, the map data comprises a map of the surroundings of a road user, and the initial trajectory data comprises an initial trajectory of the road user; Predicting a variety of road user occupancy levels over a given future period using the accessibility analysis and the input data; Determining the road user’s occupancy corridor restrictions based on a machine learning model and the input data; Determining a road user's occupancy corridor based on the road user's occupancy corridor constraints using the machine learning model and the road user's occupancy quantities using an accessibility analysis; and Controlling a host vehicle's movement based on the road user's occupancy corridor. [2] The method of claim 1, further comprising: Receiving perception data from a plurality of sensors of the host vehicle, wherein the perception data comprises information about the environment of the road user; and Fusing the perception data with the map data to generate scene data, wherein the scene data comprises a nominal route and a road centerline of a road and the road user is moving on the road. [3] The method of claim 2, wherein the initial trajectory data comprises a longitudinal acceleration and a lateral acceleration of the road user, the accessibility analysis uses the longitudinal acceleration and the lateral acceleration of the road user, and the occupancy quantities of the road user are a forward reachable quantity over the predetermined period of time. [4] The method of claim 3, wherein the prediction of the occupancy quantities of the road user further comprises: Determining that the road user’s forward reachable quantity does not cross other road users during the specified period; and Determine that the road user's forward reachable amount is within the permissible road positions. [5] The method of claim 4, further comprising: the use of a reasoner to determine the downforce accelerations of the road user based on the input data; The occupancy corridor is determined based on the downforce accelerations of the road user. [6] The method of claim 5, wherein the road user occupancy corridor constraints include the road user's maximum acceleration and the road user's minimum acceleration. [7] The method of claim 6, further comprising determining that a trajectory of the host vehicle crosses the occupancy corridor of the road user. [8] The method of claim 7, further comprising removing the trajectory of the host vehicle that crosses the occupancy corridor of the road user. [9] The method of claim 6, further comprising determining that a trajectory of the host vehicle does not cross the occupancy corridor of the road user. [10] The method of claim 9, wherein controlling the movement of the host vehicle based on the occupancy corridor of the road user comprises a movement on the trajectory of the host vehicle that does not cross the occupancy corridor of the road user.

Images