Predictive frozen scenario for the continuous operation of automatic evasive maneuver systems

Abstract

A vehicle control system comprising a sensor for detecting an object location, a memory for storing an environment map indicating a driving environment coordinated with a location of a host vehicle, a processor configured to estimate line and object quality, determine uncertainty in response to historical movement of the objects, detect loss of sensor data in response to the data signal, generate a frozen scene construction in response to the line and object quality and the uncertainty, locate a location of the host vehicle relative to the frozen scene construction, regenerate object data and lane data in response to the frozen scene and the located location of the host vehicle, and generate a vehicle movement path in response to the object data and lane data.and a vehicle controller for controlling the host vehicle in response to the vehicle movement path.

Description

introduction

[0001] The present disclosure generally relates to programming control systems for motor vehicles. More specifically, aspects of this disclosure relate to systems, methods, and apparatus for generating uninterrupted vehicle trajectories including evasive steering features despite a temporary loss of sensor data due to rapid lateral vehicle acceleration in an ADAS-equipped vehicle.

[0002] The operation of modern vehicles is becoming increasingly automated, meaning they are capable of controlling driving with less and less driver intervention. Vehicle automation has been classified into numerical levels ranging from zero, meaning no automation with full human control, to five, meaning full automation with no human control. Various advanced driver assistance systems (ADAS) such as cruise control, adaptive cruise control, and parking assistance correspond to lower levels of automation, while true "driverless" vehicles correspond to higher levels of automation.

[0003] Adaptive control systems have been developed that not only maintain the set speed but also automatically decelerate the vehicle when a slower vehicle ahead is detected using various sensors such as radar and cameras. Additionally, some vehicle systems attempt to keep the vehicle in the center of a lane. However, maintaining a lane-by-lane speed on uneven roads or on roadways with other nearby hazards, such as construction barriers, could not only cause discomfort to vehicle occupants but could potentially lead to loss of vehicle control.

[0004] Conventional approaches to active safety have been anti-lock braking systems and traction control systems, which help maintain vehicle stability by sensing road conditions and intervening in the vehicle's braking and throttle control. Automated driving systems, however, can be further enhanced, for example, by supplementing the evasive maneuver control system with strategies that intervene in vehicle control when hazards are detected on or near the roadway. However, on-board sensors, particularly cameras and radars, tend to lose track of the lane or object during evasive maneuvers. It would be desirable to address these issues to provide a method and apparatus for providing an evasive maneuver control system with a continuous stream of lane and object information during evasive maneuvers in an ADAS-equipped motor vehicle.Furthermore, other desirable features and characteristics of the present disclosure will become apparent from the subsequent detailed description and the appended claims, taken in conjunction with the accompanying drawings and the foregoing technical field and background.

[0005] DE 10 2006 000 640 A1 describes a method for controlling a driver assistance system of a vehicle with at least one environmental sensor for collecting environmental data. The data collected by the environmental sensor is stored. If a malfunction occurs during the collection of environmental data, the stored data is used to derive estimated values for currently missing environmental data.

[0006] DE 10 2013 215 100 A1 describes a method for providing an environment model in the event of a failure of a first sensor of a vehicle, wherein the vehicle comprises the first sensor, based on at least the measurements of which environment models for the vehicle are successively created, wherein the environment models each provide information regarding the occupancy of the environment by objects.The information is only provided for occupancies up to a predetermined distance limit from the vehicle, the method comprising: providing a first environment model which was created based on at least the measurements of the first sensor at a first point in time at which the first sensor was still functional; determining, at a second point in time, that the first sensor is inoperative; in response to the determining: providing a second environment model by supplementing the first environment model with information regarding the occupancy by a phantom object, namely an object not detected on the basis of sensor measurements, the providing comprising: determining the occupancy by the phantom object in the second environment model taking into account the distance limit of the first environment model. Description

[0007] Presented herein are control systems for vehicles with driver assistance systems and methods, as well as the associated control logic for providing control systems for vehicles with driver assistance systems, methods for manufacturing and operating such systems, and motor vehicles equipped with control systems for vehicles with driver assistance systems. By way of example and without limitation, various embodiments of systems for providing evasive steering control systems with a continuous stream of lane and object information during evasive maneuvers in an ADAS-equipped motor vehicle are presented, which are disclosed herein.

[0008] According to the invention, a system for performing a vehicle maneuvering algorithm includes a sensor for generating a data signal indicative of an object location, a memory for storing an environment map indicative of a driving environment coordinated with a location of a host vehicle, a processor configured to estimate line and object quality, determine uncertainty in response to historical movement of the objects, detect loss of sensor data in response to the data signal, generate a frozen scene construction in response to the line and object quality and the uncertainty, locate a location of the host vehicle relative to the frozen scene construction,regenerates object data and lane data in response to the frozen scene and the localized location of the host vehicle, and generates a vehicle movement path in response to the object data and lane data, wherein the estimation of line and object quality and the determination of uncertainty are performed in response to a prediction of an impending evasive maneuver, and a vehicle controller for controlling the host vehicle in response to the vehicle movement path.

[0009] According to another exemplary embodiment, the sensor is a vehicle-mounted camera and the object location is determined in response to an image recognition algorithm executed by an image processor.

[0010] According to another exemplary embodiment, the vehicle maneuvering algorithm is initiated in response to an acceleration detected by an inertial measurement unit exceeding a threshold.

[0011] According to another exemplary embodiment, the object data and the lane data are reconstructed depending on vehicle telemetry data and an object reconstruction model.

[0012] According to another exemplary embodiment, the object data and the lane data are reconstructed in response to an uncertainty characterization in response to a historical linear object behavior uncertainty.

[0013] According to another exemplary embodiment, the construction of the frozen scene is predicted based on measurement model probabilities and covariances and a covariance adjustment.

[0014] According to another exemplary embodiment, the sensor set comprises a LiDAR.

[0015] According to another exemplary embodiment, the sensor set comprises at least one of a radar, a LiDAR, and a camera.

[0016] According to the invention, a method for providing a vehicle maneuver comprises generating a data signal by a sensor indicative of an object location, storing an environment map in memory indicative of a driving environment coordinated with a host vehicle position, determining line and object quality by the processor, detecting a loss of sensor data in response to the data signal by a processor, determining uncertainty in response to the historical movement of the objects, generating, by the processor, a construction of a frozen scene in response to the line and object quality and the uncertainty, locating, by the processor, a location of a host vehicle with respect to the construction of the frozen scene,The processor regenerates object and lane data in response to the frozen scene and the localized location of the host vehicle; the processor generates a vehicle motion path in response to the object and lane data; and a vehicle controller controls the host vehicle in response to the vehicle motion path. Determining line and object quality and determining uncertainty are performed in response to a prediction of an impending evasive maneuver.

[0017] According to another exemplary embodiment, the sensor set includes a vehicle-mounted camera and wherein the object location is determined in response to an image recognition algorithm executed by an image processor.

[0018] According to another exemplary embodiment, in response to the vehicle speed reduction control signal, a warning is further generated for the user indicating a hazard on the roadway.

[0019] According to another exemplary embodiment, the object data and the lane data are reconstructed depending on vehicle telemetry data and an object reconstruction model.

[0020] According to another exemplary embodiment, the speed reduction on the host lane is determined in response to a user preference associated with the first hazard.

[0021] According to another exemplary embodiment, the object data and the lane data are reconstructed in response to an uncertainty characterization in response to a historical linear object behavior uncertainty.

[0022] According to another exemplary embodiment, the construction of the frozen scene is predicted based on measurement model probabilities and covariances and a covariance adjustment.

[0023] According to another exemplary embodiment, the vehicle maneuvering algorithm is initiated in response to an acceleration detected by an inertial measurement unit exceeding a threshold.

[0024] According to another exemplary embodiment, the sensor set comprises at least one of a radar, a LiDAR, and a camera.

[0025] According to another exemplary embodiment, a vehicle control system for performing an evasive steering maneuver algorithm comprises a camera for capturing an image of a field of view including an object within the field of view, a distance sensor for detecting a distance to the object, an inertial measurement unit for detecting an acceleration of a host vehicle, a processor configured to detect a loss of sensor data, invalid data, or frozen data in response to the data signal, estimate line and object quality in response to the acceleration of the host vehicle exceeding a threshold acceleration, generate a frozen scene construct in response to the line and object quality, locate a location of the host vehicle relative to the frozen scene construct,to regenerate object data and lane data in response to the frozen scene and the localized location of the host vehicle, and to generate a vehicle movement path in response to the object data and lane data, and a vehicle controller for controlling the host vehicle in response to the vehicle movement path.

[0026] According to another exemplary embodiment, the object is a nearby vehicle. Brief description of the drawings

[0027] The present disclosure will now be described in conjunction with the following drawings, wherein like reference numerals designate like elements, and wherein: Fig. 1 is a functional block diagram illustrating an ADAS-equipped motor vehicle in accordance with various embodiments for implementing a lane hazard mitigation strategy; Fig. 2 illustrates an exemplary operating environment for implementing an evasive maneuver algorithm in an ADAS-equipped motor vehicle according to an exemplary embodiment of the present disclosure; Fig. 3 shows a block diagram of an exemplary system for providing an evasive maneuver algorithm in an ADAS-equipped motor vehicle according to an exemplary embodiment of the present disclosure; and Fig. 4 shows a flowchart illustrating a method for providing an evasive maneuver algorithm in an ADAS-equipped motor vehicle according to an exemplary embodiment of the present disclosure. Detailed description

[0028] The following detailed description is merely exemplary in nature and is not intended to limit the disclosure or its application and uses. Furthermore, there is no intention to be bound by any of the theories presented in the foregoing background or the following detailed description.

[0029] Fig. 1 illustrates an exemplary operating system 100 for implementing an evasive maneuver algorithm in an ADAS-equipped motor vehicle 10, as described below in connection with the vehicle 10 of Fig. 1 and the surrounding area 200 of Fig. 2 and the implementations of Fig. 3 and Fig. 4 is described in more detail.

[0030] In various embodiments, vehicle 10 includes an automobile. Vehicle 10 may be any type of automobile, such as a sedan, station wagon, truck, or sport utility vehicle (SUV), and may have two-wheel drive (2WD) (i.e., rear-wheel drive or front-wheel drive), four-wheel drive (4WD), or all-wheel drive (AWD), and / or, in certain embodiments, various other vehicle types. In certain embodiments, vehicle 100 may also include a motorcycle or other vehicle, such as an aircraft, spacecraft, watercraft, etc., and / or one or more other types of mobile platforms (e.g., a robot and / or other mobile platform).

[0031] As in Fig. 1, the vehicle 10 generally includes a chassis 12, a body 14, front wheels 16, and rear wheels 18. The body 14 is disposed on the chassis 12 and substantially encloses components of the vehicle 10. The body 14 and the chassis 12 may together form a frame. The wheels 16-18 are each rotatably connected to the chassis 12 near a corner of the body 14.

[0032] In various embodiments, vehicle 10 may be an ADAS-equipped vehicle with an operating system 100 installed for implementing a lane hazard mitigation strategy (hereinafter referred to as vehicle 10). For example, vehicle 10 is a vehicle that can be automatically controlled to transport passengers from one location to another. Vehicle 10 is illustrated as a passenger car in the depicted embodiment, but it should be understood that any other vehicle, including motorcycles, trucks, sport utility vehicles (SUVs), recreational vehicles (RVs), watercraft, aircraft, etc., may also be used. In an exemplary embodiment, autonomous vehicle 10 is autonomous in that it provides partial or fully automated assistance to the driver operating vehicle 10.As used herein, the term operator includes a driver of the vehicle 10 and / or an autonomous driving system of the vehicle 10.

[0033] As illustrated, the autonomous vehicle 10 generally includes a propulsion 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 controller 34, and a communication system 36. The propulsion system 20, in various embodiments, may include an internal combustion engine, an electric machine such as a traction motor, and / or a fuel cell propulsion system. The transmission system 22 is configured to transmit the power of the propulsion system 20 to the vehicle wheels 16-18 according to selectable speed ratios. According to various embodiments, the transmission system 22 may include a continuously variable automatic transmission, a stepped transmission, or other suitable transmission. The braking system 26 is configured to apply braking torque to the vehicle wheels 16-18.The braking system 26 may, in various embodiments, include friction brakes, brake-by-wire systems, a regenerative braking system such as an electric machine, and / or other suitable braking systems. The steering system 24 influences a position of the vehicle wheels 16-18. Although a steering wheel is shown for illustrative purposes, the steering system 24 may not include a steering wheel in some embodiments contemplated by the present disclosure.

[0034] The sensor system 28 includes one or more sensing devices 40a-40n that sense observable conditions of the external environment and / or the internal environment of the autonomous vehicle 10. The sensing devices 40a-40n may include, but are not limited to, radars, lidars, global positioning systems, optical cameras, thermal cameras, ultrasonic sensors, inertial measurement units, and / or other sensors.

[0035] The actuator system 30 includes one or more actuator devices 42a-42n that control one or more vehicle functions, such as, but not limited to, the drive system 20, the transmission system 22, the steering system 24, and the braking system 26. In various embodiments, the vehicle features may also include interior and / or exterior features of the vehicle, such as doors, a trunk, and cabin features such as air, music, lighting, etc. (unnumbered).

[0036] The communication system 36 is configured to wirelessly transmit information to and from other entities 48, such as other vehicles (“V2V” communication), infrastructure (“V2I” communication), remote systems, and / or personal devices (described in more detail with respect to Fig. 2). In an exemplary embodiment, the communication system 36 is a wireless communication system configured to communicate over a wireless local area network (WLAN) using IEEE 802.11 standards or using cellular data communications. However, additional or alternative communication methods, such as a dedicated short-range communication channel (DSRC), are also contemplated within the scope of this disclosure. DSRC channels refer to short- to medium-range, one-way or two-way wireless communication channels specifically designed for use in motor vehicles, as well as a set of protocols and standards.

[0037] The data storage device 32 stores data for use in the automatic control of the autonomous vehicle 10. In various embodiments, the data storage device 32 stores defined maps of the navigable environment. In various embodiments, the defined maps may be predefined by and obtained from a remote system (described in more detail with respect to Fig. 2). For example, the defined maps may be compiled by the remote system and transmitted to the autonomous vehicle 10 (wirelessly and / or wired) and stored in the device 32. As can be appreciated, the data storage device 32 may be part of the controller 34, separate from the controller 34, or part of the controller 34 and part of a separate system.

[0038] The controller 34 includes at least one processor 44 and a computer-readable storage device or medium 46. The processor 44 may be any custom or off-the-shelf processor, a central processing unit (CPU), a graphics processing unit (GPU), an auxiliary processor among multiple processors connected to the controller 34, a semiconductor-based microprocessor (in the form of a microchip or chipset), a macroprocessor, any combination thereof, or generally any device for executing instructions. The computer-readable devices or media 46 may include volatile and non-volatile memory, such as read-only memory (ROM), random access memory (RAM), and keep-alive memory (KAM). KAM is volatile 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(s) 46 may be implemented using any number of known storage devices such as PROMs (programmable read-only memory), EPROMs (electrically erasable PROMs), EEPROMs (electrically erasable PROMs), flash memory, or other electrical, magnetic, optical, or combination storage devices capable of storing data, some of which represent executable instructions used by the controller 34 in controlling the vehicle 10.

[0039] The instructions may comprise one or more separate programs, each comprising an ordered collection of executable instructions for implementing logical functions. The instructions, when executed by processor 44, receive and process signals from sensor system 28, perform logic, calculations, methods, and / or algorithms to automatically control the components of autonomous vehicle 10, and generate control signals for actuator system 30 to automatically control the components of autonomous vehicle 10 based on the logic, calculations, methods, and / or algorithms. Although in Fig. 1 only one controller 34 is shown, embodiments of the autonomous vehicle 10 may include any number of controllers 34 that communicate through communication messages over any suitable communication medium or combination of communication media and that cooperate to process the sensor signals, perform logic, calculations, methods and / or algorithms, and generate control signals to automatically control features of the vehicle 10.

[0040] In various embodiments, one or more instructions of controller 34 are implemented in quality and safety rating system 100 and, when executed by processor 44, process sensor data from sensing devices 40a-40n, message data from communication medium and / or communication system 36, and / or data sent to or received from actuator devices 42a-42n, and calculate ratings and explanations about the safety and ride quality of the operator of vehicle 10.

[0041] Now with reference to Fig. 2 illustrates an exemplary operating environment 200 for an automated evasive maneuver system in an ADAS-equipped motor vehicle 210. In this exemplary embodiment of the present disclosure, the host vehicle 210 is traveling on a multi-lane roadway 205 with various hazards, such as a leading vehicle 220 and potholes 230. A predicted evasive maneuver 225 is shown. In some exemplary embodiments, the host vehicle 210 is equipped with an ADAS feature such as adaptive cruise control and automatic lane centering. During assisted driving, the ADAS controller controls the fueling, braking, and steering of the host vehicle 210 to control the vehicle speed and position to keep the host vehicle 210 within the lane.The ADAS controller may be configured to maintain a constant speed, detect nearby objects, such as the approach of a slow-moving vehicle ahead within the lane, and reduce the vehicle speed accordingly. Likewise, the ADAS controller may be configured to keep the host vehicle 210 centered in the current lane and perform evasive steering functions to control the steering of the host vehicle 210 within the lane in response to approaching vehicles, detected hazards such as debris, or the like. When the ADAS is active, the ADAS may also perform the evasive steering control algorithm to enhance safety by automatically maneuvering the host vehicle 210 to avoid obstacles and hazards.

[0042] The rapid response time of the evasive steering control is critical for successful evasive maneuvers in the ADAS-equipped vehicle 210. Faster response times allow the system to detect a threat, assess the situation, and initiate an evasive maneuver more quickly, maximizing the available maneuvering space. Delays can significantly limit the ability of the host vehicle 210 and the evasive steering control algorithm to perform precise maneuvers. A faster response results in sharper turns and more precise trajectory control, increasing the success rate of the evasive maneuver. By responding quickly, the system can initiate corrective action before the vehicle's dynamics are significantly impacted by the impending collision, improving overall stability during the evasive maneuver.

[0043] Timely sensor data is critical for the host vehicle 210 and the evasive steering control algorithm to perform precise evasive maneuvers in the shortest possible time. To maintain evasive steering functionality during momentary sensor failures caused by rapid vehicle movements, the evasive steering control algorithm and the ADAS controller can employ a multi-stage approach that prioritizes safety and maneuverability. First, upon a temporary loss of sensor data, the evasive steering control algorithm freezes the perceived environment. This creates a static snapshot of the environment based on the most recently received sensor data.Next, the evasive steering control algorithm uses vehicle telemetry data, which tracks the vehicle's movement based on wheel revolutions, to locate the host vehicle 210 within the previously detected lane markings 207, allowing the host vehicle's position within the lane to be maintained even without real-time sensor updates. Finally, the algorithm applies a predictive model to anticipate the movement of surrounding objects on the road. This prediction incorporates safety margins to ensure that sufficient separation is maintained between the ego vehicle and surrounding objects during the maneuver. By combining these steps, the algorithm enables uninterrupted trajectory generation and control.This allows the host vehicle 210 to effectively perform evasive maneuvers even in the event of a temporary loss of sensor data, which increases overall safety in critical situations.

[0044] Now with reference to Fig. Figure 3 shows a block diagram illustrating an example implementation of a system 300 for performing an evasive steering maneuver in an ADAS-equipped motor vehicle. The example system 300 may include a processor 320, a camera 340, LiDAR 343, radar 344, inertial measurement unit (IMU) 354, and GPS sensor 345. Furthermore, the processor 320 may receive stored information such as map data 350 from a memory 350 or the like, as well as user input via a user interface 353.

[0045] Camera 340 may be a low-fidelity camera with a forward-looking field of view (FOV). Camera 340 may be mounted inside the vehicle behind the rearview mirror or on the vehicle's front fascia. Camera 340 may capture images and / or videos that may be used to detect leading and following vehicles, obstacles, lane markings, road edges, pavement features, other pavement markings, and road hazards during ADAS operation. The images captured by camera 340 and the data generated from the images may be used to supplement the map data stored in memory 350.

[0046] The LiDAR 343 is configured to emit light pulses at a known altitude and azimuth and measure a time interval between the emission of the light pulse and a detected reflection of the light pulse. This time interval can be used to determine the distance to a surface at this known altitude and azimuth. The LiDAR 343 can then perform this distance measurement over a variety of sets of known azimuth and altitude values to generate a detailed point cloud of the surrounding area. The point cloud can be used to identify nearby objects based on their size and shape. This is particularly effective for static objects and larger dynamic objects. The Radar 344 emits radio waves and analyzes the reflected signals to determine the distance, angle, and relative speed of surrounding objects.The radar 344 can be used to detect the distance of preceding vehicles for automatic emergency braking algorithms and can serve as a complementary sensor to the camera 340 and the LiDAR 343 to create a comprehensive environmental map for the ADAS processor 320.

[0047] The GPS sensor 345 may be part of a global navigation satellite system (GNSS) to receive a plurality of time-stamped satellite signals containing the location data of a transmitting satellite. The GPS controller then uses this information to determine a precise location of the GPS sensor 345. The processor 320 may be configured to receive the location data from the GPS controller and store this location data in the memory 350. The memory 350 may be configured to store map data for use by the processor 320. The memory 350 may further be configured to store map data, where the map data may be high-resolution map data containing detailed representations of roadways, including precise roadway positions, lane positions, curves, elevations, known lane hazards, and other roadway details.

[0048] The processor 320 is configured to receive data from the various sensors and user input and generate an environment map centered on the host vehicle. The environment map includes positions of static objects, such as lane markings, light poles, stop signs, and the like, as well as dynamic objects, such as nearby vehicles, pedestrians, cyclists, etc. Typically, the processor 320 uses stored map data and sensor data to create a comprehensive map of the static environment. The processor 320 then detects and tracks dynamic objects within the environment and adds this data to the environment map. Using LiDAR data, precise distance measurements can be obtained, allowing algorithms to identify objects based on their size and shape. This is particularly effective for static objects and larger dynamic objects.Image data can be used to identify and classify objects such as vehicles, pedestrians, and traffic signs. Furthermore, Kalman filters can be used to track the movement of dynamic objects over time by considering the dynamic object's position, velocity, and sensor noise; Kalman filters can be used to predict the object's future position and compensate for potential inconsistencies in the sensor data. Sensor fusion algorithms combine data from LiDAR 343, camera 340, radar 344, and other sensors to create a robust and comprehensive environmental map. This combined data can compensate for the limitations of individual sensors and the temporary loss of data from one or more sensors, resulting in more accurate object detection and tracking.Processor 320 is further configured to execute algorithms for detecting and compensating for discrepancies between data from the various sensors, as well as for differences between real-time sensor data and stored map data. Because the environment is constantly changing, processor 320 continuously updates the environment map data in response to newly acquired sensor data.

[0049] The processor 320 is further configured to activate and control the ADAS in response to a user initiation of the ADAS via the user interface 353. In ADAS operation, the processor 320 may be configured to generate a desired path in response to a user input or the like, where the desired path may include lane centering, cornering, lane changing, etc. This desired path information may be determined in response to the vehicle speed, yaw angle, and lateral position of the vehicle within the lane. Once the desired path is determined, a control signal indicative of the desired path is generated by the processor 320 and coupled to the vehicle controller 330.The vehicle controller 330 is configured to receive the control signal and generate an individual steering control signal coupled to the steering controller 370, a braking control signal coupled to the braking controller 360, and a throttle control signal coupled to the throttle controller 355 to travel the desired path.

[0050] Like all vehicle systems, the sensors in the vehicle face several challenges that prevent them from perfectly perceiving the environment. These include limitations in range and resolution, occlusions, environmental sensitivity, and mechanical and electrical interference. Due to complex or cluttered environments or due to data fusion issues, ADAS algorithms can lose track of lanes / objects. Safety-critical operations, such as evasive control systems, require a continuous stream of lane and object information during operation. The processor 320 can execute various algorithms to address such sensor deficiencies, ensuring uninterrupted trajectory and control to perform an evasive maneuver and steer the vehicle along the desired path.

[0051] For example, the processor 320 may initiate an evasive steering maneuver algorithm in response to environment map and sensor data when an impending hazardous situation is detected. While executing the evasive steering maneuver algorithm, the processor 320 may then detect a loss of sensor data. In response to the detected loss of sensor data, the processor 320 may make an estimate of lane line and object quality. The lane line and object quality estimate may use vehicle telemetry and the predicted vehicle path, as well as other data, such as tire limits, to predict lane quality based on a dynamic level of vehicle movement. The processor 320 may combine the lane line and object quality to predict scene degradation over a predicted horizon.Processor 320 then uses this estimate of lane line and object quality to generate a frozen scene construction of the environment surrounding the host vehicle. Processor 320 uses vehicle telemetry data to locate the host vehicle's position within the frozen scene. In response to the frozen scene, processor 320 regenerates object and lane data, which is used to update the frozen scene construction and the host vehicle's localization. In response to the newly generated object and lane data, processor 320 then updates the object detection and tracking for use by the evasive steering maneuver algorithm. The evasive steering algorithm can then adjust the vehicle's movement path depending on the predicted object detection and tracking data.

[0052] In some example embodiments, processor 320 is configured to receive a data stream from the various sensors, vehicle controller 330, and various user interfaces, such as fuel control 355, braking control 360, and steering control 370. In response to this data, processor 320 can predict the possibility of an impending evasive maneuver. For example, processor 320 can consider factors such as the position, speed, and direction of the host vehicle, as well as static and dynamic objects in the environment, to calculate the likelihood of high-risk conditions. Advanced systems can use machine learning to analyze large amounts of traffic data to identify risky situations and pedestrian behavior patterns.

[0053] In response to predicting a potentially imminent evasive maneuver, processor 320 may begin creating an environment map to track static and dynamic objects near the host vehicle, such as lane markings, and to locate the host vehicle within the environment map. In the event of sensor data loss during an evasive maneuver, processor 320 may freeze the environment map, estimate the positions of dynamic objects depending on previous tracking data, and estimate the vehicle position within the environment depending on other vehicle telemetry data or the like. Lane geometry may be estimated using a recursive least squares method using third-order polynomial coefficients. For example, a lane polynomial from an image captured at time k for lane j is given by: yi,j(xi,j)=C0,j|k+C1,j|kxi,j+C2,j|kxi,j2+C3,j|kxi,j3

[0054] Where the unknown lane polynomial at time k+t is given by: [RHPi,j→|k+t]y=+ C0,j|k+t+C1.j|k+t[RHPi,j→|k+t]x+C2,j|k+t[RHPi,j→|k+t]x2+C3,j|k+t[RHPi,j→|k+t]x3

[0055] The recursive least squares reconstruction of the lane coefficients at time k=t (when the image data was interrupted) is then given by: [C^0,j|k+t C^1,j|k+t C^2,j|k+t C^3,j|k+t]=[[RHP0,j→|k+t]y[RHP1,j→|k+t]y⋮[RHPn,j→|k+t]y] / [11⋯1[RHP0,j→|k+t]x[RHP1,j→|k+t]x⋯[RHPn,j→ |k+t]x[RHP0,j→|k+t]x2[RHP1,j→|k+t]x2⋯[RHPn,j→|k+t]x2[RHP0,j→|k+t]x3[RHP1,j→|k+t]x3⋯[RHPn,j→|k+t]x3]

[0056] In the case of an assisted driving algorithm, the processor 320 may then generate a movement path in response to the map of the frozen environment. If no current data is available, the processor 320 may use the frozen environment and dead reckoning to create the movement path. Dead reckoning uses a known starting point, heading or compass direction, estimated speed, and elapsed time to calculate the current position. In addition, the processor 320 may calculate a safety zone around the dynamic object depending on the relative kinematics of the dynamic object. For example, if the relative kinematics of the dynamic object are greater, such as higher nonlinear object behavior, a larger bounded uncertainty may be approximated compared to that of a dynamic object with historically linear behavior.The relative kinematics of the target with respect to the host during the automatic evasive maneuver is given by:. RHT|→k+t=RHT|→k+RTT|→k+t−RHH|→k+t where R T H at time k+t is the unknown, where the camera and radar cannot reliably detect the dynamic object at time k+t due to the evasive movement of the host vehicle, R T H at time k are the data acquired by the camera and the radar at time k, R T T at time k+t is the unknown (bounded uncertainty) and R H H at time k+t is determined from the telemetry of the host vehicle between time k and time k+t.

[0057] The relative target kinematics can be approximated by linear propagation of the states at time k+t plus a limited uncertainty: RTT|→k+t=RT|→k+VT|→k⋅t+ΔRT|→k+t

[0058] Therefore, the relative position of the dynamic object when the sensors no longer detect the object can be described as follows: RHT|→k+t=RHT|→k−RHH|→k+t+RT|→k+VT→|k+ΔR→=f(RHT→|k,RHH→|k+t,RT→|k,VT→|k,ΔR→)

[0059] Similarly, velocity and acceleration can be described as follows: RHT→|k+t=f(RHT→|k,RHH→|k+t,RT→|k,VT→|k,ΔR→) VHT|→k+t=f(VHT→|k,VHH→|k+t,ωH→|k,VT→|k,AT→|k,ωT→|k,ΔV→) AHT|→k+t=f(AHT→|k,AHH→|k+t,ωH→|k,αH→|k,VT→|k,AT→|k,ωT→|k,αH→|kΔA→)

[0060] The Kalman prediction can be extended to include the host dynamics and characteristic uncertainties, as provided as follows: z^k+1|k=Az^k|k+ΔK ΔK=f(RHT→|k+t,VHT→|k+t,AHT→|k+t,ΔR→,ΔV→,ΔA→) Pk+1|k=APk|kAT+BQBT

[0061] Now with reference to Fig.Figure 4 shows a flowchart illustrating an exemplary implementation of a method 400 for generating a predicted frozen scene for the uninterrupted operation of an automatic evasive maneuver system in an ADAS-equipped motor vehicle. The proposed method 400 is configured to ensure that evasive steering functions continue to function despite the temporary loss of sensor data due to the rapid lateral dynamics of the host vehicle. The method 400 freezes the scene when sensor information is lost, localizes the host vehicle using historically detected lane lines based on odometry, predicts the dynamics of environmental actors while considering safety margins, and enables uninterrupted trajectory generation and control performance during evasive maneuvers.

[0062] The method 400 is initially configured to initiate the evasive steering maneuver algorithm at 405. In response to initiating the evasive steering maneuver algorithm, the method 400 may generate a vehicle movement path at 410 in response to environment map data, as well as detected static and dynamic positions of nearby objects and their tracking. The method 400 may receive data from a plurality of sensors and generate an environment map representative of the host vehicle's environment. The motion planning algorithm then executes a path planning algorithm that considers the received sensor data, including information about static elements such as lane markings and dynamic objects such as other vehicles, as well as additional parameters such as traffic regulations and safety margins, to create a feasible trajectory.The motion path is then refined by the motion planning algorithm, which ensures that the trajectory can be kinematically achieved by the host vehicle, taking into account its acceleration, braking and turning capabilities.

[0063] In response to the generated vehicle motion path, method 400 may then control the host vehicle along the motion path 415. Method 400 may apply a refined method for generating control signals by calculating a lateral error, the distance between the vehicle and the planned path. A trajectory optimization module refines the planned trajectory based on the lateral error and factors such as upcoming curves or obstacles detected by the sensors. The method may employ model predictive control to anticipate future situations and generate control signals, as well as a dynamic model that considers vehicle dynamics, road conditions, and weather data. The generated control signals are then translated into specific electronic instructions for steering, acceleration, and braking.These electronic signals can be transmitted to a vehicle control unit, which acts as an interface between the ADAS controller and the host vehicle's braking, throttle, and steering controllers. The vehicle control unit can translate the received signals into commands for individual components: The electric power steering adjusts the steering angle, the engine control unit (ECU) modulates engine power, and the anti-lock braking system (ABS) and electronic brake-force distribution (EBD) work together to control the individual wheel brakes.

[0064] The method next determines at 420 whether a loss of sensor data has been detected. In some example embodiments, method 400 may utilize a sensor health monitoring module integrated into the processor unit. The sensor health monitoring module receives raw data streams from each sensor within the sensor set. It then performs a multi-step analysis to identify potential sensor malfunctions or data losses. The sensor health monitoring module may perform data validity checks to determine whether the received data conforms to the predefined parameters for each sensor type. For example, it may look for implausible values or data ranges that deviate significantly from normal operation. Temporal consistency checks may be performed to analyze the temporal consistency of the incoming data.Unexpected delays, sudden drops in data frequency, or inconsistencies between timestamps from different sensors may indicate potential problems. Sensor redundancy checks can be performed on data from different sensors to compare their outputs for consistency. Significant discrepancies may indicate a malfunction in one of the sensors.

[0065] If no loss of sensor data was detected at 420, the method generates an updated motion path at 455 in response to the received sensor data and continues to follow the updated motion path at 415. If a loss of sensor data was detected at 420 during execution of the evasive steering maneuver algorithm, the method 400 may perform a lane line and object quality estimation 425. The lane line and object quality estimation may use vehicle telemetry and the predicted vehicle path, as well as other data, such as tire boundaries, to predict lane quality based on a dynamic level of vehicle movement. The method 400 may combine lane and object quality to predict scene degradation over a predicted horizon.

[0066] The method 400 then uses this lane line and object quality estimate to generate a frozen scene construction of the environment around the host vehicle at 430.

[0067] The method 400 uses vehicle telemetry data to locate the location of the host vehicle within the frozen scene at 435. The frozen scene concept reconstructs the missing lane / object information and provides a continuous data stream for evasive steering features.

[0068] In response to the frozen scene, the method 400 regenerates object and lane data at 440, which is used to update the frozen scene construction and the host vehicle localization. In some example embodiments, the method 400 may localize lane information related to the host vehicle when the sensory lane data is interrupted using host vehicle odometry, historical camera data, and recursive estimation. In some example embodiments, the object reconstruction may use vehicle telemetry data with an object reconstruction model. The object reconstruction model characterizes the degree of uncertainty of the object to be avoided. An extended Kalman model may be initialized, a prediction and state propagation model may be created, such asFor slow target dynamics or fast host dynamics, a covariance adjustment can be created, such as an armed, reconstructed, or partial update. An updated prediction can be created based on mode probabilities and covariances. Finally, the estimated convergence can be verified. In response, a reconstructed relative position, velocity, and heading of the object can be generated and incorporated into the updated environment map.

[0069] For example, lane lines in the frozen scene may be localized in response to the loss of lane information due to the evasive maneuver. The method 400 may use a historical buffer of lane information and host vehicle telemetry data, such as steering angles, IMU data, host vehicle speeds, GPS position data, and wheel speeds, as inputs. The method 400 may generate a lane polynomial from camera images and perform recursive least-squares reconstruction of the lane coefficients. The relative kinematics of the object with respect to the host vehicle during the maneuver are approximated by linearly propagating states in a time interval plus a bounded uncertainty. The maximum uncertainty bound may be interpreted as a function of the target behavior over previous sample times.The method 400 may further include supplementing a Kalman filter prediction with ego dynamics based on a trigger indicating the initiation of an evasive steering maneuver, plus the characterized uncertainties in the event of loss of sensor data.

[0070] In response to the newly generated object and lane data, method 400 then updates the object detection and tracking at 450 for use by the evasive steering maneuver algorithm. This updated object detection may also establish a safety margin to ensure safe operation of the ego vehicle during evasive maneuvers. In some example embodiments, an uncertainty band is calculated based on historical object motion and updated with sporadic sensor measurements. Method 400 may perform object tracking in the host vehicle's coordinate system during highly dynamic maneuvers using an extended filter, such as a Kalman filter, to utilize reconstructed object data in combination with host vehicle odometry.

[0071] The evasive steering maneuver algorithm may then update the vehicle's path of travel in response to the predicted object detection and tracking data at 455. In response to the updated path of travel, the method 400 may then control the host vehicle to follow the updated path of travel at 415.

[0072] It will be appreciated that the systems, vehicles, and methods may vary from those illustrated in the figures and described herein. Although at least one exemplary embodiment has been presented in the foregoing detailed description, it should be recognized that a wide variety of variations exist. The exemplary embodiment or embodiments are merely examples and are not intended to limit the scope, applicability, or configuration of the disclosure in any way. Rather, the foregoing detailed description is intended to provide those skilled in the art with a practical guide for implementing the exemplary embodiment or embodiments.It should be understood that various changes in the function and arrangement of elements may be made without departing from the scope of the disclosure as set forth in the appended claims and their legal equivalents.

Claims

[1] System (300) for performing a vehicle maneuvering algorithm, comprising: a sensor set for generating a data signal indicative of an object location and a line; a memory (350) for storing an environment map indicating a driving environment coordinated with a location of a host vehicle (210); a processor (320) configured to estimate line and object quality, determine uncertainty in response to the historical movement of the objects, detect loss of sensor data in response to the data signal, generate a frozen scene construction in response to the line and object quality and the uncertainty, locate a location of the host vehicle (210) relative to the frozen scene construction, regenerate object data and lane data in response to the frozen scene and the located location of the host vehicle (210), and generate a vehicle movement path in response to the object data and lane data, wherein the estimation of the line and object quality and the determination of the uncertainty are performed in response to a prediction of an impending evasive maneuver; and a vehicle controller (330) for controlling the host vehicle (210) in dependence on the vehicle movement path. [2] A system (300) for performing a vehicle maneuvering algorithm according to claim 1, wherein the sensor set comprises a camera mounted on the vehicle (210), and wherein the object location is determined in response to an image recognition algorithm executed by an image processor. [3] A system (300) for performing a vehicle maneuvering algorithm according to claim 1, wherein the vehicle maneuvering algorithm is initiated in response to an acceleration detected by an inertial measurement unit exceeding a threshold. [4] A system (300) for performing a vehicle maneuvering algorithm according to claim 1, wherein the object data and the lane data are reconstructed in response to vehicle telemetry data and an object reconstruction model. [5] A system (300) for performing a vehicle maneuvering algorithm according to claim 1, wherein the object data and the lane data are reconstructed in response to an uncertainty characterization in response to a historical linear object behavior uncertainty. [6] A system (300) for performing a vehicle maneuvering algorithm according to claim 1, wherein the construction of the frozen scene is predicted based on measurement model probabilities and covariances and a covariance adjustment. [7] A system (300) for performing a vehicle maneuvering algorithm according to claim 1, wherein the sensor set includes a LiDAR. [8] A system (300) for performing a vehicle maneuvering algorithm according to claim 1, wherein the sensor set comprises a radar, a LiDAR, and a camera. [9] A method for providing a vehicle maneuver, comprising: Generating a data signal indicative of an object location and a lane line position by a sensor set; Storing an environment map in a memory indicating a driving environment coordinated with the location of a host vehicle (210); Determining a lane line quality and an object quality by a processor (320); detecting a loss of sensor data in response to the data signal by the processor (320); Determination of uncertainty in response to the historical movement of objects; generating a frozen scene construction by the processor (320) in response to the lane line quality, the object quality and the uncertainty; Locating a position of the host vehicle (210) with respect to the construction of the frozen scene by the processor (320); Regenerating object data and lane data by the processor (320) depending on the frozen scene and the localized location of the host vehicle (210); generating a vehicle movement path by the processor (320) in response to the object data and the lane data; and Controlling the host vehicle (210) by a vehicle controller (330) in dependence on the vehicle movement path; wherein determining lane line and object quality and determining uncertainty are performed in response to a prediction of an impending evasive maneuver.

Images