System, method and device for estimating the hitching angle of a hitching device

Abstract

In various embodiments, methods, systems, and vehicle devices are provided. One method involves estimating the hitch articulation angle (HAA) using ultrasonic sensors (USS) while employing likelihood filtering to ensure the quality of the detected echo signal performance. This includes generating at least one set of USS data based on a set of echo signals generated by multiple ultrasonic sensors configured around a vehicle coupled to a trailer; determining, during likelihood filtering for any frontal shape of the trailer, using a selected set of geometric equations based on the set of USS data; and estimating, based on the at least one set of USS data, at least one comparison with a motion model at low speed to ensure that the motion model's result for the HAA associated with the determined trailer shape is based on a pair of detected echo signals considered to have higher signal performance.

Description

Technical Field

[0001] This technical field generally relates to vehicle kinematics modeling systems, methods, and apparatuses, and more specifically, to systems, methods, and apparatuses for estimating the hitch articulation angle (HAA) using ultrasonic sensors (USS), kinematic models, and other sensor modes during various towing dynamic events in traction operations, in order to evaluate the signal quality of the trailer front of arbitrary shapes and estimate the hitch articulation angle (HAA). Background Technology

[0002] Autonomous, semi-autonomous, and conventional vehicles can be designed to accommodate a wide range of loads for towing or trailering. Different front-end trailers include, but are not limited to, flatbed trailers, enclosed trailers, cargo trailers, campervans, boats, and sometimes other motor vehicles. Furthermore, various trailer hitches are used in towing operations, such as gooseneck hitches, weight-distribution hitches, pivot hitches, receiver hitches, and fifth-wheel hitches. Each configuration of trailer type and hitch type exhibits different vehicle dynamics. While there are (though limited) systems and devices available to enhance vehicle and trailer stability during towing operations, given the diverse combinations of trailers and hitches, there is no single, universal solution, nor even one that can cover most or almost all potential combinations of vehicles, trailers, and hitches in a coupled operation. In addition, to improve the stability of numerous coupled combinations, manufacturers have promoted basic add-ons such as vehicle trim and deflectors for ground effect, which provide additional aerodynamic stability to the vehicle. In addition, a traction control system and an automatically adjusting suspension system have been developed, which can change the vehicle height according to the load weight. The vehicle traction system may still be improved in many aspects.

[0003] Ultrasonic sensors (USS) have traditionally been used for vehicle parking assistance functions. When towing a trailer, an array of ultrasonic sensors (USS) can be used to estimate the hitch articulation angle (HAA). However, the implementation of USS is limited, partly due to practical obstacles such as high levels of interference in signal reception noise, unwanted reflections in addition to those from the front of the trailer, such as reflections from the hitch point, and uncertain trailer shape.

[0004] Therefore, it is desirable to realize methods, systems, and apparatuses with improved USS for processing raw data, as well as signal and logic filtering of an additional layer for at least accurate estimation of the hitch articulation angle (HAA) when towing a trailer.

[0005] Furthermore, other desirable features and characteristics of the invention will become apparent from the following detailed description and appended claims, taken in conjunction with the accompanying drawings and the foregoing technical field and background. Summary of the Invention

[0006] A system, method, and apparatus are disclosed for estimating the hitch articulation angle (HAA) using an ultrasonic sensor (USS) and a kinematic model, evaluating the quality of the estimated HAA value by comparing the USS, kinematic, and other modal data, and then fusing it with other sensor data to estimate the hitch articulation angle (HAA) of the front of an arbitrary-shaped trailer during various towing dynamic events in a traction operation.

[0007] In one exemplary embodiment, a system is provided including a processing unit disposed in a vehicle, the processing unit including one or more processors configured by programming instructions encoded on a non-transient computer-readable medium. The processing unit is configured to: generate at least one set of ultrasonic sensor (USS) data based on detecting at least one set of echo signals generated by a plurality of USSs configured to surround a vehicle already pointed at a trailer, to determine the hitch articulation angle (HAA) of the trailer to the vehicle; filter the at least one set of USS data based on the detection of the at least one set of echo signals, the at least one set of USS data being evaluated based on echo signal quality, consistency with driving conditions including straight-line driving, and the plausibility of the echo signal steering input; and, in response to the filtering of the at least one set of echo signals, pass... The likelihood filtering process performs calculations using a set of geometric equations selected from one of multiple sets of different geometric equations, estimates based on at least one set of USS data to determine at least an arbitrary frontal shape of the trailer; classifies the estimated arbitrary frontal shape of the trailer based at least on a comparison of measured distances from detected echo signals about the trailer to the determined frontal shape of the trailer; and generates at least one comparison with a kinematic model based on the estimation results of at least one set of USS data to ensure that the kinematic model is considered to have higher signal performance for the HAA results associated with the determined frontal shape of the trailer, wherein the higher signal performance is based on the correlation with the kinematic model at low speeds.

[0008] In at least one exemplary embodiment, the system includes a processing unit configured to use likelihood filtering on a plurality of driving conditions, including at least straight driving and steering angle consistency, and on the determination of the trailer frontal shape.

[0009] In at least one exemplary embodiment, the system includes a processing unit configured to: remove USS data inconsistent with a plurality of driving conditions, the plurality of driving conditions including at least straight driving at low speed and steering angle input; and classify the trailer frontal shape to select a geometric equation for determining the HAA.

[0010] In at least one exemplary embodiment, the system includes a processing unit configured to determine a range of likelihoods for acceptable distance measurements of USS echoes reflected and collected from the front of the trailer during straight-line driving using a set of measurements including steering wheel angle (SWA) or road wheel angle (RWA); and to establish a range of likelihoods based on the SWA or RWA measurements to remove USS echoes that fall outside the acceptable distance measurement range.

[0011] In at least one exemplary embodiment, the system includes a processing unit configured to optimize the use of USS data by comparing HAA results estimated from at least one set of USS data with HAA results derived from a kinematic model at low speeds, which can provide a quality metric for the implementation of USS HAA calculations with associated, defined trailer frontal shapes.

[0012] In at least one exemplary embodiment, the system includes a processing unit configured to compare the HAA derived from a kinematic model and the USS process HAA at low speeds to determine a quality metric for the USS based on a mean squared error calculation with sufficient data points using an acceptable threshold.

[0013] In at least one exemplary embodiment, the system includes a processing unit configured to: determine a quality index of the kinematic model by comparing a kinematic model-derived HAA and a USS process HAA under the following conditions: the kinematic model is derived based on at least a set of assumptions that trailer dynamics are considered invalid, said trailer dynamics including dynamics at higher speeds, at higher speeds with steering input, rough road conditions acting on the vehicle and trailer above a threshold, and crosswind conditions acting on the vehicle and trailer above a threshold; and use a set of values ​​already calculated for the HAA results and the determined quality index to... The HAA results from the USS process HAA and the HAA derived from the kinematic model are fused with HAA results from multiple sensor modalities, including at least visual, radar, and UWB modalities; and based on kinematic models considered to be of low quality, HAA results from the USS process HAA and related quality metrics are fused with HAA results from multiple sensor modalities and related quality metrics; and based on kinematic models considered to be of high quality, HAA results from the USS process HAA and related quality metrics are fused with HAA results from multiple sensor modalities and related quality metrics, as well as HAA results derived from the kinematic model and related quality metrics.

[0014] In yet another exemplary embodiment, a method is provided for estimating the hinge angle (HAA) of a mounting device using an ultrasonic sensor (USS) while using likelihood filtering to ensure the quality of the detected echo signal performance. The method includes: in response to attaching a trailer to a vehicle, generating at least one set of ultrasonic sensor (USS) data based on detecting at least one set of echo signals generated by at least one set of USSs already directed at the trailer and configured around the vehicle; in response to detecting the at least one set of echo signals, filtering based on the at least one set of USS data by evaluating the quality of the echo signals and the consistency with at least one set of driving conditions including straight-ahead driving and steering inputs with the echo signals; in response to filtering the at least one set of echo signals, estimating based on the at least one set of USS data by performing calculations during the likelihood filtering process using a set of geometric equations selected from one of a plurality of different sets of geometric equations to determine at least one arbitrary frontal shape of the trailer; classifying the estimated arbitrary frontal shape of the trailer based at least on a comparison of measured distances from the detected echo signals about the trailer to the determined frontal shape of the trailer; and generating at least one comparison with a kinematic model based on the estimation results of the at least one set of USS data to ensure that the results of the kinematic model for the HAA associated with the determined frontal shape of the trailer are considered to have high signal performance, wherein the high signal performance is based on the correlation with the kinematic model at low speeds.

[0015] In at least one exemplary embodiment, the method includes using likelihood filtering to determine a plurality of driving conditions, including at least straight driving and steering angle consistency, and the frontal shape of the trailer.

[0016] In at least one exemplary embodiment, the method includes removing USS data that is inconsistent with a plurality of driving conditions, the plurality of driving conditions including at least straight driving at low speeds and steering angle input; and classifying the trailer frontal shape to select a geometric equation for determining the HAA.

[0017] In at least one exemplary embodiment, the method includes using a set of measurements including steering wheel angle (SWA) or road wheel angle (RWA) to determine a range of likelihoods for acceptable distance measurements of USS echoes reflected and collected from the front of the trailer during a straight-line driving state; and establishing a range of likelihoods based on the SWA or RWA measurements, wherein USS echoes falling outside the acceptable distance measurement range are removed.

[0018] In at least one exemplary embodiment, the method includes optimizing the use of USS data by comparing HAA results estimated from at least one set of USS data with HAA results derived from a kinematic model at low speeds, which can provide a quality metric for the implementation of USS HAA calculations with associated, determined trailer frontal shapes.

[0019] In at least one exemplary embodiment, the method includes comparing the HAA derived using a kinematic model at low speed with the USS process HAA to determine a quality metric for the USS based on a mean squared error calculation with sufficient data points using an acceptable threshold.

[0020] In at least one exemplary embodiment, the method includes comparing HAA results derived from a kinematic model with a USS process HAA under the following conditions to determine a quality index of the kinematic model, wherein the kinematic model is derived based on at least a set of assumptions that trailer dynamics are considered invalid, including dynamics at higher speeds, at higher speeds with steering input, on rough road surfaces above a threshold acting on the vehicle and trailer, and on crosswinds above a threshold acting on the vehicle and trailer.

[0021] In at least one exemplary embodiment, the method includes fusing HAA results from USS process HAA and HAA derived from kinematic models with HAA results from multiple sensor modalities, using a set of values ​​already calculated for HAA results and quality metrics already determined, wherein the multiple sensor modalities include at least visual, radar and UWB modalities.

[0022] In at least one exemplary embodiment, the method includes fusing HAA results from USS process HAA and related quality metrics with HAA results from multiple sensor modes and related quality metrics, based on a kinematic model considered to be of low quality.

[0023] In at least one exemplary embodiment, the method includes fusing HAA results from USS process HAA and related quality metrics with HAA results from multiple sensor modes and related quality metrics, as well as HAA and related quality metrics derived from the kinematic model, based on a kinematic model considered to be of high quality.

[0024] In at least one exemplary embodiment, the method includes implementing a kinematic model for evaluating the quality of USS process HAA results from USS measurements at lower speeds, wherein the USS process HAA results are used to evaluate the quality of the kinematic model under a set of trailer dynamic conditions, which includes at least the dynamics at higher speeds, the dynamics at higher speeds with steering input, the dynamics of rough road conditions above a threshold acting on the vehicle and trailer, and the dynamics of crosswind conditions above a threshold acting on the vehicle and trailer.

[0025] In yet another exemplary embodiment, a vehicle apparatus is provided, including an estimated hitch articulation angle (HAA) based on an ultrasonic sensor (USS) estimation unit, the ultrasonic sensor estimation unit including one or more processors and a non-transient computer-readable medium encoded with programming instructions. The USS estimation unit is configured to: generate at least one set of ultrasonic sensor (USS) data based on detecting at least one set of echo signals generated by a plurality of USS devices configured to be directed toward a trailer, wherein the trailer is coupled to the vehicle; in response to the detection of the at least one set of echo signals, filter the at least one set of USS data based on a likelihood filtering assessment of the quality of the echo signals and consistency with driving conditions including straight-line driving and steering inputs that include the echo signals; and in response to the filtering of the at least one set of echo signals, perform a likelihood filtering process... The computation is performed using a set of geometric equations selected from one of multiple sets of different geometric equations, estimation is performed based on at least one set of USS data to determine at least an arbitrary frontal shape of the trailer; the estimated arbitrary frontal shape of the trailer is classified based at least on a comparison of measured distances from detected echo signals about the trailer to the determined frontal shape of the trailer; and at least one comparison is generated based on the estimation results of at least one set of USS data with a kinematic model to ensure that the kinematic model is considered to have higher signal performance for the HAA results associated with the determined frontal shape of the trailer, wherein the higher signal performance is based on the correlation with the kinematic model at low speeds.

[0026] In at least one exemplary embodiment, the vehicle device includes a USS estimation unit configured to implement a kinematic model for evaluating the quality of a USS process HAA result from USS measurements at lower speeds, wherein the USS process HAA result is used to evaluate the quality of a kinematic model under a set of trailer dynamic conditions, which includes at least the dynamics of higher speeds, higher speeds with steering input, rough road conditions above a threshold acting on the vehicle and trailer, and crosswind conditions above a threshold acting on the vehicle and trailer. Attached Figure Description

[0027] Exemplary embodiments will now be described in conjunction with the following accompanying drawings, wherein the same reference numerals denote the same elements, and wherein:

[0028] Figure 1 A block diagram depicting an exemplary vehicle according to an exemplary embodiment is shown. The exemplary vehicle may include a processor that calculates the hitch articulation angle (HAA) by detecting the rear ultrasonic sensor (USS) echo of any trailer of arbitrary shape using echo switching and a likelihood filter for distance modeling, steering angle consistency, and trailer front shape determination.

[0029] Figure 2 An example diagram is shown illustrating a method for accurately estimating the HAA of an arbitrary-shaped trailer by using an ultrasonic sensor (USS) system in conjunction with a kinematic model to estimate the hitch articulation angle (HAA).

[0030] Figure 3 An example diagram is shown illustrating the quality assessment of the kinematic model prior to fusion by comparing it with an estimated hook-up angle (HAA) using an ultrasonic sensor (USS) system, according to an exemplary embodiment.

[0031] Figure 4 An example diagram is shown according to an exemplary embodiment, which provides the fusion of USS data with kinematic models and / or other sensors and their respective quality assessments for estimating the hanger articulation angle (HAA) using an ultrasonic sensor (USS) system;

[0032] Figure 5 An example diagram is shown of a vehicle and trailer arrangement according to an exemplary embodiment and of the hitch hinge angle (HAA) detected by an ultrasonic sensor (USS) through the estimation of the hitch hinge angle (HAA) using the ultrasonic sensor (USS) system.

[0033] Figure 6An example diagram is shown illustrating the geometric angle of the mount hinge angle (HAA) detected by an ultrasonic sensor (USS) using an ultrasonic sensor (USS) system to estimate the mount hinge angle (HAA) according to an exemplary embodiment; and

[0034] Figure 7 An exemplary flowchart is shown of a process for using an ultrasonic sensor (USS) system to process USS data by estimating the hinge angle (HAA) of the attachment device, and then proceeding through likelihood logic, according to an exemplary embodiment. Detailed Implementation

[0035] The following detailed description is merely exemplary in nature and is not intended to limit application or use. Furthermore, it is not intended to be bound by any express or implied theory presented in the preceding technical field, background, overview, or the following detailed description.

[0036] As used herein, the term "module" means any hardware, software, firmware, electronic control components, processing logic and / or processor device, individually or in any combination, including but not limited to: application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), electronic circuits, processors (shared, dedicated or grouped) and memory, combinational logic circuits and / or other suitable components that execute one or more software or firmware programs to provide the said functionality.

[0037] Embodiments of this disclosure can be described herein in terms of functional and / or logical block components and various processing steps. It should be understood that such block components can be implemented by any number of hardware, software, and / or firmware components configured to perform specified functions. For example, embodiments of this disclosure can employ various integrated circuit components, such as memory elements, digital signal processing elements, logic elements, lookup tables, etc., which can perform various functions under the control of one or more microprocessors or other control devices. Furthermore, those skilled in the art will understand that embodiments of this disclosure can be practiced in conjunction with any number of systems, and the systems described herein are merely exemplary embodiments of this disclosure.

[0038] Autonomous and semi-autonomous vehicles are capable of perceiving their environment and navigating based on that perception. These vehicles utilize various types of sensing devices to sense their environment, such as optical cameras, radar, lidar, and other image sensors. In such vehicles, the sensed data can be fused with map data and vehicle sensors (inertial measurement units, vehicle speed sensors, etc.) to identify and track vehicle trajectory performance based on road geometry, and this disclosure is applicable to sensing data for kinematic modeling and for estimating the hitch articulation angle (HAA) to enhance traction stability when towing trailers of different shapes.

[0039] In various exemplary embodiments, this disclosure describes systems, methods, and apparatuses that associate a "kinematic model" with a trailer. The kinematic model describes the relationship between the position, orientation, and motion of the vehicle and the trailer.

[0040] In various exemplary embodiments, this disclosure describes systems, methods, and apparatuses for estimating the hitch articulation angle (HAA) of a trailer using an ultrasonic sensor (USS) and a kinematic model, comparing it with the kinematic model and other sensor data to at least assess the quality of the kinematic model, and then fusing it with other sensor data and the kinematic model to estimate the hitch articulation angle (HAA) of the front of an arbitrary-shaped trailer during various towing dynamic events in a traction operation.

[0041] In various exemplary embodiments, this disclosure describes systems, methods, and apparatuses for accurately estimating HAA using a USS for arbitrary-shaped trailers by employing enhanced raw sensor data, likelihood filtering, robust sensor pair selection, quality metric evaluation of HAA using a USS based on comparison with a kinematic model during low-dynamic towing events, quality metric evaluation of a kinematic model based on comparison of the USS with HAA during high-dynamic towing events, and then fusing the estimated HAA using the USS with the kinematic model (along with other sensor modal) signals using a quality metric.

[0042] In various exemplary embodiments, this disclosure describes systems, methods, and apparatuses for improving raw USS data filtering by omitting "implausible" data points and by configuring a trailer motion model to further enhance the acceptance criteria of USS measurements. The method includes evaluating the performance of sensor pairs to select a "winning" pair and HAA angle quality, then using USS signals during dynamic maneuvers to evaluate the quality of the kinematic model to establish a quality indicator for kinematic modeling, followed by fusion. In various exemplary embodiments, this disclosure describes a method for estimating the hitch articulation angle (HAA) using ultrasonic sensors (USS), wherein USS data is processed and then passed through likelihood logic to (1) detect the true trailer echo from unwanted reflections, (2) detect the trailer shape, and (3) perform an initial evaluation of which sensor pairs provide better information in order to calculate the HAA from their echoes. The method includes calculating the geometric equations of the raw HAA, wherein each raw HAA from a sensor pair is compared to a kinematic model at low speeds, and the pair with higher performance (i.e., greater correlation with the kinematic model) is selected.

[0043] Figure 1A block diagram depicting an exemplary vehicle according to one embodiment is shown. This exemplary vehicle may include a processor that calculates the HAA using rear USS echoes from any arbitrary-shaped trailer using echo switching and likelihood filters for straight-ahead driving, steering angle consistency, and trailer frontal shape determination. The processor may use a mean squared error threshold to select the “winning” USS sensor pair for any trailer frontal shape. The processor designs a quality metric for the USS using a low-speed comparison between the kinematic model and the USS method, with the mean squared error having sufficient data points based on an acceptance threshold. The processor designs a quality metric for the kinematic model using a comparison between the kinematic model and the USS method in cases where the kinematic model assumptions may not hold (e.g., higher speeds or rough road conditions in trailer dynamics). The processor also fuses the USS and kinematic model using methods known in the industry, such as Kalman filtering (and other sensor modalities, i.e., vision, radar, ultra-wideband (UWB)...), using the calculated HAA values ​​(i.e., USS, vision, radar, UWB...) and the designed quality metric.

[0044] like Figure 1 As shown, vehicle 10 typically includes a chassis 12, a body 14, front wheels 16, and rear wheels 18. The body 14 is mounted on the chassis 12 and substantially encloses the components of vehicle 10. The body 14 and chassis 12 may together form a frame. Each of the wheels 16-18 is rotatably connected to the chassis 12 near a corresponding corner of the body 14. In the illustrated embodiment, vehicle 10 is depicted as a passenger vehicle. However, it should be understood that any other vehicle may be used, including motorcycles, trucks, sports utility vehicles (SUVs), recreational vehicles (RVs), boats, aircraft, etc.

[0045] As shown in the figure, vehicle 10 typically 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. In this example, the propulsion system 20 may include an electric motor, such as a permanent magnet (PM) motor. The transmission system 22 is configured to transmit power from the propulsion system 20 to the wheels 16 and 18 according to a selectable speed ratio.

[0046] Braking system 26 is configured to provide braking torque to wheels 16 and 18. In various exemplary embodiments, braking system 26 may include friction brakes, brake-by-wire brakes, regenerative braking systems such as electric motors, and / or other suitable braking systems.

[0047] The steering system 24 affects the position of the wheels 16 and / or 18. Although depicted as including a steering wheel 25 for illustrative purposes, in some exemplary embodiments contemplated within the scope of this disclosure, the steering system 24 may not include a steering wheel.

[0048] The sensor system 28 includes one or more sensing devices 40a-40v that sense observable conditions of the external and / or internal environment of the vehicle 10 and generate associated sensor data. The sensor system may include, but is not limited to, rear ultrasonic sensors, rear vision sensors, rear radar sensors, vehicle speed sensors, and inertial measurement sensors.

[0049] The actuator system 30 includes one or more actuator devices 42a-42n that control one or more vehicle features, such as, but not limited to, the propulsion system 20, the transmission system 22, the steering system 24, and the braking system 26. In various exemplary embodiments, the vehicle 10 may also include… Figure 1 Interior and / or exterior vehicle features not shown, such as various doors, trunk and cabin features, such as air, music, lighting, touchscreen display components, etc.

[0050] Data storage device 32 stores data used to control vehicle 10. Data storage device 32 may be part of controller 34, separate from controller 34, or part of controller 34 and separate system.

[0051] The controller 34 includes at least one processor 44 (integrated with or connected to the system 100) and a computer-readable storage device or medium 46. The processor 44 can be any custom or commercially available processor, central processing unit (CPU), graphics processing unit (GPU), application-specific integrated circuit (ASIC) (e.g., a custom ASIC implementing a neural network), field-programmable gate array (FPGA), an auxiliary processor among several processors associated with the controller 34, a semiconductor-based microprocessor (in the form of a microchip or chipset), any combination thereof, or any device typically used for executing instructions. The computer-readable storage device or medium 46 can include, for example, volatile and non-volatile memory in read-only memory (ROM), random access memory (RAM), and keep-alive memory (KAM). KAM is persistent or non-volatile memory used to store various operational variables when the processor 44 is powered off. The computer-readable storage device or medium 46 may be implemented using any of several known storage devices, such as PROM (programmable read-only memory), EPROM (electrically programmable read-only memory), EEPROM (electrically erasable programmable read-only memory), flash memory, or any other electrical, magnetic, optical, or combined storage device capable of storing data, some of which represent executable instructions used by the controller 34 when controlling the vehicle 10.

[0052] The instructions may include one or more separate programs, each of which includes an ordered list of executable instructions for implementing logical functions. When executed by processor 44, the instructions receive and process signals (e.g., sensor data) from sensor system 28, execute logic, calculations, methods, and / or algorithms for automatically controlling components of vehicle 10, and generate control signals transmitted to actuator system 30 to automatically control components of vehicle 10 based on logic, calculations, methods, and / or algorithms. Although in Figure 1 Only one controller 34 is shown, but embodiments of vehicle 10 may include any number of controllers 34 that communicate via any suitable communication medium or combination of communication media and cooperate to process sensor signals, execute logic, calculations, methods and / or algorithms, and generate control signals to automatically control the features of vehicle 10.

[0053] For example, system 100 may include any number of additional submodules embedded in controller 34, which may be combined and / or further subdivided to similarly implement the systems and methods described herein. Additionally, inputs to system 100 may be received from sensor system 28, from other control modules (not shown) associated with vehicle 10, and / or by… Figure 1 Other sub-modules (not shown) within the controller 34 are determined / modeled. Furthermore, the input can be preprocessed, such as subsampling, noise reduction, normalization, feature extraction, and missing data reduction.

[0054] Figure 2 An example diagram is shown illustrating the accurate estimation of the HAA of an arbitrary-shaped trailer by estimating the hook-up articulation angle (HAA) using an ultrasonic sensor (USS) system according to an exemplary embodiment.

[0055] Figure 2 Including system 100 for filtering system ( Figure 1 The processing and filtering system includes a set of raw echo measurement data from ultrasonic sensors. (For example, i = 1, 2, 3, 4, representing the number of available USS), used for signal preprocessing and echo switching based on echo quality 205. The preprocessed aggregate echo signal or driving data is then processed by a set of likelihood filters 210 for robust direction (e.g., straight-line driving estimation) and acoustic filtering. For example, using straight-line driving d i straight A set of likelihood filters 215 is used for the driving data, partly to establish a "baseline" for the acceptable range of the measured distance and filtering out inconsistent time-varying measurement data. Once the output of the straight-line driving 215 is known, the likelihood filters further refine the measurement points by combining steering wheel angle (SWA) or road wheel angle (RWA). The acceptance criteria. Given the known straight-line travel distance d of any single sensor i. i straight A clockwise increase in the SWA input (box 220) will result in a corresponding increase or decrease in the actual physical distance between sensor i and the trailer front. Sensor data that does not conform to this correspondence will be discarded. This knowledge further limits the range of reflections that actually originate from the trailer surface and allows unwanted reflections, such as those from the hook-up points, to be filtered out.

[0056] Straight-line driving data based on different sensors d i straight The differences between them are used to make an initial assessment of the shape of the trailer's front. At 225, a set of straight-line driving data di is compared. The likelihood of the trailer shape is determined by whether the value is less than a threshold. If so, a set of plane geometry equations is used at 230 to estimate the hitch articulation angle (HAA). If not, another determination is performed at 235 to determine a set of straight-driving data d. i Sensor data on the same side of the vehicle Is it greater than a threshold (for the case of 4 symmetrically placed ultrasonic sensors)? If the result is yes, then a set of geometric equations for an irregular trailer shape is used at 240 to estimate the hitch articulation angle (HAA). If not, then it is assumed that the trailer is a V-Nose shape, and at 245, a set of V-Nose geometric equations is used for the likelihood of the trailer shape.

[0057] Once the trailer shape is determined and unwanted USS data points are filtered out, the determined set of geometric equations 250 is used to compare the reported distance d of sensors i and j. i filtered Filtering and d j straight (Calculate the hinge angle (HAA) φ of the mounting device using a straight line) i，j .

[0058] Once the trailer shape is determined and unwanted USS data points are filtered out, the determined set of geometric equations 250 is used to compare the reported distance d from sensors i and j. i filtered and d j straight To calculate the hinge angle (HAA) φ of the mounting device i，j The geometric equations require knowing the position of the USS sensor on the vehicle relative to the attachment point.

[0059] Depending on the determined set of geometric equations and the acceptable USS distance measurement d i filteredAvailability, at any time k At 255, a maximum of 6 calculated HAAs can be obtained (for the case of 4 ultrasonic sensors), as can be obtained from vector... shown.

[0060] At 265, this processing is performed by comparing the sensor estimation results with an independently computed low-speed motion model φ kin The results were compared and optimized using USS data. Kinematic model φ kin It was implemented as a background application, and the discovered error term was defined as the kinematic model φ that had been collected. kin The difference between the equations and the USS equations. At 260, based on the kinematic model φ... kin Consistency, processing by each sensor for output φ i，j The quality of the estimates provided by (or geometric equations) is used to identify sensor pairs that perform better in their HAA estimates. It is known that sensors on the vehicle side may perform better in providing HAA estimates when the trailer is physically closer to these sensors; therefore, the φ of the estimates can be defined. i，j Relative to φ kin Different performance bands. As an example, for φ calculated by comparing the sensors on the right side of the vehicle. i，j The HAA performance band [-20°, -10°] shows better estimated performance, while the same sensor may have reduced performance in the opposite band [+10°, +20°]. The φ value for each different performance band... kin and φ i，j The mean square error (MSE) between the pairs can indicate the "winning" pair in that band.

[0061] Once the winning pair is selected under low-speed conditions, at any time t k The average of the available winning pair estimates for any performance band is output as the HAA estimate φ via the USS process 260. USS The winning choice leads to a difference in time t. k Simply average All available φ i，j More refined HAA estimation output. Once the process of identifying winning pairs in different estimation bands is complete, the HAA estimation output is then processed at 270. To generate "holistic" USS process performance and quality metrics.

[0062] Figure 3 An exemplary schematic diagram illustrates a quantitative accuracy assessment of the kinematic model prior to fusion by estimating the hinge angle (HAA) of the attachment device using an ultrasonic sensor (USS) system, according to an exemplary embodiment. Figure 3In the middle, it is used to execute the kinematic model φ kin The input for quality assessment 305 can include various types of USS radar Sensor inputs, and camera The signal input is used for calculations by the detection module 310 based on swaying, sideslip, rough road, or wind inputs. The moving average value 315 of the sensor input is based on the kinematic model φ. kin The estimated value of 325 is calculated independently. Calculating the deviation 320 between these two independent modalities of the HAA allows for the detection of kinematic model failures, including sideslip, sway, rough road conditions, and wind effects exceeding acceptable thresholds for trailers. If the detection of sideslip, sway, rough road conditions, and wind effects is determined to be above acceptable thresholds, the quality assessment 330 of the kinematic model (1,0) is determined.

[0063] In an exemplary embodiment, if the determined speed is higher than an acceptable threshold for trailer sway based on a trailer sway detection standard, the following evaluation is performed: The estimated hitch articulation angle (HAA) using ultrasonic sensor (USS) values ​​is evaluated using (i.e., an exponential moving average or a bandpass filter for the frequency range of interest for trailer sway, i.e., 0.5 to 4 Hz) HAA signals from the USS (and / or other sensor modalities, such as vision, radar, UWB, etc.) to the kinematic model. If the difference between the kinematic model and (i.e., the USS) sensor values ​​is greater than an acceptable threshold, the likelihood of a sway-induced effect is high, and the kinematic model quality assessment at 330° is set to 0.

[0064] In another exemplary embodiment, if the speed and steering inputs exceed an acceptable threshold for trailer sideslip detection based on a trailer sideslip detection criterion, the following evaluation is performed: The filter sensor is evaluated using (i.e., exponential moving average) HAA signals from the USS (and / or other sensor modalities, such as vision, radar, UWB, etc.) to the kinematic model. If the difference between the kinematic model and (i.e., the USS) sensor values ​​is greater than the threshold, the likelihood of a sideslip effect is high, and the result of the kinematic model quality evaluation at 330 is set to 0.

[0065] In another exemplary embodiment, if the rough road wheel speed sensor algorithm generates results higher than an acceptable threshold for rough road detection based on rough road detection criteria, the following evaluation is performed: the filter sensor is evaluated using a statistical measure (i.e., standard deviation) of the error between the HAA signal from the USS (and / or other sensor modalities, such as vision, radar, UWB…) and the kinematic model. If the statistical measure is greater than the acceptable threshold, the rough road input is highly likely to have an impact, and the result of the kinematic model quality evaluation at 330 is set to 0.

[0066] In another exemplary embodiment, if the speed is above a threshold and crosswind is detected above an acceptable threshold, the following evaluation is performed: the filter sensor is evaluated using (i.e., exponential moving average) the HAA signal from the USS (and / or other sensor modalities, such as vision, radar, UWB…) to the kinematic model. If the difference between the kinematic model and (i.e., the USS) sensor value is greater than an acceptable threshold, the wind input effect is highly probable, and the result of the kinematic model quality evaluation at 330 is set to 0.

[0067] A technical example for detecting environmental crosswind conditions in a vehicle is described in U.S. Patent No. 10,679,436, entitled “Vehicle suspension system alignment monitoring”, published on June 9, 2020.

[0068] Figure 4 An exemplary schematic diagram is shown, according to an exemplary embodiment, of providing the fusion of USS data with a kinematic model and / or other sensors by estimating the hinge angle (HAA) of the attachment device using an ultrasonic sensor (USS) system. Figure 4 In one exemplary embodiment, a kinematic model is implemented to evaluate the quality of the HAA from a lower-velocity USS in a first case, and in a second case, although sensing data about the HAA from the USS is used to evaluate the quality of the kinematic model under dynamic conditions. Figure 3 The set of acceptable thresholds described herein is used to evaluate the quality of motion models based on individual operating conditions or a logically separate set of conditions.

[0069] like Figure 4 As shown, the HAA from USS (along with other sensor modes, such as USS 415) Vision (camera) 425 Radar 435 UWB (not shown), ...) and related quality indicators and kinematic models Integration. For example, in Figure 4The diagram shows a set of multiplexed inputs, including a kinematic model-based HAA and a quality index 410. Based on USS's HAA and Quality Indicator 420 Vision-based HAA and quality metrics 430. Radar-based HAA and quality indicators 440 And additional sensor modalities based on HAA and quality index 450, which are combined or fused using various techniques (such as Kalman filtering, Bayesian filtering, deep learning (i.e., CNN, RNN, etc.), machine learning, etc.), for fusion sensor processing at HAA signal fusion 460.

[0070] Based on the quality of the kinematic model, HAA signal fusion is performed under two relevant operational conditions 460. In the first case, if (or when) the kinematic model If the HAA from the USS is considered low quality, it can be fused with other sensor modalities (such as vision, radar, UWB, etc., if available) and relevant quality metrics. In this case, the fused signal can simply be reported as a moving average. In a supplementary second case, fusion occurs here based on a high-quality kinematic model: if the kinematic model is considered high quality, it can be fused with the HAA from the USS and other sensor modalities (such as vision, radar, UWB, etc., if available) and relevant quality metrics.

[0071] Figure 5 An example diagram is shown illustrating the estimation of the hitch hinge angle (HAA) by using an ultrasonic sensor (USS) system, according to an exemplary embodiment. The distance l is based on the vehicle wheelbase. w 515. The distance l between the rear axle of the vehicle and the engagement point. h 520. The distance l between the attachment point and the rear axle of the trailer. tr The HAA510 "φ" detected in the vehicle trailer 500 is calculated using a set of measurements of 525 and the vehicle's wheel angle δ535. The formula for calculating 510 "φ" is as follows:

[0072]

[0073] Where V c 530 is the longitudinal velocity of the vehicle, δ535 is the wheel angle determined based on the vehicle driver's steering angle input, and the kinematic model of the trailer is obtained by measuring the vehicle-trailer 500 over time t. w , l tr , l hThis is derived for calculating HAA (i.e., 510 "φ").

[0074] Figure 6 An exemplary schematic diagram of the geometry of a V-Nose trailer according to an exemplary embodiment is shown, the trailer having a frontal angle determined by estimating the hitch articulation angle (HAA) using an ultrasonic sensor (USS) system (HAA system).

[0075] like Figure 6 As shown, examples of geometric equations are provided below:

[0076]

[0077] Where d1 is the value measured by sensor 1. Distance, d4 is measured by sensor 4. Distance, L hitch L1 is the distance between the vehicle bumper and the attachment point, L2 is the distance between the midpoint of the vehicle bumper and sensor 1, L4 is the distance between the midpoint of the vehicle bumper and sensor 4, and α is the frontal angle of the trailer. This equation is solved numerically to obtain the value of the unknown HAA, denoted by φ.

[0078] Figure 7 An exemplary flowchart is shown illustrating a process of using an ultrasonic sensor (USS) system (HAA system) to perform USS data processing by estimating the hinge angle (HAA) of the attachment device according to an exemplary embodiment, and then processing it through likelihood logic.

[0079] Initially, USS data is processed, and then, through likelihood logic, real trailer echoes are detected from unwanted reflections, trailer shape is detected, the sensor pair geometry equations are initially evaluated, and raw HAAs are calculated. Each raw HAA from a sensor pair is compared with a kinematic model during low speeds, and those pairs with higher performance are selected.

[0080] At task 705, in response to the engagement (or engagement status, or other similar action) between the trailer and the vehicle, the system is executed or triggered, which uses an ultrasonic sensor (USS) to estimate the hitch articulation angle (HAA) while using likelihood filtering to ensure the quality of the detected echo signal performance.

[0081] At mission 710, a set of ultrasonic sensors (USS) generates an echo signal that provides USS data to the HAA system based on the detection of echo signals already pointing towards the trailer. The USS data is processed, noise is filtered through likelihood logic, and the real trailer echo is identified from unwanted reflections using data collected during straight-line travel. The noise filtering is further enhanced by incorporating steering wheel angle (SWA).

[0082] At task 715, the likelihood filter uses data collected during straight-line driving to classify trailer front-end shapes for any given trailer front-end shape. The likelihood filter logic compares the actual echo signals from straight-line driving with each other and determines the trailer shape classification.

[0083] At mission 720, based on the shape classification of trailers for arbitrary trailer front-end shapes from mission 715, a likelihood filter is used to initially evaluate which sensor pair combinations and corresponding geometric equations to select, using data collected during straight-line driving, in order to calculate the HAA.

[0084] At task 725, the original HAA set from each sensor pair is compared with the kinematic model at low speed. This is to ensure that the selected sensor pairs are those exhibiting higher performance quality. The comparison is based on a set of USS data estimates φ from each selected sensor pair. i，j , and the kinematic model φ at low speed kin Compare. φ kin and φ i，j The mean square error (MSE) between φ is used to measure the mean square error between φ. kin and φ i，j The "winning" pair is selected by comparing the pairs.

[0085] At task 730, the result is determined by combining the outcomes from the "winning" pair. The use of USS data has been further optimized. The combination of results from the winning pair can be determined using multiple methods, including arithmetic averaging. The estimated hinge angle of the hook is then... With kinematic model φ kin The estimation results at lower speeds are compared to provide a quality metric for the implementation process used in the USS HAA calculation, which has an associated, defined trailer shape. Low-speed comparisons between the kinematic model and the USS process to determine USS quality metrics can be performed using various different types of algorithms to identify quality. For example, mean squared error based on an acceptable threshold and sufficient data can be used for calculation.

[0086] At task 735, in cases where the kinematic model is derived based on assumptions that trailer dynamics (e.g., vehicle-trailer speed exceeding a certain threshold or vehicle speed and steering input exceeding a certain threshold) are considered invalid, a comparison is made between the kinematic model and the USS process to determine the quality index of the kinematic model. When the comparison exceeds the threshold, trailer swaying or sideslip may occur, and the quality index of the kinematic model is evaluated as low quality and given a value of 0.

[0087] At task 740, if the kinematic model is derived based on assumptions deemed invalid, a comparison is made between the kinematic model and the USS process, at least based on rough road dynamics or crosswind conditions, to determine the quality index of the kinematic model. When this comparison exceeds a threshold, trailer motion due to crosswinds or rough roads may exist, and the quality index of the kinematic model is evaluated as low quality and given a value of 0.

[0088] At mission 750, when the kinematic model is considered high quality, the fusion of HAA results from USS data can occur across multiple sensor modalities and kinematic models. The HAA results are determined by the fusion of USS data, multi-sensor modal data, kinematic models, and relevant quality metrics. Multi-sensor modalities can include vision, radar, and UWB modalities.

[0089] At mission 755, the fusion of HAA results from USS data can occur across multiple sensor modalities, rather than based on a kinematic model considered to be of low quality. Here, again, the HAA results are determined by the fusion of USS data, multi-sensor modal data, and relevant quality metrics. Multi-sensor modalities can include vision, radar, and UWB modalities.

[0090] At task 760, the final output based on the fused HAA angle is provided.

[0091] As briefly mentioned, the various modules and systems described above can be implemented in an HAA system for estimation steps, kinematic modeling, and likelihood processing using one or more machine learning models that undergo supervised, unsupervised, semi-supervised, or reinforcement learning. Such models can be trained to perform tasks such as classification (e.g., binary or multi-class classification), regression, clustering, dimensionality reduction, and / or similar tasks. Examples of such models include, but are not limited to, artificial neural networks (ANNs) (e.g., recurrent neural networks (RNNs) and convolutional neural networks (CNNs)), decision tree models (e.g., classification and regression trees (CART)), ensemble learning models (e.g., boosting, bootstrapped aggregation, gradient boosting machines, and random forests), Bayesian network models (e.g., Naive Bayes), principal component analysis (PCA), support vector machines (SVMs), clustering models (e.g., K-nearest-neighbor, K-means, expectation maximization, hierarchical clustering, etc.), and linear discriminant analysis models.

[0092] It should be understood that Figure 1-7 The process can include any number of additional or alternative tasks. Figure 1-7 The tasks shown do not need to be performed in the order illustrated, and Figure 1-7 This process can be integrated into a more comprehensive program or process with additional functionality not described in detail here. Furthermore, Figure 1-7 One or more tasks shown in the diagram can be obtained from Figure 1-7 The process shown in the embodiments is omitted, as long as the intended overall functionality remains intact.

[0093] The foregoing detailed description is merely illustrative in nature and is not intended to limit the embodiments of the subject matter or the application and use of such embodiments. As used herein, the word “exemplary” means “serving as an example, instance, or illustration.” Any implementation described herein as exemplary is not necessarily to be construed as being more preferred or advantageous than other implementations. Furthermore, no one is intended to be bound by any express or implied theory presented in the foregoing technical field, background, or detailed description.

[0094] While at least one exemplary embodiment has been given in the foregoing detailed description, it should be understood that numerous variations exist. It should also be understood that the one or more exemplary embodiments are merely examples and are not intended to limit the scope, applicability, or configuration of this disclosure in any way. Rather, the foregoing detailed description will provide those skilled in the art with a convenient roadmap for implementing one or more exemplary embodiments.

[0095] It should be understood that various changes may be made to the function and arrangement of the elements without departing from the scope of this disclosure as set forth in this application and its legal equivalents.

Claims

1. A system for estimating the hinge angle of a mounting device, comprising: A processing unit disposed in a vehicle, the processing unit comprising one or more processors configured by programming instructions encoded on a non-transient computer-readable medium, the processing unit being configured to: At least one set of ultrasonic sensor data is generated based on detecting at least one set of echo signals generated by a plurality of ultrasonic sensors (USS), the plurality of USS being configured to surround a vehicle that has been pointed at a trailer, to determine the hitch articulation angle (HAA) where the trailer is attached to the vehicle. In response to the detection of at least one set of echo signals, filtering is performed based on at least one set of USS data, which is evaluated based on the echo signal quality, consistency with driving conditions including straight-line driving, and the likelihood of the echo signal steering input. In response to filtering of at least one set of echo signals, estimation is performed based on at least one set of USS data by performing calculations using a set of geometric equations selected from one of multiple sets of different geometric equations during the likelihood filtering process to determine at least one frontal shape of the trailer. The estimated arbitrary frontal shape of the trailer is classified based at least on the comparison of the measured distance from the detected echo signal about the trailer to the determined frontal shape of the trailer. and At least one comparison is generated based on the estimation results of at least one set of USS data with the kinematic model to ensure that the kinematic model is considered to have higher signal performance for the HAA results associated with the determined trailer frontal shape, wherein the higher signal performance is based on the correlation with the kinematic model at low speeds.

2. The system according to claim 1, further comprising: The processing unit is configured to: Likelihood filtering is used for multiple driving conditions, including at least straight driving and consistent steering angle, as well as for determining the frontal shape of the trailer.

3. The system according to claim 2, further comprising: The processing unit is configured to: Remove USS data that is inconsistent with multiple driving conditions, including at least straight driving at low speeds and steering angle input; and The trailer frontal shape is classified to select the geometric equations used to determine the HAA.

4. The system according to claim 3, further comprising: The processing unit is configured to: A set of measurements, including steering wheel angle (SWA) or road wheel angle (RWA), is used to determine the range of likelihood values ​​for acceptable distance measurements of reflected and collected USS echoes from the front of the trailer during straight-line driving. and Based on SWA or RWA measurements, establish a range of likelihoods to remove USS echoes that fall outside the acceptable distance measurement range.

5. The system according to claim 4, further comprising: The processing unit is configured to: The use of USS data is optimized by comparing HAA results estimated from at least one set of USS data with HAA results derived from a kinematic model at low speeds, which provides a quality metric for the implementation of USS HAA calculations with associated, defined trailer frontal shapes.

6. The system according to claim 5, further comprising: The processing unit is configured to: At low speeds, a comparison is made between the HAA results derived from the kinematic model and the HAA of the USS process to determine the quality metric of the USS based on mean squared error calculation with sufficient data points using an acceptable threshold.

7. The system according to claim 6, further comprising: The processing unit is configured to: The quality index of the kinematic model is determined by comparing the HAA derived from the kinematic model with the USS process HAA under the following conditions, where the kinematic model is derived based on at least a set of assumptions that trailer dynamics are considered invalid, including dynamics at high speeds, rough road conditions acting on the vehicle and trailer above a threshold, and crosswind conditions acting on the vehicle and trailer above a threshold. Using a set of values ​​already calculated for the HAA results and established quality metrics, the HAA results from the USS process HAA and the HAA derived from the kinematic model are fused with the HAA results from multiple sensor modalities, which include at least vision, radar and UWB modalities. and Based on a kinematic model considered to be of low quality, HAA results from USS process HAA and related quality metrics are fused with HAA results from multiple sensor modes and related quality metrics. and Based on what is considered a high-quality kinematic model, HAA results from the USS process HAA and related quality indices are fused with HAA results from multiple sensor modes and related quality indices, as well as HAA and related quality indices derived from the kinematic model.

8. The system of claim 7, wherein the higher speed is a higher speed with steering input.

9. A method for estimating the hinge angle of a mounting device, comprising estimating the hinge angle (HAA) of the mounting device using an ultrasonic sensor (USS) while using likelihood filtering to ensure the quality of the detected echo signal performance, the method comprising: In response to attaching the trailer to the vehicle, at least one set of ultrasonic sensor (USS) data is generated based on detecting at least one set of echo signals generated by multiple USS configured around the vehicle and pointing towards the trailer. In response to the detection of at least one set of echo signals, filtering is performed based on at least one set of USS data by evaluating the quality of the echo signals and the likelihood filtering of the consistency with at least multiple driving conditions including straight driving and the steering input of the echo signals. In response to filtering of at least one set of echo signals, estimation is performed based on at least one set of USS data by performing calculations using a set of geometric equations selected from one of a plurality of different sets of geometric equations during the likelihood filtering process to determine at least one frontal shape of the trailer. The estimated arbitrary frontal shape of the trailer is classified based at least on the comparison of the measured distance from the detected echo signal about the trailer to the determined frontal shape of the trailer. and At least one comparison is generated based on the estimation results of at least one set of USS data and the kinematic model to ensure that the kinematic model is considered to have high signal performance for the HAA results associated with the determined trailer frontal shape, wherein the high signal performance is based on the correlation with the kinematic model at low speeds.

10. The method of claim 9, further comprising: Likelihood filtering is used for multiple driving conditions, including at least straight driving and consistent steering angle, as well as for determining the frontal shape of the trailer.

11. The method of claim 10, further comprising: Remove USS data inconsistent with multiple driving conditions, including at least low-speed straight driving and steering angle input; and The trailer frontal shape is classified to select the geometric equations used to determine the HAA.

Images