Vehicle parking pose detection device and method
By arranging an elastic protective structure and a non-uniformly distributed pressure sensor array in the vehicle parking area, and combining air pressure and piezoelectric sensors, the problems of high cost and slow response in the prior art are solved, and low-cost, high-precision vehicle parking posture detection is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-03
- Publication Date
- 2026-03-10
AI Technical Summary
Existing vehicle parking pose detection technologies rely on high-precision imaging equipment and complex algorithms, which are costly, susceptible to environmental interference, and have slow response speeds, making it difficult to achieve low-cost and rapid high-precision detection.
Employing an elastic protective structure and a non-uniformly distributed pressure sensor array, combined with pneumatic and piezoelectric sensors, the vehicle's parking posture is determined by detecting instantaneous pressure changes and pressure data at the vehicle's contact edges.
It achieves low-cost, high-precision, and fast-response vehicle parking pose detection, reduces hardware costs, improves detection stability and response speed, and adapts to different vehicle models and environmental conditions.
Smart Images
Figure CN121632026A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of unmanned car washing machine detection, and in particular to a vehicle parking pose detection device and method. BACKGROUND
[0002] In an automatic car washing system, a vehicle needs to be parked in a car washing area according to specified requirements, so as to ensure that the washing robot arm, water spraying device, cleaning brush and other equipment can safely and efficiently complete the cleaning work. Therefore, the accuracy of vehicle parking directly affects the operation efficiency and safety of the car washing system: if the vehicle parking pose deviates too much, it may cause the robot arm to collide with the vehicle body, uneven distribution of cleaning liquid or failure of the liquid recovery system, thereby causing equipment damage or user complaints.
[0003] At present, a visual recognition system is usually used for vehicle parking pose detection, which collects vehicle images through a camera or a laser radar and identifies the vehicle contour and position by combining an image processing algorithm. This method relies on high-precision imaging equipment and complex algorithms, and is high in cost and susceptible to environmental light, rain and snow weather and other disturbances, and has a slow response speed. SUMMARY
[0004] Embodiments of the present application provide a vehicle parking pose detection device and method to achieve the effect of low cost and fast response.
[0005] In a first aspect, embodiments of the present application provide a vehicle parking pose detection device, comprising:
[0006] An elastic protection structure arranged at the edge position of the target parking area for detecting the instantaneous pressure change when the vehicle contacts the edge;
[0007] A plurality of measurement units arranged in a non-uniform distribution manner in the target parking area, wherein each measurement unit comprises a pressure sensor for collecting pressure data after the vehicle enters the target parking area;
[0008] A data processing module for determining the parking pose of the vehicle according to the instantaneous pressure change and the pressure data.
[0009] In a possible implementation, the elastic protection structure comprises: a middle-through rubber airbag arranged around the target parking area;
[0010] and a gas pressure sensor built in the rubber airbag.
[0011] In a possible implementation, the shape enclosed by the rubber airbag is the same as the shape of the target parking area.
[0012] In a possible implementation, the density of the measurement units arranged in the central region of the target parking area is greater than the density of the measurement units arranged in the edge region.
[0013] In a possible implementation, the pressure sensor in the measuring unit within the diagonal area of the target parking area is arranged at a preset inclination angle.
[0014] In a possible implementation, the sensor in the measuring unit includes an air pressure sensor and a piezoelectric sensor.
[0015] The proportion of air pressure sensors in the plurality of measuring units is greater than the proportion of piezoelectric sensors.
[0016] In a possible implementation, the measuring unit further includes:
[0017] The base has a groove.
[0018] The air bag body is placed in the groove, the air bag is filled with gas, and the air pressure sensor is built-in.
[0019] In a possible implementation, the measuring unit further includes:
[0020] The top plate covers the opening of the groove, and the air bag body is located in the space formed by the top plate and the groove.
[0021] The top plate is connected to the base through the guide rod, and when there is no object extruding the air bag body through the top plate, the air bag body lifts the top plate, and when there is an object extruding the air bag body through the top plate, the top plate extrudes the air bag body along the guide rod.
[0022] In a possible implementation, the device further includes a grid guide plate, and the plurality of measuring units are embedded in the grids of the grid guide plate.
[0023] In a possible implementation, the data processing module is specifically configured to:
[0024] In the case of detecting a transient pressure change, the pressure data collected by the plurality of measuring units is acquired.
[0025] Based on the pressure data, the parking contact surface of the tire of the vehicle is fitted.
[0026] According to the parking contact surface, the parking pose of the vehicle is determined.
[0027] In a possible implementation, based on the pressure data, the parking contact surface of the tire of the vehicle is fitted, including:
[0028] The pressure data is clustered to obtain pressure data corresponding to at least one tire of the vehicle.
[0029] According to the pressure data corresponding to the at least one tire, the parking contact surface of the at least one tire and the target parking area is fitted.
[0030] In a possible implementation, the parking pose of the vehicle is determined according to the parking contact surface, including:
[0031] The center point coordinates of the parking contact surface are extracted;
[0032] The parking direction of the vehicle is calculated according to the center point coordinates.
[0033] In a second aspect, the embodiments of the present application provide a vehicle parking pose detection method, including:
[0034] In the case of detecting the instantaneous pressure change, the pressure data collected by the plurality of measuring units is acquired;
[0035] The parking contact surface of the tire of the vehicle is fitted based on the pressure data;
[0036] The parking pose of the vehicle is determined according to the parking contact surface.
[0037] The vehicle parking pose detection device and method provided by the embodiments of the present application, the device includes: an elastic protection structure arranged at the edge position of the target parking area, used for detecting the instantaneous pressure change when the vehicle contacts the edge; a plurality of measuring units arranged in a non-uniform distribution manner in the target parking area, wherein each measuring unit includes a pressure sensor for collecting pressure data after the vehicle enters the target parking area; and a data processing module for determining the parking pose of the vehicle according to the instantaneous pressure change and the pressure data. By increasing the elastic protection structure, the vehicle can be quickly responded and detected when it enters the target parking area, combined with the non-uniformly distributed pressure sensors in the target parking area, and the mixed sensor type design, the hardware cost is reduced while realizing high-precision detection of the parking pose of the vehicle. BRIEF DESCRIPTION OF DRAWINGS
[0038] The accompanying drawings, which are incorporated into and form part of the specification, illustrate embodiments consistent with the present application and, together with the specification, serve to explain the principles of the application.
[0039] Figure 1 The structural schematic diagram of the vehicle parking pose detection device provided by the present application is shown;
[0040] Figure 2 The three-dimensional shape schematic diagram of the rubber airbag and the target parking area provided by the present application is shown;
[0041] Figure 3 The distribution schematic diagram of the measuring unit provided by the present application is shown;
[0042] Figure 4 The structural sectional view of the measuring unit provided by the present application is shown;
[0043] Figure 5 The external view of the measuring unit provided by the present application is shown;
[0044] Figure 6 Effectual top view of vehicle parking position detection device provided in the present application;
[0045] Figure 7 Effectual side view of vehicle parking position detection device provided in the present application;
[0046] Figure 8 Flow chart of vehicle parking position detection method provided in the present application.
[0047] Reference signs:
[0048] 100: elastic protection structure; 200: measurement unit; 201: pressure sensor; 202: base; 203: air bag body; 204: top plate; 205: guide rod.
[0049] The specific embodiments of the present application have been shown and described in the above-described drawings, and will be described in more detail hereinafter. These drawings and the written description are not intended to restrict the scope of the present application concept in any way, but to illustrate the present application concept to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION
[0050] The exemplary embodiments will be described in detail herein with reference to the attached drawings. In the following description, the same numbers are used to indicate the same components throughout the drawings. The embodiments described in the following exemplary embodiments are not meant to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of the present application as detailed in the appended claims.
[0051] In the description of the present application, the terms "first", "second", etc. are used only for the purpose of description, and should not be understood as indicating or implying relative importance or implying the number of the technical features indicated. Therefore, the features defined as "first", "second" can explicitly or implicitly include one or more of the features. In the description of the present application, the meaning of "a plurality of" is two or more, unless otherwise specifically limited.
[0052] In the description of the present application, it should be noted that, unless otherwise specifically defined and limited, the terms "connected", "connected" should be understood broadly, for example, it can be fixedly connected, or detachably connected, or integrally connected; it can be mechanically connected, or electrically connected; it can be directly connected, or indirectly connected through an intermediate medium. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.
[0053] In an automatic car washing system, the vehicle needs to be parked in the car washing area according to the specified requirements to ensure that the car washing mechanical arm, water spraying device, cleaning brush and other equipment can safely and efficiently complete the cleaning work. Therefore, the accuracy of the vehicle parking directly affects the operation efficiency and safety of the car washing system: if the vehicle parking pose deviates too much, it may cause the mechanical arm to collide with the vehicle body, uneven distribution of cleaning liquid or failure of the liquid recovery system, thereby causing equipment damage or user complaints. In addition, a large amount of liquid generated during the car washing process needs to be collected and recovered through the drainage system, so the detection of the vehicle parking position also needs to take into account the efficiency of liquid recovery.
[0054] At present, a visual recognition system is usually used to detect the parking pose of the vehicle, which collects the image of the vehicle through a camera or a laser radar and identifies the vehicle contour and position by combining an image processing algorithm. This method relies on high-precision imaging equipment and complex algorithms, which is costly and is greatly affected by environmental light, rain and snow weather and other disturbances, and has a slow response speed. In addition, there is also a scheme of using a pressure sensor array, which detects the vehicle weight distribution to determine the parking position by laying pressure sensors on the ground of the car washing area. However, the existing scheme usually uses a uniformly distributed sensor array, but has the following problems: (1) the large number of sensors leads to high cost; (2) it is difficult to balance high precision and low cost with uniform layout; (3) the sensors are exposed to car washing liquid and are easily corroded, resulting in high maintenance cost. An infrared / ultrasonic sensing device is used: infrared or ultrasonic sensors are arranged at the edge of the car washing area to determine the position of the vehicle by reflecting signals. This method is sensitive to the surface material of the vehicle, such as dark paint or curved body, and is easily affected by car washing liquid mist, resulting in poor detection stability.
[0055] To solve the above technical problems, the embodiment of the present application provides a vehicle parking pose detection device, which is used to detect the parking pose of the vehicle to assist or ensure that the vehicle is parked in a preset area. For example, in an automatic car washing scenario, the vehicle needs to be automatically identified and determined to be parked in a preset position with a preset pose without the guidance and judgment of the staff. If it does not meet the requirements, an alarm information can be triggered to remind the driver how to park.
[0056] In one embodiment, as shown in Figure 1 The vehicle parking pose detection device comprises:
[0057] An elastic protection structure 100 is arranged at the edge position of the target parking area to detect the instantaneous pressure change when the vehicle contacts the edge;
[0058] A plurality of measurement units 200 are arranged in a non-uniform distribution manner in the target parking area, wherein each measurement unit 200 comprises a pressure sensor 201 for collecting pressure data after the vehicle enters the target parking area;
[0059] The data processing module 300 is used to determine the parking position of the vehicle based on instantaneous pressure changes and pressure data.
[0060] The target parking area refers to the pre-defined area where the vehicle is expected to be parked. For example, in an automatic car wash scenario, since the length of the robotic arm of the automatic car wash system is limited, the robotic arm can only reach every part of the vehicle's surface and achieve a complete car wash if the vehicle is parked in the designated area.
[0061] Vehicle position detection devices typically include a data acquisition side and a data processing side. The data acquisition side is usually deployed at the car wash site. For example, the aforementioned elastic protective structure 100 and multiple measurement units 200 are deployed in the target parking area to achieve data acquisition.
[0062] The elastic protective structure 100 is positioned at the edge of the target parking area, forming an encircling posture around it. The elastic protective structure 100 has a certain degree of elasticity; when a vehicle reaches the edge of the target parking area, or enters or leaves the target parking area, the elastic protective structure 100 will deform under the pressure of the vehicle's tires. If this deformation can be detected, it can be determined that the vehicle has entered or left the target parking area.
[0063] In one embodiment, this deformation can be detected by a barometric pressure sensor. Specifically, the elastic protective structure 100 includes a centrally located rubber airbag arranged around the target parking area;
[0064] And sensors built into the rubber airbag.
[0065] The rubber airbag is hollow and has a certain strength. It is filled with either gas or liquid and maintained at an initial stable pressure. For gas, when a vehicle passes over it, the tire applies downward pressure, compressing the airbag and reducing its internal volume, causing a significant, instantaneous increase in pressure. A pressure sensor connected to one end of the airbag immediately detects this pressure peak. For liquid, since liquid is almost incompressible, the pressure applied when the tire compresses it forces the liquid to flow within the airbag and connected tubing. A flow sensor or differential pressure sensor connected to the tubing detects this instantaneous flow or pressure difference.
[0066] When filling with gas or liquid, to prevent the vehicle from bursting the rubber airbag, the airbag is usually not filled completely with gas or liquid, and the maximum filling pressure of the rubber airbag must be strictly observed.
[0067] Rubber airbags are deployed around the target parking area to detect vehicles entering or leaving the target parking area from any direction.
[0068] In one embodiment, the shape enclosed by the resilient protective structure 100 is the same as the shape of the target parking area. For example... Figure 2 As shown, long, tubular rubber airbags are deployed around the edge of the target parking area.
[0069] Based on the detection principle of the elastic protective structure 100, instantaneous pressure changes can be detected when the vehicle comes into contact with the edge, at which point subsequent detection can be initiated.
[0070] Multiple measurement units 200 are distributed within the target parking area. When a vehicle enters the target parking area, the vehicle's weight triggers pressure sensors in the measurement units 200, thereby detecting the vehicle. Based on the pressure data collected by the multiple measurement units 200, combined with the instantaneous pressure changes collected by the elastic protective structure 100, the vehicle's position and posture can be detected.
[0071] Among them, pressure data refers to the data collected by multiple measuring units 200. It is based on the positional relationship of multiple measuring units in real space to form a matrix to carry positional information, thereby determining the position of the vehicle tires based on the pressure distribution.
[0072] Structurally, in this embodiment, the multiple measurement units 200 are arranged in a non-uniform distribution in the target parking area. That is, the distribution density of the measurement units 200 is different at different locations within the target parking area, and the measurement units 200 do not cover the entire target parking area. This deployment method can save costs, but if there are no other restrictions on the distribution of the multiple measurement units 200, it may affect the detection accuracy, such as if the measurement units 200 are distributed in the edge area.
[0073] Therefore, in practical applications, the target parking area has a central area and an edge area, and vehicles are usually required to park in the central area rather than the edge area. Thus, although the multiple measurement units 200 are not uniformly distributed in the target parking area, the density of measurement units 200 arranged in the central area is greater than the density of measurement units 200 arranged in the edge area. For example, a measurement unit 200 can be arranged every 5cm in the central area, while a measurement unit 200 can be arranged every 10cm in the edge area.
[0074] It should be noted that in some embodiments, the target parking area can be divided into sub-regions, each with a different density. Multiple sub-regions can be divided based on their distance from the center point of the target parking area, with the density gradually decreasing outwards from the center. Alternatively, in some embodiments, the distribution density can be determined based on the size of most vehicle models, using the distance between the four tires as a reference. Specifically, after determining the tire placement positions of various vehicle models parked in the exact center of the target parking area, the areas where the distribution density of the measuring unit 200 needs to be higher are determined, while the distribution density in other areas can be relatively lower. Figure 3 As shown, the distribution density in region A is greater than the distribution density in region B and the distribution density in region C.
[0075] In one embodiment, in the measurement unit 200 located in the diagonal region of the target parking area, the pressure sensor is arranged at a preset tilt angle.
[0076] Normally, although the multiple measurement units 200 in this embodiment are deployed in a non-uniform manner, since the target parking area is usually rectangular, the multiple measurement units 200 are also deployed in a direction parallel to the length and width of the rectangle. This deployment method is mainly sensitive to forces in the X-axis (horizontal) and Y-axis (vertical) directions. When a car is parked at an angle, the sensor points pressed by the tires will form a jagged pattern instead of a smooth line. Based on these discrete points that are only sensitive to the XY directions, calculating the vehicle's centerline and parking angle (i.e., the vehicle's orientation) becomes inaccurate, especially when the sensor density is low. The algorithm may require complex interpolation and fitting, and the results are easily affected by noise.
[0077] This embodiment introduces sensors sensitive to the diagonal direction, providing the system with a completely new and independent measurement dimension, thereby directly capturing oblique features. These pressure sensors can be the same type of pressure sensor as those deployed horizontally and vertically, such as piezoelectric, capacitive, or thin-film sensors, but deployed at a specific angle, such as a preset tilt angle of 45°. When pressure is applied to the protective layer on the sensor surface, the force is transmitted to the sensitive sensing element through the internal structure. Through the design of the mechanical structure, such as an elliptical sensing area, a special force transmission path, or the properties of the encapsulation material, this sensor is made most sensitive to the pressure component parallel to its long axis. As a result, a sensor mounted at a 45° tilt is also sensitive to force in the 45° direction. It has the highest response to pressure parallel to this direction and a weaker response to pressure perpendicular to this direction.
[0078] After reading the pressure data collected by multiple measurement units 200, the data processing module 300 creates a comprehensive "pressure distribution map." This map is no longer just a collection of brightness points on an XY grid, but a vector field containing directional information. This method is helpful in determining the parking orientation of a vehicle.
[0079] In one embodiment, the measuring unit 200 further includes:
[0080] Base 202 has a groove;
[0081] The airbag 203 is placed in the groove, filled with gas, and has a built-in air pressure sensor.
[0082] The measuring unit 200 employs a pressure sensor, which is built into the airbag body 203. The airbag body 203 is placed in a groove in the base 202. The groove opening faces upwards, and the airbag body 203 is filled with gas. Under normal circumstances, the gas-filled airbag body 203 is slightly higher than the depth of the groove. When a vehicle passes by, it presses down on the airbag body 203, causing a change in pressure inside the airbag body 203, which is collected by the pressure sensor.
[0083] Because the gas inside the airbag 203 is shapeless, it causes the vehicle to bounce when it is driven over. Therefore, a roof panel can be added above the airbag 203. Figure 4 and Figure 5 As shown, the top plate 204 covers the opening of the groove, and the airbag body 203 is located in the space formed by the top plate 204 and the groove;
[0084] The top plate 204 is connected to the base 202 via the guide rod 205. When no foreign object passes through the top plate 204 to compress the airbag body 203, the airbag body 203 lifts the top plate 204. When a foreign object passes through the top plate 204 to compress the airbag body 203, the top plate 204 compresses the airbag body 203 along the direction of the guide rod 205.
[0085] The measurement unit 200 provided in the above embodiments has low manufacturing cost and can effectively reduce the cost of the device when used in a vehicle parking posture detection device. Furthermore, by embedding the sensor within a rubber airbag or prefabricated frame, corrosion from car wash liquid is avoided, extending its service life.
[0086] In one embodiment, the sensors in the measuring unit 200 include a barometric sensor and a piezoelectric sensor;
[0087] In the multiple measurement units 200, the proportion of barometric pressure sensors is greater than that of piezoelectric sensors.
[0088] By combining low-cost, fast-response barometric pressure sensors with space-adaptive piezoelectric sensors, the complementary advantages of each are utilized to optimize detection performance. Furthermore, to save costs, barometric pressure sensors account for a larger proportion than piezoelectric sensors, for example, 80% barometric pressure sensors and 20% piezoelectric sensors.
[0089] For the measurement unit 200, if it includes a piezoelectric sensor, the sensor substrate can be made of an elastic polymer material with a conductive coating on its surface. When vehicle pressure is applied to the substrate, the change in resistance of the conductive coating is collected and used to calculate the pressure distribution. The elastic polymer substrate can adapt to complex terrain, such as the slight undulations in the ground of a car wash area, and its material cost is lower than that of traditional barometric pressure sensors, further reducing the overall cost.
[0090] The device provided in the above embodiments includes: an elastic protective structure disposed at the edge of the target parking area for detecting instantaneous pressure changes when a vehicle contacts the edge; multiple measurement units 200 arranged in a non-uniform distribution within the target parking area, wherein each measurement unit 200 includes a pressure sensor for collecting pressure data after the vehicle enters the target parking area; and a data processing module 300 for determining the vehicle's parking posture based on the instantaneous pressure changes and pressure data. By adding the elastic protective structure, rapid response detection is possible when the vehicle enters the target parking area. Combined with the non-uniformly distributed pressure sensors within the target parking area and a hybrid sensor type design, high-precision detection of the vehicle's parking posture is achieved while reducing hardware costs.
[0091] The large amount of liquid generated during the car wash process needs to be collected and recycled through the drainage system. Therefore, the vehicle parking position detection device also needs to take into account the efficiency of liquid recycling.
[0092] In one embodiment, the device further includes a grid guide plate, in which a plurality of measuring units 200 are embedded within the grid of the grid guide plate.
[0093] The grid deflector is used to collect and recover liquid generated during car washing. A measuring unit 200 is embedded within the grid of the deflector, allowing for liquid recovery without affecting the detection capabilities of the measuring unit 200. Grids without the measuring unit 200 can collect liquid.
[0094] Furthermore, embedding the measurement unit 200 within the grid of the grid deflector allows for adaptive density adjustment of the measurement unit 200. For example, the grid can be made movable, or a retractable sensor bracket can be designed in the central area to dynamically adjust the density of the measurement units 200 according to the vehicle size; for instance, 100mm spacing for small cars and 200mm spacing for large cars. By dynamically adjusting the grid density, the number of sensors can be reduced while maintaining accuracy, further lowering costs.
[0095] Based on all the above embodiments, the overall schematic diagram of the provided vehicle parking posture detection device is as follows: Figure 6 and Figure 7 As shown in the diagram, in the detection device, all sensor data is wirelessly transmitted to the data processing module 300 to generate the coordinates and angle information of the vehicle's parking position in real time.
[0096] The data processing module 300 completes the vehicle parking pose detection method. In one embodiment, such as... Figure 8 As shown, it includes:
[0097] Step 801: When an instantaneous pressure change is detected, acquire pressure data collected by multiple measurement units 200;
[0098] Step 802: Based on the pressure data, fit the parking contact surface of the vehicle's tires;
[0099] Step 803: Determine the parking position of the vehicle based on the parking contact surface.
[0100] Among them, based on pressure data, the parking contact surface of the vehicle's tires is fitted, including:
[0101] Clustering the pressure data yields the pressure data for at least one tire of the vehicle.
[0102] Based on the pressure data corresponding to at least one tire, fit the parking contact surface between at least one tire and the target parking area.
[0103] Determine the vehicle's parking position based on the parking contact surface, including:
[0104] Extract the coordinates of the center point of the parking contact surface;
[0105] Calculate the parking direction of the vehicle based on the coordinates of the center point.
[0106] Specifically, the vehicle parking pose detection device includes the following stages:
[0107] Phase 1: Data Acquisition and Preprocessing
[0108] (1) Hardware foundation: Rubber airbags and grid deflectors are laid out. Multiple measurement units are embedded in the grid deflectors. When the vehicle is parked on it, the tires will press against the sensors and generate pressure signals.
[0109] (2) Signal processing: The analog signal of the sensor is converted into a digital pressure matrix through the signal conditioning circuit and the data acquisition module.
[0110] (3) Preprocessing: Use median filtering and other methods to remove abnormal noise points in the data, normalize the pressure value to the [0, 1] interval, and eliminate the influence of absolute pressure magnitude.
[0111] (4) Tire area segmentation:
[0112] Clustering algorithms, such as the improved DBSCAN clustering algorithm, are used to automatically identify and group pressure points belonging to the same tire in the entire pressure matrix.
[0113] This step divides the pressure distribution of the entire vehicle into four independent tire pressure zones, preparing for subsequent individual processing.
[0114] Phase Two: High-Precision Tire Center Positioning – Pressure Distribution Inversion
[0115] (1) Establishing a physical model
[0116] Define a model class that describes an elliptical tire contact surface.
[0117] Model parameters include:
[0118] (x_c, y_c): The center of the ellipse, which is the center of the tire we are looking for.
[0119] a, b: The major and minor axes of the ellipse, representing the size of the tire contact patch.
[0120] theta: The rotation angle of the ellipse, i.e., the parking angle of the vehicle.
[0121] p_max: The maximum pressure value within the contact surface.
[0122] The model specifies how the pressure is distributed within the contact surface, such as the "Gaussian distribution" model, which states that the pressure decreases exponentially with distance from the center.
[0123] (2) Inversion and Nonlinear Optimization
[0124] Objective: To find a set of model parameters [x_c, y_c, a, b, theta, p_max] that minimizes the error between the theoretical pressure distribution calculated by the model and the actual sensor measurements.
[0125] Constructing the loss function: The loss function calculates the mean square error between the theoretical and measured values, and adds a regularization term to prevent the elliptic parameters (a, b) from deviating from a reasonable range (such as the physical dimensions of a tire).
[0126] Initial guess: A rough tire center is quickly calculated using the traditional weighted average method, and the initial major and minor axes of the ellipse are estimated based on the second moment of the pressure distribution, serving as the starting point for the optimization algorithm.
[0127] Iterative optimization: Efficient nonlinear least squares algorithms such as Levenberg-Marquardt are used to continuously adjust the model parameters to minimize the loss function until convergence.
[0128] Output: The optimization algorithm finally outputs the optimal model parameters, where (x_c, y_c) are the high-precision tire center coordinates we need.
[0129] Phase 3: Vehicle Angle Calculation and Verification
[0130] After obtaining the precise center coordinates of the four tires, the next step is to calculate the parking angle of the entire vehicle.
[0131] (1) Calculation of angles using multiple methods:
[0132] PCA Principal Component Analysis: Principal component analysis is performed using the four tire center points as the dataset. The direction of the first principal component is the direction of the vehicle.
[0133] Average axis angle: Calculate the angle of the line connecting the front axle (left front and right front tires) and the angle of the line connecting the rear axle (left rear and right rear tires) respectively, and then take the average value.
[0134] Average diagonal angles: Calculate the angles of the two diagonals (right front - left back, right back - left front) and then take the average.
[0135] (2) Adaptive weighted fusion
[0136] Calculate a "confidence level" for each method.
[0137] PCA confidence level: It is related to the contribution rate of the principal components. The more linear the distribution of data points, the higher the confidence level.
[0138] Axis alignment confidence: If the front and rear axles are very close in direction, the result is reliable.
[0139] Confidence level for diagonal symmetry: If the directions of the two diagonals are very symmetrical, the result is considered reliable.
[0140] Finally, a weighted average is calculated based on the confidence levels of the three angles to obtain the final high-precision vehicle angle θ_final.
[0141] Phase Four: Output and Application
[0142] The calculated precise vehicle position (centers of the four tires) and angle θ_final are output to the upper-level system for:
[0143] Guide to the automatic car wash system.
[0144] Vehicle position and posture recording in intelligent parking lots.
[0145] Precise docking of autonomous vehicles, etc.
[0146] The aforementioned pose detection method, which detects the vehicle's parking position (angle), is powerful in that it uses a continuous physical model to fit discrete sensor data, fundamentally overcoming the physical limitations of sensor spacing; the tire center obtained through optimization algorithms is unaffected by uneven pressure distribution or sensor boundary effects, resulting in highly stable and accurate results; it not only utilizes the magnitude of pressure values but also fully leverages the spatial distribution characteristics of pressure (through elliptical models and multi-scale feature fusion); and through adaptive parameter adjustment and weighted fusion mechanisms, it can automatically adapt to different vehicle types and load conditions, such as sedans, SUVs, and trucks.
[0147] Building upon the basic detection mode, an adaptive sensor adjustment function is added: the sampling frequency of the central area sensor is dynamically adjusted according to the vehicle type (e.g., sedan or SUV) (sedans: high-frequency sampling; SUVs: low-frequency sampling). Based on the pressure distribution characteristics of the elastic protective structure 100, the vehicle's parking direction (e.g., forward or reverse) is predicted, optimizing the data processing algorithm. Specifically, when the vehicle enters, the edge airbag detects the initial contact point to determine the vehicle type and parking direction; the data processing module 300 dynamically adjusts the sampling strategy of the central area sensor based on the prediction result (e.g., sedans require high-frequency sampling to capture subtle displacements); multi-dimensional data is fused in real time to output accurate parking pose information.
[0148] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.
Claims
1. A vehicle parking position detection device characterized by comprising: The device comprises: a flexible protection structure arranged at the edge of the target parking area to detect the instantaneous pressure change when the vehicle contacts the edge; a plurality of measuring units arranged in a non-uniform manner in the target parking area, wherein each measuring unit comprises a pressure sensor for collecting pressure data after the vehicle enters the target parking area; a data processing module for determining the parking pose of the vehicle according to the instantaneous pressure change and the pressure data.
2. The apparatus of claim 1, wherein, The flexible protection structure comprises a middle through rubber airbag arranged around the target parking area, and a gas pressure sensor arranged in the rubber airbag. The shape of the rubber airbag is the same as that of the target parking area.
3. The apparatus of claim 2, wherein, The density of the measuring units arranged in the central area of the target parking area is greater than that of the measuring units arranged in the edge area.
4. The apparatus of claim 1, wherein, The pressure sensors in the measuring units in the diagonal area of the target parking area are arranged at a preset inclination angle.
5. The apparatus of claim 4, wherein, The sensors in the measuring units comprise gas pressure sensors and piezoelectric sensors.
6. The apparatus of claim 1, wherein, The proportion of the gas pressure sensors in the plurality of measuring units is greater than that of the piezoelectric sensors. The measuring unit further comprises:
7. The apparatus of claim 1, wherein, a base having a groove; an airbag body placed in the groove, the airbag being filled with gas and having a gas pressure sensor arranged therein. The measuring unit further comprises:
8. The apparatus of claim 7, wherein, a top plate covering the opening of the groove, the airbag body being located in the space formed by the top plate and the groove; the top plate is connected to the base by a guide rod, when there is no external object pressing the airbag body through the top plate, the airbag body lifts the top plate, when there is an external object pressing the airbag body through the top plate, the top plate presses the airbag body along the guide rod. The device further comprises a grid deflector, and the plurality of measuring units are embedded in the grid of the grid deflector.
9. The apparatus of claim 1, wherein, The data processing module is specifically used for:
10. The apparatus of claim 1, wherein, acquiring the pressure data collected by the plurality of measuring units when the instantaneous pressure change is detected; fitting the parking contact surface of the tire of the vehicle based on the pressure data; determining the parking pose of the vehicle according to the parking contact surface. The fitting of the parking contact surface of the tire of the vehicle based on the pressure data comprises:
11. The apparatus of claim 10, wherein, clustering the pressure data to obtain the pressure data corresponding to at least one tire of the vehicle; fitting the parking contact surface of the at least one tire and the target parking area according to the pressure data corresponding to the at least one tire. The determination of the parking pose of the vehicle according to the parking contact surface comprises:
12. The apparatus of claim 10, wherein, extracting the center point coordinates of the parking contact surface; calculating the parking direction of the vehicle according to the center point coordinates. The method comprises:
13. A vehicle parking position detection method characterized by comprising: acquiring the pressure data collected by the plurality of measuring units when the instantaneous pressure change is detected; fitting the parking contact surface of the tire of the vehicle based on the pressure data; determining the parking pose of the vehicle according to the parking contact surface.