An artificial intelligence-based multi-layer three-dimensional parking garage

Abstract

The application discloses a multi-layer three-dimensional parking garage based on artificial intelligence and relates to the technical field of three-dimensional parking garage.The parking space determination module is combined with target vehicle posture data and the number of to-be-selected parking spaces to determine an optimal parking space, precise parking space matching based on vehicle posture is realized, in the process that the vehicle transportation module transports the vehicle to a target floor by using an elevator, the direction and angle of adjustment can be determined in combination with vehicle posture data, the vehicle posture is adjusted through a rotating disc on the elevator, and the positioning module is matched to accurately determine whether the vehicle reaches the target floor, and the vehicle carrying module determines a support point to complete subsequent carrying, so that the problems that the traditional lane-type three-dimensional parking garage is difficult to accurately match a parking space according to vehicle posture data, and it is inconvenient to adjust the vehicle posture in the process that the vehicle is transported by using the elevator, and it is difficult to determine whether the vehicle accurately reaches the target floor are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of stereoscopic parking garage, in particular to a multi-layer stereoscopic parking garage based on artificial intelligence. BACKGROUND

[0002] The traditional lane type stereoscopic parking garage usually adopts a control strategy based on fixed process allocation of parking spaces or manual auxiliary operation. This kind of method can basically complete the parking task under simple working conditions with sufficient parking spaces and relatively uniform vehicle posture. However, in actual application, the garage often faces complex scenes with various vehicle postures and dynamic changes of parking space resources. The existing control scheme has the following problems: first, most of the existing garages determine the parking space based on fixed parking space allocation logic, lack of fusion and dynamic adaptation of target vehicle posture data, and cannot select the parking space with the highest matching degree combined with the current vehicle heading and other posture information, resulting in the need to turn a larger angle during subsequent vehicle posture adjustment, which is not only complicated and time-consuming.

[0003] Then, the vehicle transportation and posture adjustment process are separated, and the adjustment operation is not carried out simultaneously during the process of transporting the vehicle by the elevator, the overall parking process is not smooth, and the direction and angle of posture adjustment lack precise quantitative basis based on the preset parking orientation of the optimal parking space.

[0004] Finally, the existing garage lacks a compatible mechanism and auxiliary verification means for positioning errors, and it is difficult to accurately determine whether the vehicle is truly parked at the preset position of the target floor only by using vertical position data, which is prone to misjudgment.

[0005] Therefore, there is an urgent need for a lane type multi-layer stereoscopic parking garage based on artificial intelligence, which can accurately match the parking space combined with vehicle posture data, adjust the posture during transportation, verify the floor positioning in multiple dimensions, and realize safe and efficient handling, to solve the above technical bottlenecks and improve the comprehensive performance of the garage in complex application environments. SUMMARY

[0006] In view of the deficiencies of the prior art, the present application provides a multi-layer stereoscopic parking garage based on artificial intelligence, which solves the problems that the traditional lane type stereoscopic parking garage is difficult to accurately match the parking space according to the vehicle posture data, and it is inconvenient to adjust the vehicle posture during the process of transporting the vehicle by the elevator, and it is difficult to determine whether the vehicle has accurately arrived at the target floor.

[0007] To achieve the above purpose, the present application realizes the following technical scheme: a multi-layer stereoscopic parking garage based on artificial intelligence, comprising: a parking space determination module, configured to determine an optimal parking space according to the number of target vehicles and candidate parking spaces and target vehicle posture data.

[0008] The vehicle transportation module is used for transporting the target vehicle to the target floor where the optimal parking space is located through the elevator, and determining the adjustment direction and angle in combination with the target vehicle posture data during the transportation process, and adjusting the target vehicle by using the rotating disc on the elevator.

[0009] The positioning module is used for determining whether the target vehicle reaches the target floor according to the positioning data during the transportation process.

[0010] The vehicle carrying module is used for determining the carrying support point of the target vehicle after the target vehicle posture is adjusted, and carrying the target vehicle to the optimal parking space after reaching the target floor.

[0011] Compared with the prior art, the present application has the following beneficial effects: the optimal parking space is determined by the parking space determination module in combination with the target vehicle posture data and the number of candidate parking spaces, precise parking space matching based on the vehicle posture is realized, the vehicle transportation module can determine the adjustment direction and angle in combination with the vehicle posture data during the process of transporting the vehicle to the target floor by using the elevator, and the vehicle posture is adjusted by using the rotating disc on the elevator, the positioning module accurately determines whether the vehicle reaches the target floor, and the vehicle carrying module determines the support point to complete the subsequent carrying, the parking space matching accuracy, the vehicle posture adjustment convenience and the overall parking efficiency of the lane-type multi-layer stereoscopic parking garage are improved, and the problems that the traditional lane-type multi-layer stereoscopic parking garage is difficult to accurately match the parking space according to the vehicle posture data, it is inconvenient to adjust the vehicle posture during the process of transporting the vehicle by using the elevator, and it is difficult to determine whether the vehicle accurately reaches the target floor are solved. BRIEF DESCRIPTION OF DRAWINGS

[0012] Figure 1 It is a module connection schematic diagram of the multi-layer stereoscopic parking garage based on artificial intelligence.

[0013] Figure 2 It is a structure schematic diagram of the multi-layer stereoscopic parking garage based on artificial intelligence.

[0014] Figure 3 It is a schematic diagram of determining the optimal parking space in combination with the target vehicle posture data.

[0015] Figure 4 It is a process schematic diagram of determining the horizontal adjustment direction and the horizontal adjustment distance. DETAILED DESCRIPTION

[0016] Please refer to Figure 1The embodiment of the application provides a technical scheme: a multi-layer three-dimensional parking garage based on artificial intelligence, comprising a parking space determination module, a vehicle transportation module, a positioning module and a vehicle carrying module, the parking space determination module is connected with the vehicle transportation module, the vehicle transportation module is connected with the positioning module, and the positioning module is connected with the vehicle carrying module, wherein: the parking space determination module is used for determining an optimal parking space according to a target vehicle, a number of candidate parking spaces and target vehicle attitude data.

[0017] The process of determining the optimal parking space is as follows: when the target vehicle arrives on the rotating disc, the candidate parking spaces and the number of candidate parking spaces on the lowest floor are determined by screening from bottom to top.

[0018] If the number of candidate parking spaces on the lowest floor is 1, the empty parking space is the optimal parking space.

[0019] If the number of candidate parking spaces on the lowest floor is 2, the optimal parking space is determined according to the target vehicle attitude data.

[0020] It should be noted that, as shown in Figure 2 The tunnel type three-dimensional parking garage in the embodiment has and only has two three-dimensional parking garages, and each floor of each parking garage has only one parking space. By preferentially locking the empty parking space on the lowest floor, the vertical distance of the elevator transporting the target vehicle can be maximally reduced, the transportation time is shortened, and the transportation cost and time cost increased due to the preferential selection of the parking space on the high floor are avoided.

[0021] When there is only one empty parking space on the lowest floor, the optimal parking space is directly determined, the comparison and decision links of the redundant parking spaces are omitted, the parking space matching can be quickly completed, and parking delay caused by the redundant decision process is avoided. When there are two empty parking spaces on the lowest floor, the optimal parking space is selected in combination with the vehicle attitude data, so that the selected optimal parking space is more suitable for the current attitude of the target vehicle, the difficulty and time consumption of adjusting the direction and angle of the vehicle are reduced, and the effect of accurately matching the parking space according to the vehicle attitude is realized.

[0022] As shown in Figure 3 The process of preferentially determining the optimal parking space according to the target vehicle attitude data is as follows: taking the preset parking direction of the vehicle of one of the candidate parking spaces as a reference direction, the current heading direction of the target vehicle is determined according to the target vehicle attitude data collected by the 3D vision sensor, and the included angle between the heading direction and the reference direction is calculated.

[0023] The included angle between the heading direction of the target vehicle and the preset parking direction of the vehicle of the other candidate parking space is calculated.

[0024] The sizes of the two included angles are compared, and the candidate parking space corresponding to the smallest included angle is taken as the optimal parking space.

[0025] It should be noted that the specific process of determining the head orientation is as follows: first, the 3D vision sensor first performs a full three-dimensional scanning on the target vehicle parked on the rotating disc, obtains the three-dimensional point cloud data of the target vehicle, including the spatial position information of the vehicle appearance, contour and key structure, and obtains the target vehicle attitude data; then, based on the collected three-dimensional point cloud data, the three-dimensional structure features unique to the head of the target vehicle and having direction recognition degree are screened and extracted, such as the unique three-dimensional contour and spatial distribution form of the vehicle front grille, front headlamp group and front bumper, which are the key basis for distinguishing the head and tail of the vehicle and judging the head orientation.

[0026] Then, taking the preset center point of the rotating disc as a reference, a three-dimensional space coordinate system associated with the preset parking orientation of the parking space is established to ensure that the coordinate system can form a unified angle calculation reference with the reference direction of the preset parking orientation of the selected parking space, and the extracted key features of the vehicle head are mapped into the coordinate system to determine the three-dimensional coordinate values of the key features in the coordinate system; finally, according to the three-dimensional coordinate distribution rule of the key features of the vehicle head, such as the three-dimensional coordinates of the front grille being located at the front end of the vehicle, the three-dimensional coordinates of the front headlamp group being symmetrically distributed on both sides of the front grille and being biased towards the front end of the vehicle, the spatial direction pointed by the key features of the vehicle head is determined; and the spatial direction is taken as the head orientation of the target vehicle.

[0027] The 3D vision sensor collects the target vehicle attitude data to determine the head orientation, which ensures the accurate capture of the current attitude of the vehicle, provides an accurate data basis for the matching of the parking space and the attitude, and avoids the problem of excessive deviation of the subsequent vehicle and the parking space orientation caused by the lack of attitude data support in the traditional garage.

[0028] The angles between the head orientation and the preset parking orientations of the two selected parking spaces are calculated respectively, and the parking space with the smallest angle is selected, which can minimize the deviation between the selected optimal parking space and the current attitude of the vehicle. When the subsequent vehicle transportation module adjusts the attitude of the vehicle through the rotating disc, only a minimum angle needs to be turned to make the vehicle orientation meet the preset requirements of the parking space, which greatly reduces the time consumption and operation complexity of attitude adjustment, and fundamentally solves the problem of subsequent adjustment inconvenience caused by the mismatch between the parking space and the attitude.

[0029] Compared with other adjustment methods that need to move the whole position of the vehicle, the rotating adjustment is more concise, time-consuming and shorter, and can ensure that the orientation of the vehicle after adjustment completely adapts to the preset requirements of the parking space, laying a foundation for the subsequent vehicle transportation module to accurately deliver the vehicle into the parking space, and avoiding the difficulty of subsequent transportation or misplacement caused by inaccurate attitude adjustment.

[0030] The vehicle transportation module is used to transport the target vehicle to the target floor where the optimal parking space is located through the elevator, and to determine the adjustment direction and angle in combination with the target vehicle attitude data during the transportation process, and to adjust the target vehicle by using the rotating disc on the elevator.

[0031] In order to ensure that the rotating disc adjustment action is accurately controllable, does not cause direction misalignment due to reference confusion, does not cause insufficient or excessive adjustment due to angle ambiguity, and reduces subsequent handling obstacles caused by adjustment errors.

[0032] Therefore, the process of determining the adjustment direction and angle in combination with the target vehicle attitude data during transportation is: taking the vehicle preset parking orientation of the optimal parking space as the adjustment reference direction.

[0033] The direction of rotating the head orientation of the target vehicle towards the adjustment reference direction is defined as the adjustment direction of the target vehicle.

[0034] The included angle between the head orientation of the target vehicle and the adjustment reference direction is calculated, and the included angle is taken as the adjustment angle of the target vehicle.

[0035] The optimal parking space is the final parking position determined by the previous screening, and the vehicle preset parking orientation thereof is taken as the adjustment reference, so as to ensure that the adjusted vehicle attitude directly matches the parking requirements of the target parking space and avoids the problem that the adjusted vehicle is still misaligned with the parking space orientation due to unclear reference. Taking the preset orientation of the optimal parking space as the reference avoids the problem that the adjusted vehicle orientation does not match the parking space, and there is no need for secondary adjustment after the vehicle arrives at the target floor, thereby reducing process redundancy.

[0036] The adjustment direction is directly defined as the direction of rotating the head orientation towards the reference direction, because the head needs to be finally aligned with the reference direction, so the redundant direction judgment can be skipped and the adjustment target can be directly pointed out, which meets the needs of synchronous adjustment during transportation and avoids invalid rotation caused by direction judgment errors.

[0037] The adjustment angle is determined by the included angle between the head orientation and the reference direction, because the included angle is the deviation amount of the vehicle attitude from the target state, and by quantifying the deviation, on-demand adjustment can be achieved, which avoids excessive or insufficient adjustment caused by experience-based estimation of the adjustment angle. The function of the vehicle transportation module using the rotating disc adjustment on the elevator is adapted, and the rotating disc can accurately control the rotation amplitude.

[0038] The positioning module is used to determine whether the target vehicle arrives at the target floor according to the positioning data during transportation.

[0039] S1, extracting the vertical position data of the target vehicle currently located from the real-time collected positioning data.

[0040] S2, pre-storing the vertical reference coordinates of each floor of the multi-layer stereoscopic parking garage, the vertical reference coordinates being preset vertical height values of the ground surfaces of each floor, and setting a positioning error allowable range.

[0041] S3, compare the vertical position data with the vertical reference coordinates of the target floor, if the vertical position data falls within the corresponding positioning error allowed range, it is preliminarily determined that the target vehicle approaches the target floor.

[0042] S4, extract the auxiliary positioning detection result, if the auxiliary positioning detection result shows that the elevator has stopped at the preset stopping position of the target floor, it is determined that the target vehicle reaches the target floor; if the auxiliary positioning detection result is not satisfied, continue to obtain the vertical position data and repeat S3 until the detection result is satisfied.

[0043] First, irrelevant data such as horizontal position is excluded from real-time collected positioning data, and vertical position data is directly extracted as the key basis for determining whether to reach the target floor, because the distinguishing dimension of the floor of the three-dimensional parking garage is vertical height. Only by extracting the vertical position data can a comparison relationship be established with the vertical reference coordinates of each floor. If redundant data such as horizontal position is retained, the data processing complexity will increase.

[0044] It should be noted that the pre-stored vertical reference coordinates provide a clear judgment standard for whether the target vehicle has reached the target floor, avoiding the confusion caused by the lack of a unified comparison reference, such as the inability to distinguish between different floors without a fixed height reference.

[0045] The positioning error allowed range needs to consider the following two factors: first, the vertical height deviation of the ground of each floor of the multi-story three-dimensional parking garage is usually small, and a narrow error range is needed to ensure the accuracy of the determination of approaching the target floor, avoiding misjudgment of adjacent floors as the target floor; second, the stopping accuracy of the elevator at the target floor can usually be controlled to be in millimeter level, and the error range needs to match the stopping accuracy, avoiding both the frequent repeated detection caused by too strict judgment conditions and the judgment failure caused by too wide error range.

[0046] Therefore, the positioning error allowed range is usually approximately between 5mm-50mm.

[0047] After obtaining the vertical position data, a preliminary screening is achieved by comparing the vertical position data with the vertical reference coordinates of the target floor, first quickly locking the state that the vehicle has approached the target floor, narrowing the judgment range, avoiding starting auxiliary positioning detection when the vehicle is still in the low floor or high floor, reducing unnecessary auxiliary detection operation, and saving resources; At the same time, it provides a pre-laying for the subsequent final judgment, making the judgment process more hierarchical. However, vertical position data alone cannot confirm 100% that the vehicle has actually stopped, such as the elevator may be slightly shaking in the vertical direction due to inertia, causing the vertical position data to fall within the error range but not to be positioned correctly. Therefore, preliminary judgment is designed instead of direct judgment of arrival, which not only utilizes vertical position data to quickly narrow the range, but also leaves enough space for subsequent more accurate auxiliary positioning detection, balancing judgment efficiency and accuracy.

[0048] Finally, it should be noted that the auxiliary positioning detection result is obtained by using a photoelectric sensor or other detectors, such as a limit switch, an infrared alignment detector, provided on the rotating disc. Taking the photoelectric sensor as an example, a photoelectric sensor transmitter is installed on the rotating disc, and a photoelectric sensor receiver is installed at the corresponding position of the preset parking position of each target floor of the multi-layer stereo parking garage. When the elevator transports the target vehicle to the target floor, the photoelectric sensor transmitter continuously emits light signals. If the elevator is accurately parked at the preset parking position of the target floor, the transmitter and the receiver will be accurately aligned, and the light signal can be transmitted to the receiver without obstruction. At this time, the photoelectric sensor outputs a normal detection result, indicating that the auxiliary positioning detection result shows that the elevator has been parked at the preset parking position of the target floor.

[0049] Through the double verification of the preliminary judgment of the vertical position data and the final confirmation of the auxiliary positioning detection, the accuracy of determining whether the target vehicle has arrived at the target floor is greatly improved, avoiding the misjudgment caused by relying only on the vertical position data, preventing the subsequent vehicle handling module from starting handling when not reaching the target floor, and causing equipment collision, vehicle mispositioning and other risks.

[0050] The vehicle handling module is used to determine the handling support point of the target vehicle after the pose adjustment of the target vehicle, and to transport the target vehicle to the optimal parking space after reaching the target floor.

[0051] The process of determining the handling support point of the target vehicle by the starting handling device is: starting the handling device during transportation, and acquiring the target vehicle chassis image by using the image acquisition module on the handling device.

[0052] According to the vehicle type of the target vehicle, the position distribution and center point of the four load-bearing beams in the target vehicle chassis image are determined, and the center point is taken as the handling support point.

[0053] It should be noted that when the elevator receives the target vehicle and starts to transport it to the target floor, the system triggers the handling device starting instruction synchronously. Since the handling device is built into the rotating disc, its starting is linked with the elevator transportation action, and there is no need to wait for the transportation to stop, ensuring parallel operation during transportation. The image acquisition module on the handling device is fixedly installed on the side of the handling device facing the target vehicle chassis, and the number and installation angle of the modules are pre-set and debugged to ensure that the complete area of the target vehicle chassis is covered and the area where the load-bearing beams are located is not missed.

[0054] After the carrier starts, the image acquisition module acquires the target vehicle chassis image in real time at a preset frequency, such as 1-2 times per second. During the acquisition process, because the relative positions of the carrier, rotating disc and vehicle are fixed, image blurring caused by vibration during transportation can be avoided, ensuring that the acquired chassis image is clear and complete. The acquired chassis image is transmitted to the image data processing unit built into the system in real time, only basic noise reduction is performed to remove image noise points caused by minor interference during transportation, no complex operation is performed, and the data is quickly stored to provide a clear original image basis for subsequent support beam identification.

[0055] The system pre-stores a chassis structure feature database of common target vehicles. The database contains the typical position distribution law and contour features of the four support beams of different vehicle models, such as long strip structure and relative distance from other chassis components. Before acquiring the chassis image, the system has determined the specific vehicle model of the target vehicle through early interaction, such as vehicle owner input and pre-identification by the vehicle recognition module. At this time, the corresponding support beam feature data of this vehicle model in the database is directly called as the basis for subsequent identification.

[0056] The system compares and matches the acquired and noise-reduced chassis image with the pre-stored corresponding vehicle support beam feature data, accurately locates the complete contour range of each of the four support beams in the chassis image through image contour recognition and relative position verification, such as confirming the distance between the recognized structure and the chassis tire, transmission shaft and other components, and whether it is consistent with the pre-stored features, and excludes the interference of other non-support components of the chassis.

[0057] For each positioned support beam contour, the system determines the center point thereof through geometric calculation. Specifically, the circumscribed rectangle of the support beam contour is extracted, and the intersection point of the two diagonals of the rectangle is calculated. The intersection point is the center point of the single support beam. After the center points of the four support beams are all calculated, the system directly marks the four center points as the support points of the carrier, and generates coordinate data based on the support points of the carrier's own spatial coordinate system.

[0058] The carrier is started and the support point determination work is carried out simultaneously during the transportation of the target vehicle by the elevator, realizing parallel operation of transportation and support point preparation. The carrier does not need to be started, images acquired and support points determined separately after the vehicle reaches the target floor, avoiding time waste in process connection and effectively shortening the overall time from vehicle transportation to subsequent transportation.

[0059] Using the image acquisition module on the carrier to acquire the target vehicle chassis image can intuitively and clearly present the details of the chassis structure, and can more accurately capture the chassis features than relying on experience or simplified data, laying a reliable data foundation for subsequent positioning of support beam positions and reducing misjudgment of support points due to insufficient information.

[0060] According to the position distribution of the four load-bearing beams in the chassis image of the target vehicle, the design differences of the chassis load-bearing of different vehicle models can be adaptively adapted, so as to avoid misidentifying non-load-bearing structures as support points due to ignoring the differences between vehicle models, prevent damage or instability of the vehicle chassis during transportation, and make the support force of the transporter evenly distributed on the vehicle, further ensuring the stability of the vehicle during transportation, avoiding risks such as tilting and shaking, and providing a safe premise for accurately transporting the vehicle to the optimal parking space.

[0061] The process of transporting the target vehicle to the optimal parking space is that the transporter supports the target vehicle according to the support points, and then lifts up after locking.

[0062] The horizontal adjustment direction and distance are determined, and the target vehicle is horizontally adjusted to the center of the rotating disc. Then the transporter horizontally transports the target vehicle to the optimal parking space, and the locking state is evaluated during the transportation. If the locking state is abnormal, the transportation is stopped until the locking state is normal.

[0063] The transporter supports the vehicle according to the load-bearing beam center point determined in the early stage, which can accurately correspond to the vehicle chassis load-bearing structure and avoid damage to the chassis caused by the deviation of the support point from the load-bearing area. After supporting, the vehicle is locked before being lifted up, which can prevent the vehicle from sliding or tilting relative to the transporter during lifting, and ensure the smoothness of the lifting action.

[0064] First, the vehicle is horizontally adjusted to the center of the rotating disc, and then transported to the optimal parking space. The vehicle posture can be calibrated by the center of the rotating disc as a reference position, which can avoid the deviation of the transportation path caused by the initial position deviation of the vehicle, ensure the accuracy of the horizontal transportation direction and distance to the optimal parking space, and reduce the risk of collision between the vehicle and the edge of the parking space or the risk of not being able to smoothly enter the parking space. The locking state is continuously evaluated during the transportation, which can timely detect abnormal conditions such as loosening of the locking structure. Once the abnormality is found, the transportation is immediately stopped, and then the transportation is continued after the locking is restored to normal, which can effectively prevent safety accidents such as falling and mispositioning of the vehicle during transportation, and ensure the safety of the vehicle and the garage equipment.

[0065] As shown in Figure 4 The process of determining the horizontal adjustment direction and distance is as follows: the X-axis coordinate X1 and Y-axis coordinate Y1 of the center of the rotating disc in the actual space coordinate system of the transporter are obtained, and the X-axis coordinate X2 and Y-axis coordinate Y2 of the center support point in the actual space coordinate system of the transporter are also obtained.

[0066] The straight line direction from (X2, Y2) to (X1, Y1) is determined as the horizontal adjustment direction of the target vehicle.

[0067] The distance between (X2, Y2) and (X1, Y1) is determined as the horizontal adjustment distance of the target vehicle.

[0068] It should be noted that the center support point is the geometric center of the four carrying support points, and the coordinates of the center of the rotating disc and the center support point are obtained through a unified carrier actual space coordinate system, so as to avoid direction confusion or distance calculation deviation caused by non-uniform coordinate system, make the straight line direction of (X2, Y2) pointing to (X1, Y1) and the distance between the two points directly reflect the relative position relationship between the vehicle center and the reference center of the rotating disc, provide accurate quantitative basis for horizontal adjustment, and avoid the problem that the vehicle cannot be aligned with the center of the rotating disc due to deviation of the adjustment direction or inaccurate distance estimation.

[0069] The direction is directly determined by the line connecting the two points, and the distance is directly determined by the distance between the two points, without complex direction conversion or distance estimation. The operation logic is intuitive and easy to understand, can be quickly converted into horizontal adjustment action instruction of the carrier, meets the efficient demand of synchronous adjustment in the transportation process, reduces the redundant links of the adjustment process, and improves the overall parking efficiency. Adjusting the vehicle to the center of the rotating disc can align the vehicle center with the center of the rotating disc, so that the center of the rotating disc becomes the intermediate reference for subsequent carrying to the optimal parking space, avoids the path skew during subsequent carrying to the parking space due to the initial position deviation of the vehicle, ensures the accurate docking of the subsequent horizontal carrying direction and the optimal parking space, and reduces the risk of collision between the vehicle and the edge of the parking space.

[0070] The quantitative method of determining the straight line direction and the distance between the two points is the continuation of the precise logic of the overall technical scheme, which avoids subjective errors caused by non-quantitative basis of adjustment direction and distance, ensures that the horizontal adjustment result is controllable and verifiable, and meets the core demand of multi-layer stereo parking garage for vehicle parking accuracy.

[0071] The process of performing the locking state evaluation during the carrying process is: collecting the locking pressure value of each support point on the target vehicle, and simultaneously collecting the horizontal displacement amount of the target vehicle relative to the carrier.

[0072] The collected locking pressure values of each support point are compared with the locking pressure threshold range, and the collected horizontal displacement amount is compared with the displacement threshold.

[0073] If the locking pressure values of all support points are within the locking pressure threshold range, and the horizontal displacement amount is less than the displacement threshold, it is determined that the locking state is normal, and the carrier continues to perform the carrying operation; otherwise, it is determined that the locking state is abnormal, a carrier stopping instruction is triggered, and after the locking structure is readjusted, the carrying operation is resumed.

[0074] By simultaneously collecting the locking pressure value of each support point and the horizontal displacement amount of the vehicle relative to the carrier, the clamping stability of the support points on the vehicle is ensured from the locking force level, and it is verified whether the locking effect can prevent the vehicle from sliding from the actual displacement level, avoiding the monitoring blind spot of single dependence on pressure value or single dependence on displacement amount, and fully covering the core risk points of the locking state.

[0075] Adopting the locking pressure threshold range and displacement threshold as the judgment standard instead of subjective experience judgment can make the locking state evaluation result unified and quantifiable.

[0076] Need to be explained, the locking pressure threshold range needs to refer to the vehicle model parameters of the target vehicle, to ensure that the locking pressure can provide enough clamping force to prevent the vehicle from sliding, and not exceed the upper limit of the bearing beam resistance to cause chassis deformation. For ordinary passenger cars, the locking pressure threshold range of a single support point is usually between 5-15k, and the total locking force of four support points needs to cover the vehicle weight and leave a certain safety redundancy; for heavier vehicles such as SUVs, the locking pressure threshold range of a single support point will be correspondingly increased to 8-20kN to ensure the adaptation to larger vehicle weight.

[0077] The displacement threshold is usually in the range of 5-10mm to ensure the accuracy of the carrying.

[0078] Once it is determined that the locking state is abnormal, the carrying device will immediately trigger a stop command, and the carrying will be resumed after the locking structure is adjusted again, which can effectively prevent accidents such as vehicle tilting, falling, or vehicle and garage equipment collision caused by continuous carrying in the locking failure state, and provides key protection for the safety of vehicles and garage systems.

[0079] By collecting the locking pressure value of each support point, it is ensured that each support point meets the locking force requirement, avoiding uneven force on the vehicle due to insufficient pressure of a single support point, which may further cause tilting. The front design of four bearing beam center points as support points is directly adapted. The carrying process is to adjust the vehicle horizontally to the center of the rotating disc and then horizontally carry it to the optimal parking space, i.e. the vehicle is always in a horizontal motion state. Therefore, the horizontal displacement is collected instead of the vertical displacement, which is the core safety hazard in the horizontal carrying scene, to ensure that the locking effect can resist the interference of vibration, inertia, etc. in horizontal motion.

[0080] The parking space determination relies on attitude data and angle calculation, positioning relies on vertical reference coordinates and error range, and support point determination relies on chassis image and vehicle model matching. This locking state evaluation process uses quantitative threshold comparison to achieve accurate judgment, which is a continuation of the overall technical logic, avoiding fuzzy control in the key carrying safety link, and ensuring that the whole process from parking space selection to final carrying is in a controllable and accurate state.

[0081] The carrying device is built into the interior of the rotating disc.

[0082] It needs to be explained that the rotating disc is rotatably arranged on the elevator and is located at the center position of the elevator, the carrier is arranged in the interior of the rotating disc, the vehicle is parked on the rotating disc conveniently, and is closed through the closing cover on the rotating disc. When the carrier needs to be started, the closing cover can be moved downward and then retracted twice in the interior of the rotating disc, and the carrier is lifted through the lifting structure. The lifting structure can also move the carrier transversely or longitudinally on the horizontal plane through the horizontal reciprocating motion structure, so that the center point is aligned and the carrying process is realized. The locking structure arranged on the carrier needs to act on four tires respectively, and the telescopic arm is also arranged on the carrier, the support structure corresponding to the carrying support point is arranged on the end of the telescopic arm, and the telescopic arm can also rotate relative to the carrier body.

[0083] The above embodiments can be realized wholly or partially by software, hardware, firmware or any combination thereof. When realized by software, the above embodiments can be realized wholly or partially in the form of a computer program product.

[0084] Those skilled in the art can realize that the modules and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solutions. Those skilled in the art can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0085] In addition, the functional modules in each of the embodiments of the present application can be integrated in one processing module, or each module can exist physically independently, or two or more modules can be integrated in one module.

[0086] The above is merely specific implementation of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical range disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0087] Finally, the above is merely preferred embodiments of the present application, and is not used to limit the present application, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. An artificial intelligence-based multi-story stereoscopic parking garage, characterized by, The application relates to a parking lot management system and method. The application comprises: a parking space determination module for determining an optimal parking space according to the number of candidate parking spaces and the target vehicle attitude data of a target vehicle; a vehicle transportation module for transporting the target vehicle to a target floor where the optimal parking space is located by using an elevator, and determining an adjustment direction and angle in combination with the target vehicle attitude data during the transportation process, and adjusting the target vehicle by using a rotating disc on the elevator; a positioning module for determining whether the target vehicle reaches the target floor according to positioning data during the transportation process; a vehicle handling module for determining a handling support point of the target vehicle after the target vehicle attitude is adjusted, and handling the target vehicle to the optimal parking space after the target vehicle reaches the target floor; the process of starting the handler to determine the handling support point of the target vehicle is as follows: starting the handler during the transportation process, and acquiring a target vehicle chassis image by using an image acquisition module on the handler; 2. The multi-tiered vertical parking garage based on artificial intelligence according to claim 1, characterized in that: determining the position distribution and center point of four bearing beams in the target vehicle chassis image according to the vehicle type of the target vehicle, and taking the center point as the handling support point. the process of determining the optimal parking space is as follows: after the target vehicle reaches the rotating disc, the lowest floor candidate parking space and the number of candidate parking spaces are determined by traversing and screening from bottom to top; if the number of candidate parking spaces of the lowest floor is 1, the empty parking space is the optimal parking space; 3. The multi-tiered vertical parking garage based on artificial intelligence according to claim 2, characterized in that, if the number of candidate parking spaces of the lowest floor is 2, the optimal parking space is determined according to the target vehicle attitude data. the process of determining the optimal parking space according to the target vehicle attitude data is as follows: taking the preset parking orientation of the vehicle of one candidate parking space as a reference direction, determining the current vehicle head orientation of the target vehicle according to the target vehicle attitude data collected by a 3D vision sensor, and calculating the included angle between the vehicle head orientation and the reference direction; calculating the included angle between the vehicle head orientation of the target vehicle and the preset parking orientation of the vehicle of another candidate parking space; 4. The multi-tiered vertical parking garage based on artificial intelligence according to claim 3, characterized in that, comparing the sizes of the two included angles, and taking the candidate parking space corresponding to the smallest included angle as the optimal parking space. the process of determining the adjustment direction and angle in combination with the target vehicle attitude data during the transportation process is as follows: taking the preset parking orientation of the vehicle of the optimal parking space as an adjustment reference direction; taking the direction of rotating adjustment of the vehicle head orientation of the target vehicle to the adjustment reference direction as the adjustment direction of the target vehicle; 5. The multi-tiered vertical parking garage based on artificial intelligence according to claim 1, wherein, calculating the included angle between the vehicle head orientation of the target vehicle and the adjustment reference direction, and taking the included angle as the adjustment angle of the target vehicle. the process of determining whether the target vehicle reaches the target floor according to the positioning data during the transportation process is as follows: S1, extracting the vertical position data of the target vehicle from the real-time collected positioning data; S2, pre-storing the vertical reference coordinates of each floor of the multi-layer stereoscopic parking garage, the vertical reference coordinates being preset vertical height values of the ground of each floor, and simultaneously setting a positioning error allowable range; S3, comparing the vertical position data with the vertical reference coordinates of the target floor, if the vertical position data falls within the corresponding positioning error allowable range, it is preliminarily determined that the target vehicle approaches the target floor; S4, extracting an auxiliary positioning detection result, if the auxiliary positioning detection result shows that the elevator has stopped at the preset stopping position of the target floor, it is determined that the target vehicle reaches the target floor. If the auxiliary positioning detection result does not meet the requirement, the vertical position data is continuously acquired and the step S3 is repeated until the detection result meets the requirement.

6. The multi-tiered vertical parking garage based on artificial intelligence according to claim 1, wherein, The process of carrying the target vehicle to the optimal parking space is as follows: The carrier supports the target vehicle according to the support points, and lifts up after locking; The horizontal adjustment direction and distance are determined, the target vehicle is horizontally adjusted to the center of the rotating disc, and then the carrier carries the target vehicle to the optimal parking space, and the locking state is evaluated during the carrying process. If the locking state is abnormal, the carrying process is stopped until the locking state is normal.

7. The multi-tiered vertical parking garage based on artificial intelligence according to claim 6, characterized in that, The process of determining the horizontal adjustment direction and distance is as follows: The X-axis coordinate X1 and Y-axis coordinate Y1 of the center of the rotating disc in the actual space coordinate system of the carrier are acquired, and the X-axis coordinate X2 and Y-axis coordinate Y2 of the center support point in the actual space coordinate system of the carrier are also acquired; The straight line direction of (X2, Y2) pointing to (X1, Y1) is determined as the horizontal adjustment direction of the target vehicle; The distance between (X2, Y2) and (X1, Y1) is determined as the horizontal adjustment distance of the target vehicle.

8. The multi-tiered vertical parking garage based on artificial intelligence according to claim 6, wherein, The process of evaluating the locking state during the carrying process is as follows: The locking pressure values of each support point on the target vehicle are collected, and the horizontal displacement amount of the target vehicle relative to the carrier is also collected; The collected locking pressure values of each support point are compared with the locking pressure threshold range, and the collected horizontal displacement amount is compared with the displacement threshold; If the locking pressure values of all support points are within the locking pressure threshold range, and the horizontal displacement amount is less than the displacement threshold, it is determined that the locking state is normal, and the carrier continues to perform the carrying operation; otherwise, it is determined that the locking state is abnormal, a carrier stopping instruction is triggered, and after the locking structure is adjusted again, the carrying operation is resumed.

9. The multi-tiered vertical parking garage based on artificial intelligence according to claim 6, wherein, The carrier is built in the interior of the rotating disc.

Images