Parking lot autonomous valet parking system and method based on intelligent network connection technology

Abstract

The invention discloses a parking lot autonomous valet parking system and method based on an intelligent network connection technology. The parking lot autonomous valet parking system comprises a vehicle end unit, a field end unit, a cloud end unit and an APP unit. The vehicle end unit is used for providing a parking request and a vehicle taking request, and acquiring a navigation path to enable a vehicle to automatically drive into a target parking space or a vehicle taking place; the field end unit comprises a sensing device arranged in the parking lot and used for sensing information of vehicles and moving obstacles in the parking lot. The cloud unit predicts traffic flow data of the day through traffic flow historical data, plans parking lot vacancies and access channels in different time periods, generates a parking navigation path or a vehicle taking navigation path and sends the path to the vehicle end unit; a user sends a parking request or a car taking request through an operation interface of the APP unit. The parking lot autonomous valet parking system and method have the advantages that the parking path is scheduled and planned in a unified mode, drivers do not need to participate in parking, the traffic flow efficiency is improved, and the intelligent level and the operation efficiency of the parking lot are improved.

Description

Technical Field

[0001] The present invention relates to an automatic parking technology, and in particular to an autonomous valet parking system and method for a parking lot based on intelligent network connection technology. Background Art

[0002] As a product of the integration of multiple industries, autonomous driving has been valued by countries around the world. As a technology that develops from conditional autonomous driving to highly autonomous driving, the valet parking system has great research value and practical significance. The main function of the valet parking system is that in parking mode, the user gets off the car at the designated drop-off point and issues a parking instruction through the mobile phone APP. After receiving the instruction, the vehicle can automatically drive to the parking space in the parking lot without user operation and monitoring; the user issues a pick-up instruction through the mobile phone APP, and the vehicle can automatically drive from the parking space to the designated pick-up point after receiving the instruction.

[0003] Parking is one of the most common driving behaviors, and parking lots are one of the most common usage scenarios for valet parking. Therefore, autonomous valet parking in parking lots has a huge market prospect. In the prior art, valet parking mainly includes two technical directions.

[0004] First, improve the intelligence level of the single vehicle and realize automatic parking with high-level autonomous driving functions. The APA (AutoParking Assist) automatic parking assistance function is generally operated by the user on the vehicle's central control screen when the vehicle is driving at a low speed (<15km / h) or stationary. It uses a surround camera as the core and is supplemented by sensing devices such as ultrasonic radar to perceive the surrounding parking spaces in real time, including different types of parking spaces such as straight garages, diagonal garages, and side garages, and display real-time images on the central control screen. The domain control will plan the path to park in the nearby empty parking spaces in real time based on the detection results of the perception system. If the planning is successful, the user will be prompted on the central control screen, otherwise there will be no prompt. After the user clicks on the parking space that can be parked according to the needs, the vehicle can automatically park in the parking space according to the planned path. After the function is completed, it automatically switches to P gear and exits the function.

[0005] VPA (Valet Parking Assist) parking lot memory parking, based on APA, adds a low-speed automatic driving function for a fixed route. At the entrance of the parking lot, when entering the VPA function for the first time, the user drives the vehicle to the target parking space and parks it. During this process, the speed must be kept below 15km / h. The domain control will learn and record the path from the starting point. When the function is enabled again, enter the system near the starting point, and the vehicle will automatically drive to the target parking space. During this process, it will automatically bypass obstacles and park in the parking space through APA after arriving near the parking space.

[0006] In the existing technology, automatic parking functions such as Xiaopeng's APA and VPA are limited by the vehicle's perception technology and intelligence level, and still require the driver to participate in the whole process and take over when necessary. The driver is still the main person responsible and cannot leave the vehicle to achieve remote autonomous parking. At the same time, due to cost constraints, the performance of its perception system is limited, the perception range is small, and the domain control computing power is also limited, so it is difficult to complete more complex path planning and cannot cover all parking scenarios.

[0007] Secondly, the automated parking lot is equipped with an automatic car moving robot, which transports the vehicle to complete parking. Most of the existing parking lots are relatively old, and there are unreasonable aspects in the parking space planning. The existing technology for parking lot transformation is divided into two steps. First, the user selects a parking space on the interactive interface. After the parking space is selected, there is no need to rely on the intelligence of the single vehicle. Instead, the parking lot is equipped with a consignment robot, which is in the shape of a plate and automatically slides under the vehicle, lifts the vehicle and consigns it to the designated parking space. During this process, the driver can directly turn off the engine and get off the vehicle, and the consignment robot will take over to complete the automatic valet parking.

[0008] In order to meet the needs of efficient parking for multiple vehicles in multiple batches, automated parking lots usually require high-cost transport robots. At the same time, their operating speed must also be taken into account. However, high-frequency and high-load work can easily damage the robots, posing a safety hazard. Summary of the invention

[0009] The present invention is to avoid the deficiencies existing in the above-mentioned prior art and to provide a parking lot autonomous valet parking system and method based on intelligent network connection technology, so as to uniformly schedule and plan parking routes, eliminate the need for driver participation in parking, improve traffic flow efficiency, and enhance the intelligence level of parking lots.

[0010] The present invention adopts the following technical solutions to solve the technical problems.

[0011] The present invention discloses an autonomous valet parking system for parking lots based on intelligent network connection technology, which includes a vehicle-end unit, a parking-end unit, a cloud-end unit and an APP unit;

[0012] The vehicle-side unit is used for sending vehicle information to the cloud unit after the vehicle issues a parking request in the passenger drop-off area, receiving the target parking space and the parking navigation path from the current position to the target parking space sent by the cloud unit, and automatically driving to the target parking space according to the parking navigation path to complete the parking work; when receiving the vehicle pickup request, the vehicle pickup location and the vehicle pickup navigation path from the parking space where the vehicle is located to the vehicle pickup location sent by the cloud unit, driving to the vehicle pickup location in the delivery area according to the vehicle pickup navigation path to complete the vehicle delivery work;

[0013] The field end unit includes a sensing device arranged in the parking lot, which is used to sense the information of vehicles and moving obstacles in the parking lot to achieve obstacle avoidance and positioning of the vehicle, send the real-time road conditions in the parking lot to the cloud unit, and provide real-time monitoring in the parking lot;

[0014] The cloud unit stores historical traffic flow data, and the traffic flow prediction model predicts the traffic flow data of the day through the historical traffic flow data, and plans parking lot vacancies and entry and exit channels at different time periods; receives the parking request sent by the APP unit, selects the target parking space according to the current position of the vehicle, generates a parking navigation path from the current position of the vehicle to the target parking space, and sends it to the vehicle-side unit; receives the pick-up request and pick-up location sent by the APP unit, generates a pick-up navigation path from the parking space where the vehicle is located to the pick-up location, and sends it to the vehicle-side unit;

[0015] The APP unit is set on the mobile terminal, and is registered with the real name system and bound to the license plate number; the user sends a parking request or a vehicle pickup request through the operation interface of the APP unit.

[0016] The parking lot autonomous valet parking system and method based on intelligent network connection technology of the present invention are also characterized in that:

[0017] Furthermore, the vehicle-end unit and the field-end unit communicate with each other in a V2X manner.

[0018] Furthermore, the vehicle information includes vehicle size data and a license plate number.

[0019] Furthermore, the sensing device includes a camera and a radar.

[0020] Furthermore, the field end unit converts the position information of the mobile obstacle into coordinate information relative to each nearby vehicle in the vehicle coordinate system, and converts the position information of each vehicle into global coordinate information of the parking lot, and sends it to each vehicle to achieve obstacle avoidance and positioning of the vehicle.

[0021] Furthermore, the cloud unit obtains the traffic flow data of the day for iterating the traffic flow prediction model.

[0022] Furthermore, the cloud unit also predicts the traffic volume at a designated location based on time nodes of the day, week, month, and year.

[0023] Furthermore, the APP unit includes a map display module for displaying a parking navigation path or a car pickup navigation path.

[0024] Furthermore, the APP unit is used to locate the vehicle or obtain vehicle monitoring images.

[0025] The present invention also discloses a parking method of an autonomous valet parking system of a parking lot based on intelligent network connection technology, which is characterized by comprising the following steps:

[0026] Step 1: The driver checks the available parking spaces in the parking lot near the destination through the APP unit, selects the target parking lot and drives to the drop-off area of ​​the target parking lot;

[0027] Step 2: The driver sends a parking request through the APP unit, and the vehicle terminal unit sends the vehicle information to the cloud unit;

[0028] Step 3: Select a target parking space according to the current position of the vehicle, generate a parking navigation path from the current position of the vehicle to the target parking space, and send it to the vehicle end unit;

[0029] Step 4: The vehicle-side unit receives the target parking space and the parking navigation path from the current position to the target parking space sent by the cloud unit, and automatically drives to the target parking space according to the parking navigation path to complete the parking work;

[0030] Step 5: The driver sends a pick-up request through the APP unit, and the cloud unit receives the pick-up request and the pick-up location sent by the APP unit, generates a pick-up navigation path from the parking space where the vehicle is located to the pick-up location and sends it to the vehicle-side unit;

[0031] Step 6: The vehicle-end unit drives to the pick-up location in the delivery area according to the pick-up navigation route to complete the delivery.

[0032] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0033] The present invention discloses an autonomous valet parking system and method for a parking lot based on intelligent network connection technology. The system includes a vehicle-end unit, a parking lot-end unit, a cloud unit and an APP unit. The vehicle-end unit is used to send vehicle information to the cloud unit after the vehicle issues a parking request in the passenger drop-off area, receive the target parking space and the parking navigation path from the current position to the target parking space sent by the cloud unit, and automatically drive to the target parking space according to the parking navigation path to complete the parking work; when receiving the vehicle pickup request, the vehicle pickup location and the vehicle pickup navigation path from the parking space where the vehicle is located to the vehicle pickup location sent by the cloud unit, the vehicle is driven to the vehicle pickup location in the delivery area according to the vehicle pickup navigation path to complete the vehicle delivery work; the parking lot-end unit includes a sensing device arranged in the parking lot, which is used to sense the information of vehicles and moving obstacles in the parking lot to realize the obstacle avoidance and positioning of the vehicle The cloud unit stores historical traffic data, and the traffic flow prediction model predicts the traffic flow data of the day through the historical traffic flow data, and plans the parking lot vacancies and entry and exit channels at different time periods; receives the parking request sent by the APP unit, selects the target parking space according to the current position of the vehicle, generates a parking navigation path from the current position of the vehicle to the target parking space, and sends it to the vehicle-side unit; receives the pick-up request and pick-up location sent by the APP unit, generates a pick-up navigation path from the parking space where the vehicle is located to the pick-up location, and sends it to the vehicle-side unit; the APP unit is set on the mobile terminal, and is registered with the real name system and bound to the license plate number; the user sends a parking request or a pick-up request through the operation interface of the APP unit.

[0034] The parking lot autonomous valet parking system and method based on intelligent network connection technology of the present invention has the following technical features.

[0035] 1. The interaction between the cloud control center of the parking lot cloud unit and the navigation system and its data processing can make predictions and analyses based on historical parking records and the navigation traffic flow of the day, rationally plan the allocation of parking spaces for each period of the day, publish real-time occupancy status, and update the traffic flow prediction model;

[0036] 2. The cloud control center interacts with the mobile APP unit. Users can access the parking lot’s cloud through the APP when leaving the car and remotely operate the parking and pick-up operations;

[0037] 3. The cloud control center interacts with the vehicle-side unit and performs multi-vehicle planning. The vehicle will send its own vehicle geometry information to the cloud unit, and the cloud unit will simultaneously plan multi-vehicle paths based on real-time parking requirements;

[0038] 4. The field end unit interacts with the vehicle end unit of the vehicle, senses the surrounding mobile obstacle information, performs relevant coordinate conversion calculations, and then sends it to the vehicle, so that the vehicle can avoid obstacles during autonomous parking.

[0039] The parking lot autonomous valet parking system and method based on intelligent network connection technology of the present invention has the advantages of unified scheduling and planning of parking routes, no need for driver participation in parking, improved traffic flow efficiency, and improved parking lot intelligence level and operation efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a schematic diagram of an autonomous valet parking system for a parking lot based on intelligent network connection technology of the present invention.

[0041] Figure 2 It is a schematic diagram of a parking lot global coordinate system, a field equipment coordinate system, and a vehicle coordinate system of an autonomous valet parking system in a parking lot based on intelligent network connection technology of the present invention.

[0042] Figure 3 This is a chain model structure diagram of the LSTM neural network of a parking lot autonomous valet parking system based on intelligent networking technology of the present invention.

[0043] The present invention will be further described below through specific implementation modes in conjunction with the accompanying drawings. DETAILED DESCRIPTION

[0044] See also Figure 1 to Figure 3 , an autonomous valet parking system for parking lots based on intelligent network connection technology of the present invention comprises a vehicle end unit, a field end unit, a cloud unit and an APP unit;

[0045] The vehicle-side unit is used for sending vehicle information to the cloud unit after the vehicle issues a parking request in the passenger drop-off area, receiving the target parking space and the parking navigation path from the current position to the target parking space sent by the cloud unit, and automatically driving to the target parking space according to the parking navigation path to complete the parking work; when receiving the vehicle pickup request, the vehicle pickup location and the vehicle pickup navigation path from the parking space where the vehicle is located to the vehicle pickup location sent by the cloud unit, driving to the vehicle pickup location in the delivery area according to the vehicle pickup navigation path to complete the vehicle delivery work;

[0046] The field end unit includes a sensing device arranged in the parking lot, which is used to sense the information of vehicles and moving obstacles in the parking lot to achieve obstacle avoidance and positioning of the vehicle, send the real-time road conditions in the parking lot to the cloud unit, and provide real-time monitoring in the parking lot;

[0047] The cloud unit stores historical traffic flow data, and the traffic flow prediction model predicts the traffic flow data of the day through the historical traffic flow data, and plans parking lot vacancies and entry and exit channels at different time periods; receives the parking request sent by the APP unit, selects the target parking space according to the current position of the vehicle, generates a parking navigation path from the current position of the vehicle to the target parking space, and sends it to the vehicle-side unit; receives the pick-up request and pick-up location sent by the APP unit, generates a pick-up navigation path from the parking space where the vehicle is located to the pick-up location, and sends it to the vehicle-side unit;

[0048] The APP unit is set on the mobile terminal, and is registered with the real name system and bound to the license plate number; the user sends a parking request or a vehicle pickup request through the operation interface of the APP unit.

[0049] like Figure 1 The present invention provides an autonomous valet parking system and method for a parking lot based on intelligent networking technology, wherein the cloud unit includes a historical database and a cloud control host. The historical database can store historical traffic data, and retrieve historical data for learning to predict the traffic volume of the day, thereby scheduling parking spaces in different areas of the parking lot in advance and planning the channels for parking and picking up vehicles on the day. The sensing equipment on the field side is mainly visual monitoring cameras and radars, which can sense moving obstacles and send their relative positions to the vehicle through V2X technology for use in obstacle avoidance planning. At the same time, the field side equipment can provide monitoring services, which are uniformly managed by the field side staff, i.e., the administrator. The driver can obtain monitoring information through the APP to ensure the safety of his own vehicle. Both the vehicle side and the APP side can access the parking lot cloud control center server, access data and interact, thereby realizing four-end intelligent networking and completing the autonomous valet parking task.

[0050] In a specific implementation, the vehicle-end unit and the field-end unit communicate with each other in a V2X manner.

[0051] When the user arrives near the parking lot, the vehicle's vehicle-side unit communicates with the parking lot's field-side unit based on V2X technology. In specific implementation, the vehicle's vehicle-side unit and the parking lot's field-side unit will automatically connect within a certain range.

[0052] During specific implementation, the vehicle information includes vehicle size data and a license plate number.

[0053] The driver enters the passenger drop-off area and sends a parking request to the cloud unit through the APP unit. At the same time, the vehicle's vehicle-side unit sends the vehicle's size information and license plate number to the cloud unit. The cloud unit will automatically search for the target parking space and plan a global parking navigation path to the target parking space and send it to the vehicle-side unit of the vehicle. The vehicle will automatically drive to the target parking space according to the parking navigation path, and turn off the engine to inform the cloud unit to complete the autonomous parking work. The route of the parking navigation path can be displayed in real time on the APP unit according to the vehicle positioning. An animated route. When the user needs to pick up the car, specify the pick-up location on the APP and send a pick-up request to the cloud unit. The cloud unit will automatically plan the route and generate a pick-up navigation path and send it to the vehicle. The vehicle will automatically drive to the pick-up location in the delivery area according to the pick-up navigation path to complete the delivery. In the present invention, a relatively mature path planning technology in the prior art is used to plan the navigation path. In the present invention, the path planning task is deployed in the cloud, and the paths of multiple vehicles are uniformly planned and sent to each vehicle.

[0054] In a specific implementation, the sensing device includes a camera and a radar.

[0055] In specific implementation, the field end unit converts the position information of the mobile obstacle into the coordinate information of each nearby vehicle in the self-vehicle coordinate system, and converts the position information of each vehicle into the global coordinate information of the parking lot, and sends it to each vehicle to achieve obstacle avoidance and positioning of the vehicle.

[0056] In the present invention, the calculation formula for converting the position information of the moving obstacle into the vehicle coordinate system is shown in the following formula (1).

[0057]

[0058] In formula (1), (x i ,y i ,θ i ) is the position information of the moving obstacle in the vehicle's ego-vehicle coordinate system; It is the position information of a moving obstacle in the field equipment coordinate system; is the position information of the reference vehicle in the field equipment coordinate system. This coordinate transformation operation converts the position of the mobile obstacle in the field equipment coordinate system to the reference vehicle's own coordinate system. i , The equal angles are constrained to the range [-π, π].

[0059] In the present invention, the calculation formula for converting the position information of each vehicle into the global coordinates of the parking lot is shown in the following formula (2).

[0060]

[0061] In formula (2), (X car ,Y car,φ car ) is the vehicle’s position information in the parking lot’s global system, (X m ,Y m ,ψ m ) is the position information of the field equipment in the global system of the parking lot (this information is fixed during installation). is the position information of the vehicle in the field equipment coordinate system. car ,ψ m , The equal angles are constrained to the range [-π, π].

[0062] The field end unit includes sensing devices with integrated V2X functions, such as vision or radar, which are arranged at fixed points in the parking lot. The sensing device is used to sense the relative position information of vehicles and moving obstacles in and near the parking lot, and convert the position information of moving obstacles into the coordinate information of the vehicle system relative to nearby vehicles, and convert the vehicle position information into the global coordinate information of the parking lot, and then realize communication with the vehicle based on V2X technology, and send the above information to each vehicle to achieve local obstacle avoidance and vehicle positioning.

[0063] The terminal unit uploads the real-time traffic conditions in the field to the cloud unit. In case of emergencies or renovations that hinder normal traffic, the cloud unit will be assisted to adjust the parking space planning and clear the traffic in time to avoid congestion.

[0064] The monitoring equipment of the field end unit can provide a security monitoring terminal to prevent damage and theft of vehicles, and is managed uniformly by the administrator, that is, the parking lot staff.

[0065] In a specific implementation, the cloud unit obtains the vehicle flow data of the day for iterating the vehicle flow prediction model.

[0066] In specific implementation, the cloud unit also predicts the traffic volume at a designated location based on time nodes of the day, week, month, and year.

[0067] The cloud unit can be a cloud control center for parking lots. The tasks of the cloud unit mainly include the following three parts:

[0068] 1. Based on the time nodes such as day, week, month, and year, the possible traffic flow of a certain place and the length of time vehicles stay in the parking lot within each hour can be inferred. Therefore, the traffic flow of the day is predicted based on the stored historical data, and the empty parking spaces and entry and exit channels of the parking lot at different times are planned in advance to improve the operation efficiency. At the same time, the empty parking space information is sent to the server in real time, which can be accessed and selected by the driver through the mobile APP unit. The daily data is stored in the cloud and can be used to iterate the prediction model;

[0069] In the present invention, daily data from the past few years are taken, with date x s1 、Lunar calendar date xs2 、Whether it is a holiday x s3 、Day of the week x s4 、Weather s5 , the average number of vehicles entering and leaving each hour on the previous day x s6 、The previous Japanese parking lot parking fee for each type of charge x s7 (such as membership, points, direct payment, etc.), the number of vehicles entering and leaving the warehouse on the previous day x s8 , the average parking time of the previous day, etc. s9 , forming a sequence X according to the time sequence of natural days t =[x s1 ,x s2 ,x s3 ,x s4 ,x s5 ,x s6 ,x s7 ,x s8 ,x s9 ], select an appropriate sliding window width, input sequence information in a rolling window manner, build an LSTM neural network prediction model for training, and output the predicted parking time sequence of vehicles entering the warehouse every hour in the next day to achieve the prediction of the traffic flow on the day. The input and output sequences predicted in the present invention, and the parking time of vehicles entering the warehouse every hour in the next day are used to characterize the traffic flow.

[0070] After 24:00 every day, the data of the day will enter the training set, and the LSTM neural network prediction model will be iteratively upgraded. The LSTM neural network prediction model of the present invention adopts a conventional and mature iterative algorithm. Those skilled in the art can refer to relevant papers, which will not be repeated here.

[0071] LSTM (Long Short-Term Memory) neural network is a special type of recurrent neural network. The so-called recurrent neural network, or RNN, is the most natural architecture for time series data. It uses the "internal loop" of the neural network to retain the contextual information of the time series, and can use past signal data to infer the understanding of the current signal. In theory, RNN can retain information from any moment in the past. But in actual use, we may not only need to use the information from the most recent moment, but also the information from a very long time ago. However, in RNN, the transmission of information will gradually decay due to the long time interval, so it has no good way to deal with the problem of long-term dependence on information. The special structural design of LSTM allows it to avoid long-term dependence problems. Remembering information from very early moments is its default behavior, without having to pay a high price for it. It constructs a chain model through forget gates, memory gates, and output gates. Figure 3 As shown, the processing and updating of the current input, the last moment information and the past information are realized.

[0072] Figure 3 This is a chain model structure diagram of the LSTM neural network of the present invention. Figure 3 The three blocks in the middle, left, center and right have the same structure, but they are connected in series.

[0073] Figure 3 In, X t is the input information at the current moment, h t is the output information at the current moment;

[0074] Sigmoid function

[0075] tanh function

[0076] Human activities often have a certain periodicity. The traffic flow in a specific place during a specific period of time is a typical example. Therefore, we consider using LSTM neural network for predictive analysis.

[0077] LSTM neural network and sliding window method are commonly used methods for processing time series problems. The present invention emphasizes the use of time series information as input and the construction of a multivariate input network for traffic flow prediction.

[0078] The LSTM neural network first calculates the forget gate vector f based on the result of the previous moment. t , where W f is the forget gate weight matrix, b f To forget the gate bias matrix, the calculation process is as follows.

[0079] f t =σ(W f ·[h t-1 ,x t ]+b f )

[0080] Next, calculate the input gate vector i t and the input gate candidate vector Where W i is the input gate weight matrix, b i is the input gate bias matrix, W C is the input gate candidate weight matrix, b C is the input gate candidate bias matrix, and the calculation process is as follows.

[0081] i t =σ(W i ·[h t-1 ,x t ]+b i )

[0082]

[0083] By forgetting the gate vector f t and the input gate vector i t Calculate the candidate value C t

[0084]

[0085] Finally, calculate the output gate vector o t And the final output h t , where W o is the output gate vector weight matrix, b o is the output gate vector bias matrix.

[0086] o t =σ(W o [h t-1 ,x t ]+b o )

[0087] h t =o t *tanh(C t )

[0088] 2. The cloud unit receives the parking request sent by the mobile APP unit in real time, and plans the parking navigation path and trajectory from the current position of the car to the selected target parking space, and sends it to the car;

[0089] 3. The cloud unit receives the car pickup request sent by the mobile APP unit in real time, and plans the car pickup navigation path and trajectory from the current parking space to the pickup location in the delivery area, and sends it to the car.

[0090] The traffic volume output by the LSTM neural network prediction model of the present invention is not the size of the traffic volume at a designated location, but rather the predicted parking time of vehicles entering the parking lot every hour within 24 hours in the next day.

[0091] The output is the final output ht after the chain network calculation, which is a sequence of [ht1, ht2, ..., ht24], with a total of 24 elements, representing the predicted values ​​of the average parking time of vehicles entering the parking lot in each time period (24-hour system) from 0 to 1 o'clock, 1 to 2 o'clock, ..., 23 to 24 o'clock, that is, the definition of the traffic flow described in the present invention.

[0092] In a specific implementation, the APP unit includes a map display module for displaying a parking navigation path or a vehicle pickup navigation path.

[0093] During specific implementation, the APP unit is used to locate the vehicle or obtain vehicle monitoring images.

[0094] The APP unit is a mobile APP software. The APP software has a built-in map service, which provides an operating platform for users to select parking lots, pick up cars, pay fees, and access parking lot cloud data. APP registration adopts a real-name system and is bound to the license plate number. When making a parking request, you need to first select the license plate for this trip, and then you can remotely operate parking and picking up the car at the selected parking lot. The parking lot deducts the parking fee directly from the account by scanning the license plate. In addition, users can also check the location of the vehicle through the APP. If the field monitoring equipment finds damage or theft, the cloud control center or administrator will also report to the user through the APP.

[0095] The present invention also discloses a parking method of an autonomous valet parking system of a parking lot based on intelligent network connection technology, which is characterized by comprising the following steps:

[0096] Step 1: The driver checks the available parking spaces in the parking lot near the destination through the APP unit, selects the target parking lot and drives to the drop-off area of ​​the target parking lot;

[0097] Step 2: The driver sends a parking request through the APP unit, and the vehicle terminal unit sends the vehicle information to the cloud unit;

[0098] Step 3: Select a target parking space according to the current position of the vehicle, generate a parking navigation path from the current position of the vehicle to the target parking space, and send it to the vehicle end unit;

[0099] Step 4: The vehicle-side unit receives the target parking space and the parking navigation path from the current position to the target parking space sent by the cloud unit, and automatically drives to the target parking space according to the parking navigation path to complete the parking work;

[0100] Step 5: The driver sends a pick-up request through the APP unit, and the cloud unit receives the pick-up request and the pick-up location sent by the APP unit, generates a pick-up navigation path from the parking space where the vehicle is located to the pick-up location and sends it to the vehicle-side unit;

[0101] Step 6: The vehicle-end unit drives to the pick-up location in the delivery area according to the pick-up navigation route to complete the delivery.

[0102] The present invention provides a parking lot autonomous valet parking system and method based on intelligent network connection technology, and the specific detailed working process is as follows:

[0103] Step S1: The cloud control host of the cloud unit predicts the traffic flow of the day based on historical data, plans parking areas and entry and exit channels, and updates the available parking spaces in real time.

[0104] Step S2: Using the result of step S1, the user checks whether there are any empty parking spaces in the parking lot near the destination through the APP, selects the target parking lot, and navigates to the drop-off area of ​​the parking lot.

[0105] Step S3: After arriving at the parking lot according to the navigation in step S2, the vehicle-side unit establishes communication with the cloud-side unit. After the user gets off the vehicle in the drop-off area, the user sends a parking instruction through the mobile phone APP, and the vehicle-side unit sends the vehicle information to the cloud-side unit.

[0106] Step S4: After receiving the instruction of step S3, the cloud unit performs path planning according to the dispatched parking space and vehicle information, and packages the target parking space and the planned path into an instruction set and sends it to the vehicle-side unit.

[0107] Step S5: The vehicle-side unit receives the parking instruction set of step S4, establishes communication with the field-side unit, obtains the vehicle positioning information, and starts to execute path following according to the instruction set.

[0108] Step S6: After completing the parking in step S5, the vehicle-end unit turns off the engine and sends the information to the cloud unit.

[0109] Step S7: The user sends a pick-up instruction to the cloud unit through the mobile phone APP. After the cloud unit completes the planning, it sends a parking path instruction set to the vehicle-side unit and directly deducts the parking fee from the account by binding the license plate.

[0110] Step S8: After the vehicle receives the instruction set of step S7, the terminal unit provides positioning, drives to the delivery area, and completes the vehicle pickup.

[0111] The present invention provides an autonomous valet parking system and method for a parking lot based on intelligent network connection technology, which has the following technical features.

[0112] 1. It improves the intelligence level of parking lots, facilitates management, and can effectively increase utilization rates and increase parking lot revenue.

[0113] 2. It has improved the level of smart transportation. Users can select the target parking lot through the APP, and the parking lot cloud control center will uniformly schedule and plan the parking route to solve the problem of difficulty in parking.

[0114] 3. The cloud control center uses a time series algorithm based on deep learning to learn the historical data of parking lot traffic, predict the parking demand for the day, and plan parking spaces in advance.

[0115] 4. The cloud control center uniformly plans the entry and exit routes of multiple vehicles to improve traffic efficiency and avoid congestion.

[0116] 5. The field equipment communicates with the vehicle based on V2X technology, and the vehicle performs obstacle avoidance actions based on the obstacle information provided by it.

[0117] 6. The field equipment has monitoring function to ensure the safety of the field.

[0118] 7. The driver does not need to participate in the parking process and can park the vehicle remotely through the APP at the drop-off area and pick-up area.

[0119] The parking lot autonomous valet parking system and method based on intelligent network connection technology of the present invention aims at the problems that the current intelligence level of a single vehicle is not enough to complete the autonomous valet parking task, and it is difficult to find empty parking spaces in the parking lot, and proposes a solution for the autonomous valet parking system of the intelligent network connection parking lot. The present invention focuses on the combination of automated parking lot transformation and improvement of the intelligent human-computer interaction function on the vehicle side. Instead of using a consignment robot, it uses the cloud control center and the field-side equipment to work together with the vehicle through V2X communication technology and intelligent network connection technology to complete efficient autonomous valet parking tasks. In this way, it can be achieved using the current intelligence level of a single vehicle, and the cloud control center and the field-side equipment can provide navigation information and obstacle avoidance reminders for multiple vehicles at the same time, which can effectively improve the operation efficiency of the parking lot.

[0120] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

[0121] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

Claims

1. An autonomous valet parking system based on intelligent network technology, characterized in that: Including vehicle-side unit, field-side unit, cloud-side unit and APP unit; The vehicle-side unit is used for sending vehicle information to the cloud unit after the vehicle issues a parking request in the passenger drop-off area, receiving the target parking space and the parking navigation path from the current position to the target parking space sent by the cloud unit, and automatically driving to the target parking space according to the parking navigation path to complete the parking work; when receiving the vehicle pickup request, the vehicle pickup location and the vehicle pickup navigation path from the parking space where the vehicle is located to the vehicle pickup location sent by the cloud unit, driving to the vehicle pickup location in the delivery area according to the vehicle pickup navigation path to complete the vehicle delivery work; The field end unit includes a sensing device arranged in the parking lot, which is used to sense the information of vehicles and moving obstacles in the parking lot to achieve obstacle avoidance and positioning of the vehicle, send the real-time road conditions in the parking lot to the cloud unit, and provide real-time monitoring in the parking lot; The cloud unit stores historical traffic flow data, and the traffic flow prediction model predicts the traffic flow data of the day through the historical traffic flow data, and plans parking lot vacancies and entry and exit channels at different time periods; receives the parking request sent by the APP unit, selects the target parking space according to the current position of the vehicle, generates a parking navigation path from the current position of the vehicle to the target parking space, and sends it to the vehicle-side unit; receives the pick-up request and pick-up location sent by the APP unit, generates a pick-up navigation path from the parking space where the vehicle is located to the pick-up location, and sends it to the vehicle-side unit; The APP unit is set on the mobile terminal, and is registered with the real name system and bound to the license plate number; the user sends a parking request or a vehicle pickup request through the operation interface of the APP unit.

2. The parking lot autonomous valet parking system based on intelligent network connection technology according to claim 1 is characterized in that: The vehicle-end unit and the field-end unit communicate with each other in a V2X manner.

3. The parking lot autonomous valet parking system based on intelligent network connection technology according to claim 1 is characterized in that: The vehicle information includes vehicle size data and a license plate number.

4. The parking lot autonomous valet parking system based on intelligent network connection technology according to claim 1 is characterized in that: The sensing device includes a camera and a radar.

5. The parking lot autonomous valet parking system based on intelligent network connection technology according to claim 1 is characterized in that: The field end unit converts the position information of the moving obstacle into the coordinate information of each nearby vehicle in the self-vehicle coordinate system, and converts the position information of each vehicle into the global coordinate information of the parking lot, and sends it to each vehicle to achieve obstacle avoidance and positioning of the vehicle.

6. The parking lot autonomous valet parking system based on intelligent network connection technology according to claim 1 is characterized in that: The cloud unit obtains the vehicle flow data of the day for iterating the vehicle flow prediction model.

7. The parking lot autonomous valet parking system based on intelligent network connection technology according to claim 1 is characterized in that: The cloud unit also predicts the traffic volume at a designated location based on time nodes of the day, week, month, and year.

8. The parking lot autonomous valet parking system based on intelligent network connection technology according to claim 1 is characterized in that: The APP unit includes a map display module for displaying a parking navigation path or a vehicle pickup navigation path.

9. The parking lot autonomous valet parking system based on intelligent network connection technology according to claim 1 is characterized in that: The APP unit is used to locate the vehicle or obtain vehicle monitoring images.

10. A parking method for an autonomous valet parking system in a parking lot based on intelligent network connection technology according to any one of claims 1 to 9, characterized in that: The steps include: Step 1: The driver checks the available parking spaces in the parking lot near the destination through the APP unit, selects the target parking lot and drives to the drop-off area of ​​the target parking lot; Step 2: The driver sends a parking request through the APP unit, and the vehicle terminal unit sends the vehicle information to the cloud unit; Step 3: Select a target parking space according to the current position of the vehicle, generate a parking navigation path from the current position of the vehicle to the target parking space, and send it to the vehicle end unit; Step 4: The vehicle-side unit receives the target parking space and the parking navigation path from the current position to the target parking space sent by the cloud unit, and automatically drives to the target parking space according to the parking navigation path to complete the parking work; Step 5: The driver sends a pick-up request through the APP unit, and the cloud unit receives the pick-up request and the pick-up location sent by the APP unit, generates a pick-up navigation path from the parking space where the vehicle is located to the pick-up location and sends it to the vehicle-side unit; Step 6: The vehicle-end unit drives to the pick-up location in the delivery area according to the pick-up navigation route to complete the delivery.

Images