Disclosure of Invention
In view of this, a first aspect of the present invention is to provide a parking system, including: a memory, a sensor processing unit; the sensor is used for acquiring satellite positioning navigation information and vehicle surrounding environment information; the processing unit forms parking path information based on the satellite positioning navigation information and the vehicle surrounding environment information; the memory is used for storing the parking path information and pre-built parking path information;
the processing unit is used for processing the parking path information and/or the pre-built parking path information stored in the memory and automatically parking the vehicle based on a preset scene and/or a finished self-learning scene;
the preset scene is a scene setting for completing automatic parking of the vehicle based on the pre-built parking path information;
the completed self-learning scene is the scene setting of automatic parking completed by the vehicle based on the parking path information formed by the vehicle surrounding environment information.
Preferably, the vehicle drives to the target parking space from the expected future traffic point, and when the system collects 100% of the vehicle surrounding environment information, the parking path information is completed and stored in the memory; and the processing unit processes the parking path information in the memory, automatically parks the vehicle based on the completed self-learning scene, and marks the completed self-learning scene.
Preferably, in the self-learning scenario, if the system does not collect 100% of vehicle surrounding environment information, the memory temporarily stores incomplete parking path information, and the processing unit marks the incomplete self-learning scenario.
Preferably, when the system does not collect 100% of the information of the surrounding environment of the vehicle, the processing unit completes automatic parking based on the preset scene.
Preferably, the preset scene is a scene set for completing automatic parking based on satellite positioning navigation information acquired by the sensor or parking path information built in the system; the built-in parking path information is data information required by a parking path which is collected in advance.
Preferably, the system further comprises an obstacle detection unit, and when the vehicle is automatically parked, the system can pause or quit the automatic parking if the obstacle detection unit detects an obstacle.
Preferably, the sensor comprises a visual sensor.
A second aspect of the present invention provides a parking method using the above-described system to perform the steps of:
step S1: detecting whether pre-built parking path information is stored in the memory, if any scene of the preset scene and the self-learning scene is detected, directly entering step S3, and if not, entering step S2;
step S2: selecting a self-learning mode in the central processing unit, and driving a vehicle to a target parking space by a driver; in the driving process, the satellite positioning navigation information and the vehicle surrounding environment information acquired by the sensor complete parking path information and are stored in the memory; ending the parking;
step S3: when the processing unit only detects the preset scene, the vehicle completes automatic parking based on the pre-built parking path information of the memory; and when the processing unit detects the completed self-learning scene, the vehicle forms parking path information based on the vehicle surrounding environment information to complete automatic parking.
In a third aspect of the present invention, there is provided a vehicle recall system including: a memory, a sensor, and a processing unit; the sensor is used for acquiring satellite positioning navigation information and vehicle surrounding environment information; the system forming recall path information based on the satellite positioning navigation information and/or the vehicle ambient information; the memory is used for storing the recall path information and/or pre-built-in recall path information;
the processing unit processes the recall path information and/or the pre-built recall path information stored in the memory and automatically recalls based on a preset scene and/or a finished self-learning scene;
the preset scene is set for the scene that the vehicle completes automatic recall based on the pre-built recall path information;
the completed self-learning scenario is a scenario setting in which the vehicle completes an automatic recall based on recall path information formed including the vehicle surrounding environment information.
Preferably, when the vehicle runs from a parking space to a future vehicle transfer point, the system collects 100% of vehicle surrounding environment information to finish the recall path information and stores the recall path information in the memory; the processing unit processes the recall path information in the memory, automatically recalls based on the completed self-learning scenario, and marks the completion of the self-learning scenario.
Preferably, in the self-learning scenario, if the system does not collect 100% vehicle surrounding environment information, the memory temporarily stores incomplete recall path information, and the processing unit marks the incomplete self-learning scenario.
Preferably, when the system does not collect 100% of the vehicle surrounding environment information, the processing unit completes automatic recall based on the preset scene.
Preferably, the preset scene is a scene set for completing automatic parking based on satellite positioning navigation information acquired by the sensor or recall path information built in the system; the built-in recall path information is data information required by a recall path which is acquired in advance.
Preferably, the system further comprises an obstacle detection unit, and the system can suspend or exit the automatic recall if the obstacle detection unit detects an obstacle while the vehicle is in the automatic recall.
Preferably, the sensor comprises a visual sensor.
In a fourth aspect of the present invention, there is provided a vehicle recall method, characterized in that: using the system of any one of claims 7-13 to perform the steps of:
step S1: detecting whether a pre-built recall path information is stored in the memory, if any scene of the preset scene and the self-learning scene is detected, directly entering the step S3, and if not, entering the step S2;
step S2: sending information to a person to inform of the parking space position, and selecting a self-learning mode in the processing unit; driving the vehicle to drive from the parking space to a future vehicle transfer point by a driver; in the driving process, the satellite positioning navigation information and the vehicle surrounding environment information collected by the sensor finish recalling path information and are stored in the memory; the recall is finished;
step S3: when the processing unit only detects the preset scene, the vehicle completes automatic recall based on recall path information built in the system; when the processing unit detects the self-learning scene, the vehicle forms recall path information based on the vehicle surrounding environment information to complete automatic recall.
In a fifth aspect of the present invention, a vehicle driving system is provided, which includes the above vehicle parking system and the above vehicle recall system.
The invention is characterized by the following points but not limited to the following points:
(1) the system comprises a self-learning scene, and intelligent driving can be realized even if a parking area does not have a high-precision map; in the prior art, even though automatic driving is adopted, the GPS is often relied on, but for some scenes, such as underground garage environment, the GPS signal is weak, and automatic navigation cannot be accurately realized. Other sensors such as an inertial meter and the like cannot realize accurate automatic navigation in environments such as a ground library and the like due to the accuracy problem. The invention adopts the vision sensor to collect the surrounding environment of the vehicle and carries out positioning, thereby overcoming the technical problems.
(2) By adopting the self-learning map building method, a common map data collecting vehicle for building the map is not needed, and a vehicle entering the ground library is a map data collector, so that the convenience is greatly improved, and the map is preassembled when the vehicle enters the ground library.
(3) The parking path information adopts two sources, namely self-learning and pre-built-in, in addition, the environmental information around the vehicle and possible satellite positioning information are integrated, the advantages of the two are utilized, the diversification of the information sources is realized, and the high compatibility is realized;
(4) the parking and recalling system solves the problem that a driver needs to find a parking space, only needs to hand over or wait for a vehicle at a preset place, is driven fully automatically, and improves convenience. In the prior art, parking and recalling are not effectively integrated with a high-precision map, and particularly, automation of parking or recalling cannot be realized in a complex environment of an underground garage.
DETAILED DESCRIPTION OF EMBODIMENT (S) OF INVENTION
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the following embodiments and accompanying drawings. The exemplary embodiments and descriptions of the present invention are provided to explain the present invention, but not to limit the present invention.
A method for unmanned parking and recall mainly comprises the following steps:
a parking path autonomous learning method; a recall path autonomous learning method; an automatic parking method; an automatic recall method; a method of stopping automatic parking or recall; an obstacle detection method; provided is a parking space line detection method.
Each step will be described in detail below.
The parking path learning method in step 1 comprises the following steps:
1.1) the driver pauses from driving the vehicle to the desired future point of delivery.
1.2) selecting a parking self-learning mode on the vehicle-mounted terminal of the unmanned system.
1.3) clicking a 'start learning' button on the vehicle-mounted terminal of the unmanned system, enabling the unmanned system to enter a parking learning mode, and enabling the vehicle-mounted terminal to display a 'parking self-learning' prompt.
1.4) the driver drives the automobile to a target parking space at the speed of less than 10km/h, and the automobile is parked in the parking space line.
1.5) after the driver puts the vehicle into position and stops the vehicle stably, clicking to finish learning, finishing the parking learning mode by the unmanned system, and storing data collected by current learning; meanwhile, the vehicle-mounted terminal displays the learning progress of the unmanned system for parking in the current path. For example, a progress bar of 30% at the first "end learning" may be displayed; the progress bar at the time of "end of learning" for the second time is 60%, and the progress bar at the time of "end of learning" for the third time is 100%. In some embodiments 100% refers to all data information around the vehicle that is required to complete the auto-park path through self-learning. In some embodiments, a visual sensor is used, and the images acquired by it are used to form a map, which is important because the use of GPS is ineffective in some situations, such as an underground garage environment. Other types of sensors, such as an inertial meter, may not be accurate and may not be able to accurately achieve automatic parking. The step after forming the map further comprises locating, i.e. determining the exact position of the vehicle in the map, including the determination of the vehicle pose.
1.6) if the system does not acquire enough learning data in the secondary parking process, the vehicle-mounted terminal displays an option of storing the uncompleted parking path; if the driver selects this option, the unmanned system will store the current time data, flag the current parking path as "partially learned complete," and reuse it when learning the same parking path in the future.
1.7) if the path stored as 'partial learning completed' in the unmanned system is entered again, the unmanned system prompts an 'historical parking path continued learning' option at the vehicle-mounted terminal. If the driver selects this option, the drone system will repeat steps 1.3) -1.6). If the driver chooses to ignore this option, the drone system will delete the last learned data for this path stored in the system and repeat the steps 1.3) -1.6).
1.8) if the system collects enough learning data in the current parking process, the unmanned system prompts the completion of the parking path at the vehicle-mounted terminal, can use the autonomous parking function and prompts the driver to store the autonomous parking path option. If the driver selects this option, the unmanned system will store the current autonomous parking path. And finishing the parking path autonomous learning process.
Step 2, the method for autonomously learning the recall path comprises the following steps:
2.1) the driver drives the parking space of the vehicle parking lot by himself.
2.2) selecting a 'recall self-learning' mode on the vehicle-mounted terminal of the unmanned system.
2.3) clicking a 'start learning' button on the vehicle-mounted terminal of the unmanned system, enabling the unmanned system to enter a recall learning mode, and enabling the vehicle-mounted terminal to display a 'recall self learning in progress' prompt.
2.4) the driver drives the automobile to a future expected receiving place at the speed of less than 10 km/h.
2.5) after the driver stops the vehicle, clicking to finish learning, finishing the learning mode by the unmanned system, and storing data collected by current learning; and simultaneously, the vehicle-mounted terminal displays the learning progress recalled by the unmanned system under the current path. In some embodiments, the learning progress is displayed on the vehicle-mounted terminal in a progress bar mode. For example, a progress bar of 30% at the first "end learning" may be displayed; the progress bar at the time of "end of learning" for the second time is 60%, and the progress bar at the time of "end of learning" for the third time is 100%. In some embodiments 100% refers to all data information around the vehicle that is required to complete the auto-park path through self-learning. In some embodiments, a visual sensor is used, and the images acquired by it are used to form a map, which is important because the use of GPS is ineffective in some situations, such as an underground garage environment. Other types of sensors, such as an inertial meter, may not be accurate and may not be able to accurately achieve automatic parking. The step after forming the map further comprises locating, i.e. determining the exact position of the vehicle in the map, including the determination of the vehicle pose.
2.6) if the system does not acquire enough learning data in the process of the current recall, the vehicle-mounted terminal displays a 'storage unfinished recall path' option; if the driver selects this option, the drone system will store the current recall path data, mark the current recall path as "partially learned complete," and be reused for future studies of the same recall path.
2.7) if the route stored as "partial learning completed" in the unmanned system is re-entered, the unmanned system will prompt a "history recall route continue learning" option at the in-vehicle terminal. If the driver selects this option, the drone system will repeat steps 2.3) -2.6). If the driver chooses to ignore this option, the drone system will delete the last learned data for this path stored in the system and repeat the steps 2.3) -2.6).
2.8) if the system collects enough learning data in the current recall process, the unmanned system prompts the 'recall path completion, an autonomous recall function can be used' at the vehicle-mounted terminal, and prompts the driver to select an 'autonomous recall path storage' option. If the driver selects this option, the drone system will store the own recall path. The recall path autonomous learning process is completed.
And 3, dividing the automatic parking method into two scenes.
Scene one: and (3) a preset scene means that the path of the scene to be parked is built in the unmanned system in advance, and automatic parking can be performed without the need of a user to adopt the parking path autonomous learning method in the step 1 again. The preset scene is a scene setting for completing automatic parking of the vehicle based on the pre-built path information; advanced map-gathering specialized tools may be employed including map-gathering vehicles to draw detailed map of the property. These map of the garage are transmitted to the system of the vehicle when the vehicle enters the garage. The processing unit is used for positioning by calling peripheral information, particularly visual images, and obtaining vehicle postures and other operations.
Scene two: and (2) a self-learning scene, namely a path of a scene to be parked is not built in advance, and automatic parking can be performed only after the unmanned system fully learns the path by adopting the parking path autonomous learning method in the step (1).
In scenario two, for a certain parking path, if the unmanned system has finished learning the path, the automatic parking step is the same as scenario one. The method comprises the following steps:
3.1) the driver starts the "automatic parking function".
And 3.2) confirming that the parking path of the scene is stored in the system in advance by the unmanned system according to the satellite positioning navigation information and the surrounding environment information, and feeding back the parking path information to the driver for confirmation through the vehicle-mounted terminal.
3.3) the driver confirms that the scene is correct, selects a confirmation option on the vehicle-mounted terminal, and the system enters an automatic parking preparation state.
And 3.4) the driver gets off the vehicle, is connected with the unmanned system through the mobile phone client, and initiates an autonomous parking instruction.
And 3.5) after the unmanned system receives the automatic parking instruction, carrying out safety check on the vehicle, confirming that the vehicle can enter a driving state and an automatic parking operation state.
3.6) the unmanned system monitors the surrounding environment through a camera and calculates the driving path in the automatic parking operation process; parameters and captured information in the vehicle driving process are sent to the mobile phone client; a driver can check the running state of the unmanned system, various parameters of the vehicle and videos shot by the vehicle camera through the mobile phone client.
3.7) after the automatic parking system finishes parking, locking the vehicle and sending an automatic parking finishing signal to the mobile phone client; while transmitting the location of the vehicle to the mobile client.
And 4, the automatic recall method is also divided into two scenes.
Scene one: and (3) a preset scene means that the path of the scene to be recalled is built in the unmanned system in advance, and automatic recall can be performed without the user adopting the recall path autonomous learning method in the step (2) again.
Scene two: and (3) a self-learning scene, namely, a path of the scene to be recalled is not built in advance, and automatic recall can be performed only after the unmanned system fully learns the path by adopting the recall path autonomous learning method in the step (2).
In scenario two, for a certain recall path, if the unmanned system has completed learning the path, the automatic recall steps are the same as scenario one. The same steps are as follows:
4.1) a driver logs in the unmanned system through the mobile phone client, selects a scene that the driver wants to pick up the vehicle, and activates an automatic recall function.
And 4.2) the unmanned system confirms that the recall path of the scene is pre-stored in the system according to the satellite positioning navigation information and the surrounding environment information, and feeds back the recall path information to the driver for confirmation through the mobile phone client.
4.3) the driver confirms that the scene is correct, selects a 'confirmation' option on the mobile phone client, and the system carries out safety check on the vehicle, confirms that the vehicle can enter a driving state and enters an automatic recall running state.
4.4) the unmanned system monitors the surrounding environment through a camera and calculates a driving path in the automatic recall operation process; parameters and captured information in the vehicle driving process are sent to the mobile phone client; a driver can check the running state of the unmanned system, various parameters of the vehicle and videos shot by the vehicle camera through the mobile phone client.
4.5) the automatic recall system runs to a preset car taking place, and a double-flash indicator lamp is turned on; and simultaneously sending the position of the vehicle and a 'recall completion, waiting for vehicle receiving' instruction to the mobile client.
The above process includes special cases: for example, if a built-in recall path is not found, it can be understood that the same is true for the parking process, and the system sends out information to remind the driver. In some embodiments, the parking position of the vehicle is informed, so that the driver can find the vehicle to finish driving conveniently. Reminding the driver to switch to a non-automatic parking state in the parking process, and the like.
Step 5, the method for stopping automatic parking or recalling refers to the fact that an unmanned system is provided with an obstacle detection function, and when a driving area in front of a vehicle meets an obstacle in the parking or recalling process, the system can automatically perform stable braking, and the vehicle is guaranteed to be in a safe state. If the front obstacle is not cleared within a certain time, the system can remind a driver of parking for the valet or interruption of a recall function through the mobile phone terminal, and the position of the vehicle is sent to the mobile phone client terminal to remind the driver of returning to the vehicle for taking over; meanwhile, the vehicle can automatically turn on the double-flash indicator lamp to exit the automatic parking/recalling mode.