Parking emergency dispatching method, device and equipment based on multi-source triggering and vehicle

Abstract

The application provides a multi-source trigger-based parking emergency scheduling method, device, equipment and vehicle, and relates to the technical field of vehicle control. The method comprises the following steps: in response to a parking command request of autonomous parking of a vehicle in a parking lot, determining a digital map corresponding to the parking lot, performing abnormality detection based on vehicle sensor data to obtain a risk state signal; when the risk state signal indicates that there is a sudden risk in the current parking, interrupting the parking command request, and switching the scheduling priority of the vehicle from responding to the parking command request to responding to an emergency response request; in response to the emergency response request, searching for a reference motion trajectory for the vehicle to drive away from the parking lot in the digital map according to the current initial pose of the vehicle, so as to generate a driving control instruction to control the vehicle to drive away from the parking lot. The application is used to solve the problem that the existing automatic parking scheme has a single scheduling target and lacks an emergency response mechanism.

Description

Technical Field

[0001] This invention relates to the field of vehicle control technology, specifically to a parking emergency dispatch method, device, equipment, and vehicle based on multi-source triggering. Background Technology

[0002] With the widespread adoption of Level 3-4 autonomous driving technologies, automated parking (including valet parking, AVP) has become a core function for enhancing user experience. Existing automated parking solutions primarily focus on improving parking efficiency and multi-vehicle collaborative scheduling, resolving path conflicts through optimized trajectory planning algorithms or global state perception. However, these solutions suffer from fundamental flaws: First, their scheduling objectives are singular; all autonomous parking calculation logic serves only the vehicle's parking, failing to handle emergency scenarios such as passenger occupant malfunctions, external fires, or ground structure collapses. Second, they lack emergency response mechanisms; the automated parking command is set to the highest priority during parking, only triggering braking or warnings to stop parking in emergency situations, failing to mitigate the threats posed to the vehicle or occupants. Summary of the Invention

[0003] In view of this, it is necessary to provide a parking emergency dispatch method, device, equipment and vehicle based on multi-source triggering to solve the problem that existing automatic parking solutions have a single dispatch target and lack an emergency response mechanism.

[0004] To address the aforementioned problems, this invention provides a parking emergency dispatch method based on multi-source triggering, comprising: In response to a parking command request from a vehicle to autonomously park in a parking lot, the system determines the digital map corresponding to the parking lot and acquires data from onboard sensors. Anomaly detection is performed based on the data from the vehicle-mounted sensors to obtain a risk status signal; When the risk status signal indicates that there is a sudden risk in the current parking, the parking command request is interrupted, and the vehicle scheduling priority is switched from responding to the parking command request to responding to the emergency response request. In response to the emergency response request, a reference motion trajectory for the vehicle to leave the parking lot is retrieved from the digital map based on the vehicle's current initial pose. Based on the reference motion trajectory, a driving control command for the vehicle is generated, and the vehicle is controlled to leave the parking lot using the driving control command.

[0005] In one possible implementation, the onboard sensor data includes vehicle cabin monitoring data, external environment data, and its own state data. The step of performing anomaly detection based on the onboard sensor data to obtain a risk state signal includes: Based on the data from the vehicle-mounted sensors, cabin vital signs detection, environmental disaster detection, and ground anomaly detection are performed respectively to obtain detection results; Risk decisions are made based on the detection results to obtain the risk status signal.

[0006] In one possible implementation, the step of making a risk decision based on the detection result to obtain the risk status signal includes: When the detection value corresponding to each detection result is not greater than the corresponding preset detection threshold, it is determined that the current parking is in a normal state, and a corresponding risk state signal is generated. When the detection value corresponding to any detection result is greater than the corresponding preset detection threshold and the duration is greater than the preset time threshold, it is determined that there is a sudden risk in the current parking and a corresponding risk status signal is generated.

[0007] In one possible implementation, retrieving a reference motion trajectory for the vehicle to leave the parking lot from the digital map based on the vehicle's current initial pose includes: Retrieve multiple candidate safety exits for leaving the parking lot from the digital map; Multiple sets of candidate paths are determined from the current initial pose to each of the multiple candidate safety exits, wherein each set of candidate paths corresponds one-to-one with each of the multiple candidate safety exits; For each candidate path, calculate the collision cost of encountering obstacles, the path distance cost, and the smoothing cost of path curvature. The collision cost, path distance cost, and smoothing cost are weighted according to preset weights, and the weighted results are summed to obtain the cost function value of each candidate path. The candidate path with the smallest cost function value in each group of candidate paths is determined as the target path for each candidate safe exit. The kinematic feasibility of the target path for each candidate safety exit is verified, and the motion trajectory corresponding to the successfully verified target path is determined as the reference motion trajectory for the vehicle to leave the parking lot.

[0008] In one possible implementation, the kinematic feasibility of the target path is verified in the following way: When the target path is detected to meet the target conditions, the kinematic feasibility check is determined to be successful, and the target path is retained. When it is detected that the target path does not meet at least one of the target conditions, the kinematic feasibility check is determined to have failed, and the target path is deleted. The target conditions include: The curvature of the motion trajectory is less than or equal to the maximum curvature of the vehicle, where the maximum curvature of the vehicle is the maximum curvature allowed by the vehicle hardware when the vehicle is turning; The rate of change of the continuous steering trajectory is within the allowable rate of change when the vehicle steering actuator performs continuous steering; The distance between the target path's trajectory and the obstacle is greater than a preset safe distance threshold.

[0009] In one possible implementation, when controlling the vehicle to leave the parking lot via the driving control command, the method further includes: Acquire real-time external environment data collected by vehicle-mounted sensors, and perform obstacle detection based on the real-time external environment data; If the obstacle detection result indicates that there is an obstacle that will collide with the reference trajectory within a certain period of time in the future, then a local path to the safety exit in the reference trajectory is replanned based on the vehicle's current position. The driving control command is updated according to the local path, and the vehicle is controlled to continue leaving the parking lot based on the updated driving control command.

[0010] In one possible implementation, after controlling the vehicle to leave the parking lot via the driving control command, the method further includes: When the aforementioned emergency risk is the presence of vital signs in the cabin, a Level 1 warning is triggered. When the aforementioned emergency risk is a fire outside the vehicle, a Level 2 warning is triggered; When the aforementioned emergency risk is a risk of ground collapse where the vehicle is located, a Level 3 warning is triggered; The warning information of the first-level warning, the second-level warning, and the third-level warning is sent to the target system. The target system includes the online management system of the parking lot, the emergency response system linked to the parking lot, and the application terminal system corresponding to the vehicle owner. The warning priority of the second-level warning is higher than that of the third-level warning, but lower than that of the first-level warning.

[0011] In one possible implementation, after controlling the vehicle to leave the parking lot via the driving control command, the method further includes: When any one of the target events is triggered, the sudden risk is determined to be resolved; The target event includes: The vehicle has left the parking lot via the aforementioned driving control command; The vehicle receives a risk clearance signal from the outside world; The vehicle owner sends a reset command to the vehicle after confirming that it is safe.

[0012] The present invention also provides a parking emergency dispatching device based on multi-source triggering, comprising: The parking initialization module is used to respond to parking command requests from vehicles to autonomously park in parking lots, determine the digital map corresponding to the parking lot, and acquire onboard sensor data. A multi-source anomaly perception module is used to detect anomalies based on the vehicle-mounted sensor data and obtain a risk status signal; The dynamic switching module is used to interrupt the parking command request and switch the vehicle's scheduling priority from responding to the parking command request to responding to the emergency response request when the risk status signal indicates that there is a sudden risk in the current parking. An emergency route planning module is used to respond to the emergency response request by retrieving a reference trajectory for the vehicle to leave the parking lot from the digital map based on the vehicle's current initial position. The vehicle driving control module is used to generate driving control commands for the vehicle based on the reference motion trajectory, and to control the vehicle to leave the parking lot through the driving control commands.

[0013] The present invention also provides an electronic device, including a memory and a processor, wherein the memory is used to store a program; the processor is coupled to the memory and is used to execute the program stored in the memory to implement the steps of the above-described parking emergency dispatch method based on multi-source triggering.

[0014] The present invention also provides a vehicle, including an on-board sensor, an on-board device, and the aforementioned electronic device, wherein the electronic device, the on-board sensor, and the on-board device all communicate via a vehicle network; the on-board sensor is used to collect on-board sensor data and store the sensor data in the memory of the electronic device; the on-board device is used to obtain driving control commands for the vehicle after the processor of the electronic device executes the program stored in the memory, and control the vehicle to leave the parking lot through the driving control commands.

[0015] The beneficial effects of the above implementation method are as follows: The parking emergency dispatch method, device, equipment, and vehicle based on multi-source triggering provided by this invention, when executing a parking command request for autonomous parking, utilize onboard sensor data for anomaly detection to identify risk status signals. Upon determining that a sudden risk exists during parking, an emergency response mechanism is activated, proactively triggering a switch in vehicle dispatch priority. This allows for the execution of an emergency response request for vehicle escape, interrupting parking. This ensures that the computational logic of autonomous parking not only serves to complete vehicle parking but also allows for flexible emergency handling through dispatch priority switching when risks are detected, rather than passively waiting for a stop. Through the emergency response mechanism, the core task objective of the vehicle is switched from completing parking to safely departing, enabling the vehicle to automatically enter a preset, safer state under risk conditions. This forms a closed-loop safety mechanism, ensuring the safety of the vehicle and its occupants, and is more conducive to rescue operations compared to passively waiting for a stop. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating the parking emergency dispatch method based on multi-source triggering provided by the present invention. Figure 2 A schematic diagram illustrating the architecture of the emergency dispatch process provided by this invention; Figure 3 This is a schematic diagram of the vehicle priority-based scheduling process provided by the present invention; Figure 4 This is a schematic diagram of the parking emergency dispatch device based on multi-source triggering provided by the present invention; Figure 5 A schematic diagram of an embodiment of the electronic device provided by the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0019] In the description of the embodiments of this application, unless otherwise stated, "a plurality of" means two or more.

[0020] In this embodiment of the invention, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, apparatus, product or device that includes a series of steps or modules is not necessarily limited to those steps or modules that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to such process, method, product or device.

[0021] The naming or numbering of steps in the embodiments of the present invention does not mean that the steps in the method flow must be executed in the time / logical order indicated by the naming or numbering. The execution order of the named or numbered process steps can be changed according to the technical purpose to be achieved, as long as the same or similar technical effect can be achieved.

[0022] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0023] The parking emergency dispatch method based on multi-source triggering provided in this invention can be applied to emergency parking scenarios. The executing entity can be the vehicle's onboard system, or various servers, terminals, or remote cloud devices that communicate with the onboard system via the vehicle network. When the vehicle owner sends a parking command request to the vehicle for autonomous parking in a parking lot through the automatic parking function, the method provided by this invention is invoked to complete the emergency parking dispatch. Finally, the vehicle is controlled to leave the parking lot through driving control commands. This ensures that in the event of a sudden risk during parking, the vehicle can effectively avoid the sudden risk and automatically enter a preset safer state, ensuring the safety of the vehicle and its occupants, and facilitating rescue operations.

[0024] Figure 1 This is a flowchart illustrating the parking emergency dispatch method based on multi-source triggering provided by the present invention, as shown below. Figure 1 As shown, the parking emergency dispatch method based on multi-source triggering can be implemented through the following steps 101 to 105, which are explained in detail below.

[0025] Step 101: In response to the parking command request for the vehicle to autonomously park in the parking lot, determine the digital map of the parking lot and acquire onboard sensor data.

[0026] Here, when the vehicle owner uses the automatic parking function to autonomously park in a parking lot, the application terminal system sends a parking command to the vehicle. In response to this command, the system performs parking initialization, initiating the parking operation. During the parking process, a digital map of the parking lot is acquired for route planning, while continuous monitoring of emergency triggering sources is maintained, specifically by acquiring data from onboard sensors.

[0027] The digital map is not a regular navigation map, but rather a structured data set containing semantic information with centimeter-level precision, denoted as an HD Map. The digital map includes parking space coordinates and attributes (availability, charging location), geometric parameters of lane and traffic area boundaries, precise locations of all parking lot exits and entrances, locations of ramps and speed bumps, and location labels for key safety facilities such as fire hydrants, fire lanes, and emergency exits, all used for route planning.

[0028] like Figure 2 As shown, in the perception layer, when responding to a parking command request for autonomous parking of a vehicle in a parking lot, the parking command is parsed and verified. After confirming that the parking command is valid, the corresponding HD Map of the parking lot layer is first obtained from local storage or the cloud and loaded.

[0029] The vehicle-mounted sensor data includes cabin monitoring data, external environmental data, and the vehicle's own status data. At the perception layer, this vehicle-mounted sensor data is collected in real time through cabin vital signs sensor groups, environmental disaster sensor groups, and vehicle status sensor groups. The cabin vital signs sensor group includes vital signs monitoring sensors installed in the cabin, UWB anchor points, millimeter-wave bio-radar, and infrared thermal imaging modules. The environmental disaster sensor group includes smoke sensors, carbon monoxide sensors, and triaxial vibration accelerometers. The vehicle status sensor group includes sensors controlled by the controller area network that measure vehicle speed, yaw rate, steering wheel angle, brake master cylinder pressure, and the locking status of doors and trunk.

[0030] During parking initialization, the onboard sensors that collect data need to perform a self-test. When the self-test passes, the initial motion trajectory from the vehicle's current position to the parking lot is planned on the digital map.

[0031] Specifically, after the vehicle performs a parking operation and determines the corresponding digital map of the parking lot, the vehicle begins the parking task. During the parking initialization process, a sensor self-check and initialization procedure needs to be performed on the vehicle. The sensor self-check involves self-testing the onboard sensors that collect data. For example, on one hand, the surround-view camera, millimeter-wave radar, and ultrasonic radar used for environmental perception are calibrated and zero-point calibrated to ensure that the coordinate system of the fused perception is consistent. On the other hand, the vital signs monitoring sensors, ultra-wideband (UWB) anchor points, millimeter-wave bio-radar, and infrared thermal imaging module in the cockpit undergo power-on self-tests to confirm that their operating frequency and sensitivity parameters are within the normal range. On the other hand, the smoke sensor, carbon monoxide sensor, triaxial vibration accelerometer, and other environmental disaster sensors undergo zero-point calibration and sensitivity testing. The calibration error of all sensors must be controlled within ±0.5° or an equivalent level of accuracy. If any critical sensor, such as the forward millimeter-wave radar used for obstacle detection, fails its self-test, the system will not enter a highly automated parking mode, or will only operate in a degraded mode after using redundant sensors, such as surround-view cameras and ultrasonic radar, to replace the forward radar function to achieve the minimum safety performance.

[0032] Once it is confirmed that all onboard sensors meet the corresponding requirements, the self-test is considered passed, and the vehicle enters a highly automated parking mode to execute the parking task. During this process, an initial trajectory is planned on the digital map from the vehicle's current pose to the center of the target parking space. Specifically, starting from the vehicle's current pose and ending at the center of the target parking space, and combining information on static obstacles such as pillars and walls from the HD Map with preliminary information on dynamic obstacles such as other moving vehicles and pedestrians sensed in real time, a path search algorithm based on the vehicle's kinematics model is used to generate an initial trajectory consisting of a series of discrete pose points from the starting point to the ending point. This initial trajectory must satisfy the vehicle's nonholonomic constraints, such as the minimum turning radius, and ensure that each point on the trajectory maintains a safe distance margin of at least 0.5 meters from any known static obstacle.

[0033] Once the initial trajectory is determined, it can be converted into a corresponding driving control command to control the vehicle to drive from its current position into the center of the target parking space in the parking lot, thereby completing the parking task.

[0034] Step 102: Perform anomaly detection based on vehicle sensor data to obtain risk status signals.

[0035] When a vehicle responds to a parking command request and performs a parking task, it continuously performs anomaly detection using real-time acquired onboard sensor data. The purpose of anomaly detection is to detect whether there are any risk sources in the current vehicle, and then make risk decisions based on the risk sources to obtain risk status signals.

[0036] In one possible implementation, anomaly detection based on vehicle sensor data is used to obtain risk status information. This can be achieved in the following ways, which are explained in detail below.

[0037] Based on data from vehicle-mounted sensors, cabin vital signs detection, environmental disaster detection, and ground anomaly detection are performed to obtain detection results. Risk decisions are made based on the detection results to obtain the risk status signal.

[0038] The vehicle's sensor data includes cabin monitoring data, external environmental data, and its own status data. During anomaly detection, it continuously scans and assesses three distinct risk sources: vital signs within the cabin, environmental hazards, and ground anomalies. The first channel is cabin monitoring data, comprised of channel impulse response signals generated by UWB ultra-wideband anchor arrays deployed in the roof or A-pillars, micro-Doppler spectrum maps generated by millimeter-wave bio-radar in the 79GHz band, and temperature distribution matrices generated by infrared thermal imaging cameras. The second channel is external environmental data, including the percentage of smoke concentration measured by a laser particle counter, carbon monoxide concentration measured by an electrochemical sensor, and ground triaxial acceleration values ​​measured by a high-precision MEMS vibration sensor. The third channel is vehicle status data, acquired in real-time via the Controller Area Network (CAN) bus, including vehicle speed, yaw rate, steering wheel angle, brake master cylinder pressure, and door and trunk locking status.

[0039] For the above three data streams, during risk source detection, cabin vital signs detection, environmental disaster detection, and ground anomaly detection are performed independently and separately.

[0040] For cabin vital sign detection, the system first preprocesses the UWB CIR signal to extract features reflecting the periodic fluctuations of the chest cavity in a living being. Simultaneously, millimeter-wave radar spectral data is used to analyze micro-motion features caused by breathing and heartbeat, while thermal imaging data is used to locate and segment heat source contours within the cabin. These three types of information are fed into a feature fusion classifier based on a deep neural network. This classifier, trained on massive amounts of data from various scenarios including adults, children, and pets, can distinguish with extremely high confidence between the presence of a living being inside the vehicle and the absence of moving inanimate objects, while minimizing false alarms caused by interference sources such as swaying hanging clothing or air conditioning vents. For environmental disaster detection, the processing logic is relatively straightforward but the threshold requirements are strict. By continuously monitoring smoke and carbon monoxide concentrations, a fire risk warning is triggered when either value exceeds a preset safety threshold within multiple consecutive sampling periods—for example, smoke concentration > 0.1% vol or carbon monoxide concentration > 50 ppm. To address the risk of ground collapse, vibration sensors monitor abnormal ground vibration patterns. Combined with pre-recorded data on the parking lot's inherent frequency characteristics, the amplitude and spectrum of the vibration signals are analyzed to determine if there are structural risks caused by external construction, geological activity, or explosions in the vicinity. Before entering the decision logic, all raw sensor data undergoes a spatiotemporal synchronization and fusion filter based on an extended Kalman filter (EKF) or particle filter. This aligns the spatiotemporal differences caused by installation locations and sampling delays of different sensors and filters out random noise, thereby improving detection accuracy and robustness, and enhancing the overall accuracy of the fused detection.

[0041] Ultimately, the anomaly detection results include cockpit vital sign detection results obtained by performing cockpit vital sign detection on cockpit monitoring data, environmental disaster detection results obtained by performing environmental disaster detection on external environmental data, and ground anomaly detection results obtained by performing ground anomaly detection on its own status data.

[0042] Furthermore, risk decisions are made based on the detection results to obtain risk status signals. This step determines whether the current parking situation is normal or if there is an emergency risk based on the detection results, and then generates a corresponding risk status signal based on the vehicle's status.

[0043] like Figure 2 As shown, after the perception layer uses on-board sensors to detect risk sources, the decision layer performs dynamic priority scheduling, multi-objective path planning, and linkage early warning based on the detection results.

[0044] First, risk decisions need to be made based on the detection results of risk sources to determine whether there are three types of risks when the vehicle is parked: life signs in the cabin, environmental disasters, and ground anomalies. Then, a risk status signal is generated to adjust the vehicle's scheduling priority.

[0045] In one possible implementation, risk decisions are made based on the detection results to obtain a risk status signal. This can be achieved through the following methods, which are explained in detail below.

[0046] When the detection value corresponding to each detection result is not greater than the corresponding preset detection threshold, it is determined that the current parking is in a normal state, and a corresponding risk state signal is generated. When the detection value corresponding to any detection result is greater than the corresponding preset detection threshold and the duration is greater than the preset time threshold, it is determined that there is a sudden risk in the current parking and a corresponding risk status signal is generated.

[0047] Here, the detection results include cockpit vital sign detection results obtained by detecting cockpit vital signs from cockpit monitoring data, environmental disaster detection results obtained by detecting environmental disasters from external environmental data, and ground anomaly detection results obtained by detecting ground anomalies from self-status data.

[0048] Specifically, the cabin vital signs detection result is the confidence level of the presence of vital signs within the cabin. When this confidence level exceeds a preset confidence threshold, such as 90%, a risk to life in the cabin can be determined. For example, the confidence level for a child trapped in a vehicle is 95%. Conversely, if the confidence level is lower, there is no risk to life in the cabin. Environmental disaster detection results include smoke concentration and carbon monoxide concentration. When the smoke concentration exceeds a preset first concentration threshold and / or the carbon monoxide concentration exceeds a preset second concentration threshold, an external fire risk can be determined. Conversely, there is no external fire risk. Ground anomaly detection results include the amplitude or spectrum of vibration signals from the ground where the vehicle is located. When the amplitude or spectrum exceeds the corresponding amplitude or spectrum value threshold, a risk of ground collapse can be determined. Conversely, there is no risk of ground collapse.

[0049] When making risk decisions, if at least one of the following is detected: vital signs in the cabin, external fire risk, or ground collapse risk, then a sudden risk is identified for the current parking situation, and a corresponding risk status signal is generated. If none of the three sudden risks are detected, then the current parking situation is determined to be normal, and a corresponding risk status signal is generated.

[0050] Furthermore, during risk decision-making, the detection processes for the three risk sources are parallel and independent. When risk decisions are required, the results of the three detections are collected simultaneously for judgment. The priority of the judgments is preset from high to low as cabin life risk, external fire risk, and structural collapse risk, which facilitates early warning operations at different levels.

[0051] In this embodiment of the invention, when a vehicle is performing a parking task, risk source perception is performed in real time by acquiring onboard sensor data, which can capture sudden risk signals in real time, thereby providing a reliable basis for vehicle scheduling priority.

[0052] Step 103: When the wind status signal indicates that there is a sudden risk to the current parking, interrupt the parking command request and switch the vehicle dispatch priority from responding to the parking command request to responding to the emergency response request.

[0053] like Figure 3 As shown, Figure 3 The process of vehicle priority scheduling is demonstrated. After responding to the parking command request of the vehicle to park autonomously in the parking lot, parking initialization is performed. Then, the on-board sensors are used to continuously monitor risk sources and generate risk status signals through risk decision-making.

[0054] Furthermore, risk assessment is performed based on risk status signals, specifically determining whether a sudden risk exists. If not, it indicates that the current vehicle parking is in a normal state, and the vehicle can continue to perform the parking task. Then, it is assessed whether the vehicle has reached the safe area, and the process ends once the vehicle has reached the safe area and completed parking. If so, it indicates that a sudden risk exists for the current vehicle parking. When the risk status signal indicates a sudden risk for the current parking, a dynamic vehicle priority switch is executed, interrupting the parking command request and switching the vehicle's scheduling priority from responding to parking command requests to responding to emergency response requests.

[0055] Specifically, once a risk status signal indicates a sudden risk to the current parking maneuver, the task scheduler immediately sends a high-priority "soft interrupt" or "task suspension" command to all lower-level controllers responsible for performing routine parking trajectory tracking—steering controller, drive controller, and braking controller. This command forcibly suspends any ongoing parking task, whether it's the initial parking search, steering adjustments during reversing into the parking space, or final fine-tuning of the parking space. Furthermore, the vehicle's scheduling priority changes from parking to departure, switching from responding to parking command requests to responding to emergency response requests. Thus, the vehicle's priority is no longer rigidly executing parking tasks, but rather responding to emergency response requests and executing departure tasks.

[0056] The change in vehicle scheduling priority atomically modifies the core task objective in the vehicle's global state machine from "Task_Goal_ParkIn" to "Task_Goal_EmergencyEgress". This modification is not merely changing the value of a software variable; it means that the computing power and bandwidth originally allocated to modules such as path replanning and parking optimization for parking tasks will be released, and the message priorities on the communication bus will be reconfigured to ensure that all instructions and sensor data related to emergency departure have the highest transmission priority and are instead used for departure tasks. Furthermore, a dedicated emergency resource pool will be activated in response to emergency response requests, and computing resources will be pre-allocated for upcoming operations such as escape route planning and high-precision positioning.

[0057] After the risk status signal indicates that there is a sudden risk in the current parking and interrupts the parking command request, the parking task is interrupted because the vehicle scheduling priority has been adjusted. Therefore, a task interruption snapshot of the vehicle relative to the initial movement trajectory can be generated.

[0058] The sole purpose of a mission interruption snapshot is to provide a vehicle with the possibility of resuming parking from the point of failure after an emergency event involving a sudden risk has been completely resolved and confirmed by human intervention. For example, in some special and possible situations, the detection results of the risk source may be instantaneous, such as a short-term collapse risk caused by temporary construction in a parking lot. In such cases, after the sudden risk has been completely resolved and confirmed by human intervention, the vehicle can resume parking at the point of failure.

[0059] Step 104: In response to the emergency response request, retrieve a reference motion trajectory for the vehicle to leave the parking lot from the digital map based on the vehicle's current initial pose.

[0060] See also Figure 3 After the vehicle priority dynamic switching is completed, in response to the emergency response request, the vehicle activates the emergency control mode and obtains the vehicle's current motion state context, including but not limited to: the vehicle's current precise position, fused GNSS, IMU and wheel speedometer information, heading angle, real-time vehicle speed, yaw rate, and the current state of the conventional parking task actuators, such as the steering motor target angle, drive torque value, braking pressure value, etc. The vehicle's current motion state context is used to perform departure path planning in emergency control mode.

[0061] In emergency control mode, escape path planning needs to devise a "safe, fast, and feasible" escape route for the vehicle. Path planning first requires determining the vehicle's current initial pose, that is, the vehicle's pose information when the parking command request was interrupted, including the vehicle's current precise position (x, y) and heading angle θ. Secondly, it requires a digital HD map from the parking lot, especially a list of coordinates and current status of all safety exits (including regular exits, fire exits, and connecting ramps). Furthermore, the path planning decision also needs to consider real-time point cloud or map information of dynamic and static obstacles around the vehicle, as well as the vehicle's current kinematic parameters, such as wheelbase, minimum turning radius, maximum steering angle, and vehicle body dimensions.

[0062] During specific path planning, the reference trajectory for the vehicle to leave the parking lot is retrieved from the digital map based on the vehicle's current initial pose. Since there are multiple safety exits in the parking lot, several different candidate paths can be planned. By combining factors such as obstacles, path distance, and path curvature, the optimal reference trajectory can be selected from the candidate paths for the vehicle to leave the parking lot.

[0063] In one possible implementation, the reference motion trajectory for the vehicle to leave the parking lot is retrieved from the digital map based on the vehicle's current initial pose. This can be achieved in the following ways, which are explained in detail below.

[0064] Search for multiple candidate safety exits for leaving the parking lot on the digital map; Determine multiple sets of candidate paths from the current initial pose to multiple candidate safety exits; For each candidate path, calculate the collision cost of encountering obstacles, the path distance cost, and the smoothing cost of path curvature. The collision cost, path distance cost, and smoothing cost are weighted according to preset weights, and the weighted results are summed to obtain the cost function value of each candidate path. The candidate path with the smallest cost function value in each group of candidate paths is determined as the target path for each candidate safe exit. The kinematic feasibility of the target path for each candidate safety exit is verified, and the motion trajectory corresponding to the successfully verified target path is determined as the reference motion trajectory for the vehicle to leave the parking lot.

[0065] Specifically, during path planning, a fast global exit search is first performed on the digital map. Taking the vehicle's current location as the center, all reachable safety exits within a certain radius are searched in the HD Map, and a preliminary sorting is performed based on geometric straight-line distance and path complexity to generate a candidate safety exit list.

[0066] Then, multiple sets of candidate paths are determined from the current initial pose to multiple candidate safety exits, wherein each set of candidate paths corresponds one-to-one with each candidate safety exit in the multiple candidate safety exits.

[0067] This paper employs an improved multi-objective decision optimization algorithm to plan a preliminary target path for each of the top-ranked candidate safety exits. The planning is based on the cost of adapting to emergency scenarios. Traditional parking path planning cost functions may primarily optimize for the shortest path length or the fewest turns. However, since each candidate safety exit corresponds to multiple candidate paths, the cost function value of each candidate path is first calculated and compared to select the target path corresponding to each candidate safety exit. In this embodiment, the cost function value F_total of the candidate path is constructed as the sum of three weighted sub-costs, expressed as:

[0068] Here, C_collision is the collision cost, representing the cost of hitting an obstacle. It is mapped from the distance from each point on the candidate path to the nearest obstacle. The closer the distance, the higher the cost. Its weight w1 is set to 0.6, making it absolutely dominant. This ensures that path planning prioritizes routes away from obstacles under all circumstances. C_distance is the distance cost, representing the cost of path distance. It is proportional to the remaining length of the candidate path to the candidate safe exit. Its weight w2 is set to 0.25, guiding the vehicle to evacuate efficiently. C_curvature is the curvature cost, representing the cost of smoothing the path curvature. It reflects the smoothness of the candidate path. Overly sharp turns can lead to difficulty in control or even instability. Its weight w3 is set to 0.15, ensuring the vehicle's stability and controllability. Of course, the weight settings are not fixed. For example, weight w1 can be set to more than 0.6, while w2 can be set to slightly less than 0.25, and weight w3 can be adjusted accordingly. The sum of the three weights is 1.

[0069] After the path cost evaluation is completed, one or more candidate paths with the smallest cost function value F_total are selected from these candidate paths as the target path for the corresponding candidate security exit. Thus, for each candidate security exit, a target path can be determined from a set of candidate paths.

[0070] Finally, the kinematic feasibility of each candidate safety exit path is verified to check whether the path can safely avoid obstacles and whether it meets the vehicle's turning and driving conditions, in order to filter the target paths. Ultimately, the motion trajectory corresponding to the successfully verified target path is determined as the reference motion trajectory for the vehicle to leave the parking lot.

[0071] This invention, in an emergency control mode, rapidly executes exit path planning and selects the safest, shortest, and most stable target path by calculating the cost function value of the planned path. This allows the vehicle to leave the parking lot as safely and quickly as possible, reducing the impact of risk sources and minimizing losses, thus solving the problem of dynamic path generation in emergency exit scenarios. Furthermore, the cost function value prioritizes obstacle avoidance safety, effectively ensuring the vehicle's ability to escape from narrow and chaotic environments and maintaining driving stability, significantly reducing secondary risks.

[0072] In one possible implementation, the kinematic feasibility of the target path is verified in the following manner, which is explained in detail below.

[0073] When the target path is detected to meet the target conditions, the kinematic feasibility check is successful, and the target path is retained. If the target path is found to fail to meet at least one of the target conditions, the kinematic feasibility check is determined to have failed, and the target path is deleted.

[0074] See also Figure 3 After planning the departure route, for each candidate safety exit in the candidate safety exit list, it is necessary to determine whether the target path passes the kinematic feasibility check. If it does, meaning the kinematic feasibility check is successful, the target path for that candidate safety exit is retained, and the corresponding motion trajectory can be determined as the reference motion trajectory for the vehicle to leave the parking lot, used for emergency departure. If it does not, meaning the kinematic feasibility check fails, the target path for that candidate safety exit is deleted and filtered out. Then, departure route planning can be re-executed, and the path feasibility check can be performed again to determine whether it passes the kinematic feasibility check.

[0075] The target conditions include: the curvature of the motion trajectory is less than or equal to the maximum curvature of the vehicle, which is the maximum curvature allowed by the vehicle hardware when the vehicle is turning; the continuous steering change rate of the motion trajectory is within the allowable change rate range when the vehicle steering actuator performs continuous steering; and the distance between the motion trajectory of the target path and the obstacle is greater than the preset safe distance threshold.

[0076] During the actual verification process, the verifier simulates the vehicle traveling along the target path of the candidate safety exit, checks whether the path curvature is always less than or equal to the maximum curvature physically allowed by the vehicle (i.e., the reciprocal of the minimum turning radius), and checks whether the rate of change of the continuous steering angle of the target path is within the capability range of the steering actuator, that is, whether it is within the allowable rate of change range when the vehicle steering actuator continuously turns.

[0077] In addition, the vehicle contour model can be used to sweep along the target path to ensure that, with a safety margin (such as 0.8 meters), the vehicle will not interfere with the predicted trajectory of static obstacles in the HD Map or dynamic obstacles perceived in real time during the entire driving process. That is, the distance between the vehicle and the obstacle on the target path must not be less than the preset safe distance threshold (such as 0.8 meters), thereby ensuring that the vehicle maintains a certain safe distance from the obstacle and avoids collision.

[0078] The process of kinematic feasibility verification needs to be completed in a very short time to minimize the impact of risk sources on the vehicle and its occupants.

[0079] This invention provides a multi-objective path optimization mechanism through kinematic feasibility verification, which can further filter out kinematically feasible paths based on the target paths, ensuring the reliability and safety of path planning.

[0080] Step 105: Generate driving control commands for the vehicle based on the reference motion trajectory, and control the vehicle to leave the parking lot using the driving control commands.

[0081] The reference trajectory is a validated and feasible spatiotemporal trajectory consisting of a series of sequential pose points (x, y, θ, k, v) and the corresponding target exit number. The trajectory includes not only the position (x, y) and heading angle θ, but also the desired path curvature k and recommended driving speed v, providing a clear reference for vehicle control.

[0082] like Figure 2 As shown, after the decision-making layer completes multi-objective path planning, the execution layer can then be entered to implement vehicle control. During vehicle control, the digitized reference motion trajectory needs to be converted into actual vehicle motion commands, including steer-by-wire, brake-by-wire, and power control. This conversion process also requires the use of the vehicle's current motion state context, including but not limited to: the current state of the regular parking task actuator when the parking task is interrupted, such as the steering motor target angle, drive torque value, and braking pressure value.

[0083] Based on the reference trajectory and the vehicle's current real-time status, a set of control commands can be calculated to minimize the error between the vehicle's actual trajectory and the reference trajectory. For lateral control, a combination of feedforward and feedback strategies is typically used. The feedforward part calculates the desired steering angle based on the path curvature, while the feedback part eliminates lateral position and heading deviations caused by model errors and external disturbances. For longitudinal control, speed PID control or more advanced algorithms are used to control the vehicle speed by coordinating the drive motor and braking system, based on the recommended speed v at each point on the path. Throughout the entire departure process, the vehicle's maximum speed is strictly limited to below 5 km / h. This speed is determined comprehensively based on the typical spatial dimensions (column spacing, passage width) and visibility conditions of an underground parking garage, aiming to balance escape efficiency and safe emergency braking distance.

[0084] Upon receiving a driving control command, the vehicle can be controlled to leave the parking lot along the reference trajectory and at the recommended speed, thereby reaching a safe area and completing emergency avoidance.

[0085] In one possible implementation, when controlling the vehicle to leave the parking lot via driving control commands, the method further includes: acquiring real-time external environment data collected by onboard sensors and performing obstacle detection based on the real-time external environment data; If the obstacle detection result indicates that there is a collision between an obstacle and the reference trajectory within a certain period of time in the future, a local path to the safe exit in the reference trajectory is replanned based on the vehicle's current position. The driving control instructions are updated based on the local path, and the vehicle is controlled to continue leaving the parking lot based on the updated driving control instructions.

[0086] Specifically, during the vehicle's controlled departure process, the system continuously senses the external environment, acquires real-time environmental data from onboard sensors, and performs obstacle detection based on this data. For example, it continuously performs collision detection by comparing the latest perception results, such as LiDAR point clouds or visual detection boxes, with the vehicle's reference trajectory. If a collision risk with a newly appearing dynamic obstacle is detected within a short timeframe (e.g., within 2 seconds) while the vehicle is following the current driving control commands, a local replanning is immediately triggered. Based on the vehicle's current position, a local path to the safe exit in the reference trajectory is replanned.

[0087] During the replanning process, starting from the vehicle's current state, a large number of feasible short trajectories are sampled in the velocity-steering angle space. The safety, proximity to the global target, and smoothness of each trajectory are quickly evaluated, and a new, safe local path to the safe exit in the reference trajectory is selected within tens of milliseconds. Furthermore, the driving control commands are updated based on the selected local path, and then the vehicle is controlled to continue leaving the parking lot based on the updated driving control commands. This forms a closed-loop mechanism of planning, execution, perception, and replanning, ensuring the vehicle's driving safety in a dynamic and unknown environment with potential risks.

[0088] During vehicle control, logs are generated based on external environment perception and path replanning results to record the vehicle's departure process. Furthermore, the response latency of driving control commands should be less than 100 milliseconds, and the total latency from detecting a sudden obstacle to completing local path replanning and starting execution should be less than 500 milliseconds to achieve the fastest possible response.

[0089] In one possible implementation, after controlling the vehicle to leave the parking lot via driving control commands, the system also includes: triggering a Level 1 warning when a sudden risk is identified as the presence of vital signs in the cabin; When the sudden risk is a fire outside the vehicle, a level-two warning is triggered; When a sudden risk is identified as a risk of ground collapse where the vehicle is located, a Level 3 warning is triggered. The warning information for Level 1, Level 2, and Level 3 warnings is sent to the target system.

[0090] like Figure 2 As shown, a linkage early warning system is also implemented at the decision-making level. The linkage early warning system can generate hierarchical early warning information and send it to the external linkage layer. The external linkage layer is the target system that needs to be conveyed to by the hierarchical early warning information, specifically including the online management system of the parking lot, the emergency response system linked with the parking lot, and the application terminal system corresponding to the vehicle owner.

[0091] Specifically, the present invention provides an early warning mechanism in its embodiments, see below. Figure 3 After emergency departure is initiated—that is, after the vehicle is controlled to leave the parking lot via driving control commands—a tiered warning system is implemented. Different levels of warning are issued based on different types of emergencies. These emergencies include cabin life risk, external fire risk, and structural collapse risk. Different risks have different decision priorities: cabin life risk has a higher decision priority than external fire risk, and external fire risk has a higher decision priority than structural collapse risk. Therefore, the warning levels include Level 1, Level 2, and Level 3 warnings. The priority of Level 2 warning is higher than that of Level 3 warning but lower than that of Level 1 warning.

[0092] If the sudden risk is the presence of vital signs in the cabin, it indicates that there is a cabin life risk with the highest decision priority, and a first-level warning is triggered at this time. For example, when a child is trapped in the vehicle, a first-level risk with the highest warning priority is generated, and the warning information includes the type of risk event, the real-time position of the vehicle accurate to the parking space level, the key vital sign summary (such as respiratory rate) obtained by fusion of cabin sensors, the existence confidence, and the anonymized contact code of the corresponding vehicle owner.

[0093] If the sudden risk is the presence of a fire outside the vehicle, it indicates that there is an external fire risk with a medium decision priority, and a second-level warning is triggered at this time. For example, when a fire occurs, a second-level risk with a medium warning priority is generated, and the warning information includes the suspected fire position (i.e., the vehicle position), the smoke concentration and carbon monoxide concentration detected by the sensors, and a brief description of the surrounding environment of the vehicle.

[0094] If the sudden risk is the presence of a collapse risk on the ground where the vehicle is located, it indicates that there is a structural collapse risk with the lowest decision priority, and a third-level warning is triggered at this time. For example, when construction is carried out outside, a third-level risk with the lowest warning priority is generated, and the warning information includes the structural collapse position (such as the construction position, geological activity position, and explosion position), the amplitude and spectrum of the ground vibration signal, and a brief description of the occurrence of the ground structural risk.

[0095] After the warning information of each level is encrypted to generate a warning message data packet, it is synchronously broadcast and sent through the Internet of Vehicles communication stack to the online management system corresponding to the parking lot, the emergency response system linked with the parking lot, and the application program terminal system corresponding to the vehicle owner. These three target systems can receive the warning information at the same time, so as to realize the linkage between the parking lot, the vehicle owner and the third-party emergency, which is used for rapid notification and handling of sudden risks.

[0096] In the embodiments of the present invention, through the warning mechanisms of different levels, the emergency response status and intention of the parking lot and the vehicle are broadcast to the outside in the form of warning information. Compared with the delay of several minutes or even more than ten minutes relying on the traditional mobile network and manual alarm, the efficiency of sudden risk handling is greatly improved, which creates decisive conditions for the early evacuation of the parking lot and the early response of fire rescue.

[0097] In a possible implementation manner, after the vehicle is controlled to leave the parking lot through the driving control instruction, the method further includes: after the vehicle is controlled to leave the parking lot through the driving control instruction, when any target event is triggered, it is determined that the sudden risk is relieved.

[0098] Here, once the vehicle has left the parking lot, it marks the official end of a single emergency response process. It is necessary to determine whether the sudden risk has been eliminated in order to safely, smoothly, and orderly restore the vehicle from a highly tense emergency state to a normal state suitable for new tasks, and to complete the knowledge retention of this incident.

[0099] To determine whether a sudden risk can be resolved, it is first necessary to determine whether a target event has been triggered. Target events include: the vehicle has left the parking lot via driving control commands; the vehicle has received a risk clearance signal from the outside world; and the vehicle owner has sent a reset command to the vehicle confirming that it is safe.

[0100] The target events serve as clear confirmation signals of a "safe state," acting as the basis for determining whether a sudden risk has been eliminated. Specifically, the first target event is the vehicle successfully arriving at and stably parking in a pre-designated "safe area" along the planned route, such as a designated location at an open ground exit or a wide-open fire evacuation area. The second target event is the arrival of external rescue personnel who send a "risk eliminated" signal to the vehicle via V2I or a dedicated interface. The third target event is the vehicle owner actively confirming vehicle safety and issuing a reset command via a mobile application.

[0101] When any target event is detected, it means that the sudden risk has been eliminated. At this time, the emergency control mode is exited, the vehicle returns to normal status, and it can wait for the next parking command request in the parking lot to continue to execute the parking task.

[0102] Here, the primary criterion for determining whether a sudden risk has been resolved is whether the first target event has been met, i.e., whether the vehicle has reached a safe area. The second and third target events are usually triggered only after the first target event has occurred. External rescue personnel and vehicle owners primarily learn of the sudden risk in the parking lot through early warning information; therefore, determining whether the sudden risk has been resolved occurs after the tiered early warning system is activated.

[0103] like Figure 3 As shown, after completing the tiered warning, the system then determines whether the vehicle has reached the safe zone. If so, the emergency risk has been eliminated, the emergency control mode is exited, the system is reset, and the process ends. If not, the emergency risk has not yet been eliminated. In this case, the vehicle's onboard sensors continue to monitor the risk source, assess the emergency risk, and plan a route to leave the parking lot to ensure the vehicle reaches the safe zone.

[0104] In this embodiment of the invention, after controlling the vehicle to leave the parking lot, a sudden risk resolution judgment is performed to cause the vehicle to exit the emergency control mode and then be used to perform the parking task normally. This forms a logical closed mechanism that enables the vehicle to flexibly perform emergency avoidance during autonomous parking, fundamentally filling the safety gap of high-level automatic parking in the closed scenario of parking lots.

[0105] In summary, when executing a parking command request for autonomous parking, this embodiment of the invention utilizes onboard sensor data for multi-source risk perception. When a sudden risk is detected—such as a passenger cabin occupant's vital signs, an environmental disaster, or a ground anomaly—an emergency response mechanism is activated. This proactively triggers a switch in vehicle scheduling priorities, executing an emergency response request to escape the vehicle while interrupting parking. This ensures that the autonomous parking calculation logic not only serves to complete parking but also flexibly executes emergency handling through priority switching when risks are perceived, rather than passively waiting for the vehicle to stop. Through the emergency response mechanism, the core objective of the vehicle is shifted from completing parking to safely departing, ensuring that the vehicle automatically enters a preset, safer state under risk conditions. This forms a closed-loop safety mechanism, ensuring the safety of the vehicle and its occupants, and is more conducive to rescue operations compared to passively waiting for parking.

[0106] The following is a detailed description of the parking emergency dispatch device based on multi-source triggering provided by the present invention.

[0107] Figure 4 This is a schematic diagram of the parking emergency dispatching device based on multi-source triggering provided by the present invention, as shown below. Figure 4 As shown, the parking emergency dispatch device based on multi-source triggering specifically includes: a parking initialization module 401, a multi-source anomaly perception module 402, a dynamic switching module 403, an emergency path planning module 404, and a vehicle driving control module 405.

[0108] Specifically, the parking initialization module 401 is used to respond to a parking command request from the vehicle to autonomously park in the parking lot, determine the digital map corresponding to the parking lot, and acquire onboard sensor data; the multi-source anomaly perception module 402 is used to perform anomaly detection based on the onboard sensor data to obtain a risk status signal; the dynamic switching module 403 is used to interrupt the parking command request and switch the vehicle's scheduling priority from responding to the parking command request to responding to the emergency response request when the risk status signal indicates that there is a sudden risk in the current parking; the emergency path planning module 404 is used to respond to the emergency response request and retrieve a reference motion trajectory for the vehicle to leave the parking lot from the digital map based on the vehicle's current initial pose; the vehicle driving control module 405 is used to generate a vehicle driving control command based on the reference motion trajectory and control the vehicle to leave the parking lot through the driving control command.

[0109] The vehicle driving control module 405 is further configured to, after controlling the vehicle to leave the parking lot via the driving control command, trigger a Level 1 warning when the sudden risk is the presence of vital signs in the cabin; trigger a Level 2 warning when the sudden risk is a fire outside the vehicle; and trigger a Level 3 warning when the sudden risk is a risk of ground collapse at the location of the vehicle. The warning information of the Level 1, Level 2, and Level 3 warnings is sent to a target system, which includes an online management system for the parking lot, an emergency response system linked to the parking lot, and an application terminal system corresponding to the vehicle owner. The priority of the Level 2 warning is greater than that of the Level 3 warning, but less than that of the Level 1 warning.

[0110] In one possible implementation, the vehicle driving control module 405 is further configured to determine that the sudden risk has been cleared when any target event is triggered after the vehicle has been controlled to leave the parking lot by the driving control command; the target event includes: the vehicle has left the parking lot by the driving control command; the vehicle has received a risk clearance signal sent from the outside; the vehicle owner has sent a reset command to the vehicle confirming that it is safe.

[0111] The parking emergency dispatching device based on multi-source triggering provided in the above embodiments can realize the technical solutions described in the above embodiments of the parking emergency dispatching method based on multi-source triggering. The specific implementation principles of each module or unit can be found in the corresponding content in the above embodiments of the parking emergency dispatching method based on multi-source triggering, and their technical effects can also be referred to each other, which will not be repeated here.

[0112] This invention also provides a vehicle, including an on-board sensor, an on-board device, and an electronic device 500, wherein the electronic device 500 communicates with the on-board sensor and the on-board device via a vehicle network.

[0113] like Figure 5 As shown, the electronic device 500 includes a processor 501, a memory 502, and a display 503. Figure 5 Only some components of the electronic device 500 are shown, but it should be understood that it is not required to implement all the components shown, and more or fewer components may be implemented instead.

[0114] In some embodiments, memory 502 may be an internal storage unit of electronic device 500, such as a hard disk or memory of electronic device 500. In other embodiments, memory 502 may also be an external storage device of electronic device 500, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc. equipped on electronic device 500.

[0115] Furthermore, the memory 502 may include both internal storage units of the electronic device 500 and external storage devices. The memory 502 is used to store application software and various types of data installed on the electronic device 500.

[0116] In some embodiments, processor 501 may be a central processing unit (CPU), microprocessor, or other data processing chip, used to run program code stored in memory 502 or process data, such as the multi-source trigger-based parking emergency dispatch method of the present invention.

[0117] In some embodiments, display 503 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. Display 503 is used to display information from electronic device 500 and to display a visual user interface. Components 501-503 of electronic device 500 communicate with each other via a system bus.

[0118] In some embodiments of the present invention, when the processor 501 executes the computer program in the memory 502, the following steps can be implemented: in response to a parking command request for autonomous parking of a vehicle in a parking lot, determining the digital map corresponding to the parking lot and acquiring onboard sensor data; performing anomaly detection based on the onboard sensor data to obtain a risk status signal; when the risk status signal indicates that there is a sudden risk in the current parking, interrupting the parking command request and switching the vehicle's scheduling priority from responding to the parking command request to responding to an emergency response request; in response to the emergency response request, retrieving a reference motion trajectory for the vehicle to leave the parking lot from the digital map according to the vehicle's current initial pose; generating a driving control command for the vehicle based on the reference motion trajectory, and controlling the vehicle to leave the parking lot through the driving control command.

[0119] It should be understood that when the processor 501 executes the computer program in the memory 502, in addition to the functions described above, it can also perform other functions, as can be found in the description of the corresponding method embodiments above.

[0120] Furthermore, the embodiments of the present invention do not specifically limit the type of electronic device 500 mentioned. Electronic device 500 can be a mobile phone, tablet computer, personal digital assistant (PDA), wearable device, laptop computer, or other portable electronic device. Exemplary embodiments of portable electronic devices include, but are not limited to, portable electronic devices running iOS, Android, Microsoft, or other operating systems. The aforementioned portable electronic device can also be other portable electronic devices, such as a laptop computer with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in some other embodiments of the present invention, electronic device 500 may not be a portable electronic device, but rather a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0121] The vehicle-mounted sensor is used to collect vehicle-mounted sensor data and store the sensor data in the memory 502 of the electronic device 500; the vehicle-mounted device is used to obtain the vehicle driving control command after the processor 501 of the electronic device 500 executes the program stored in the memory 502, and control the vehicle to leave the parking lot through the driving control command.

[0122] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware, and the program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0123] The above provides a detailed description of the parking emergency dispatch method, device, equipment, and vehicle based on multi-source triggering provided by the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A parking emergency dispatch method based on multi-source triggering, characterized in that, include: In response to a parking command request from a vehicle to autonomously park in a parking lot, the system determines the digital map corresponding to the parking lot and acquires data from onboard sensors. Anomaly detection is performed based on the data from the vehicle-mounted sensors to obtain a risk status signal; When the risk status signal indicates that there is a sudden risk in the current parking, the parking command request is interrupted, and the vehicle scheduling priority is switched from responding to the parking command request to responding to the emergency response request. In response to the emergency response request, a reference motion trajectory for the vehicle to leave the parking lot is retrieved from the digital map based on the vehicle's current initial pose. Based on the reference motion trajectory, a driving control command for the vehicle is generated, and the vehicle is controlled to leave the parking lot using the driving control command.

2. The parking emergency dispatch method based on multi-source triggering according to claim 1, characterized in that, The vehicle-mounted sensor data includes vehicle cabin monitoring data, external environment data, and its own status data. The anomaly detection based on the vehicle-mounted sensor data to obtain a risk status signal includes: Based on the data from the vehicle-mounted sensors, cabin vital signs detection, environmental disaster detection, and ground anomaly detection are performed respectively to obtain detection results; Risk decisions are made based on the detection results to obtain the risk status signal.

3. The parking emergency dispatch method based on multi-source triggering according to claim 2, characterized in that, The process of making risk decisions based on the detection results to obtain risk status signals includes: When the detection value corresponding to each detection result is not greater than the corresponding preset detection threshold, it is determined that the current parking is in a normal state, and a corresponding risk state signal is generated. When the detection value corresponding to any detection result is greater than the corresponding preset detection threshold and the duration is greater than the preset time threshold, it is determined that there is a sudden risk in the current parking and a corresponding risk status signal is generated.

4. The parking emergency dispatch method based on multi-source triggering according to claim 1, characterized in that, The step of retrieving a reference trajectory for the vehicle to leave the parking lot from the digital map based on the vehicle's current initial pose includes: Retrieve multiple candidate safety exits for leaving the parking lot from the digital map; Multiple sets of candidate paths are determined from the current initial pose to the multiple candidate safety exits, wherein each set of candidate paths corresponds one-to-one with each candidate safety exit in the multiple candidate safety exits; For each candidate path, calculate the collision cost of encountering obstacles, the path distance cost, and the smoothing cost of path curvature. The collision cost, path distance cost, and smoothing cost are weighted according to preset weights, and the weighted results are summed to obtain the cost function value of each candidate path. The candidate path with the smallest cost function value in each group of candidate paths is determined as the target path for each candidate safe exit. The kinematic feasibility of the target path for each candidate safety exit is verified, and the motion trajectory corresponding to the successfully verified target path is determined as the reference motion trajectory for the vehicle to leave the parking lot.

5. The parking emergency dispatch method based on multi-source triggering according to claim 4, characterized in that, The kinematic feasibility of the target path is verified using the following methods: When the target path is detected to meet the target conditions, the kinematic feasibility check is determined to be successful, and the target path is retained. When it is detected that the target path does not meet at least one of the target conditions, the kinematic feasibility check is determined to have failed, and the target path is deleted. The target conditions include: The curvature of the motion trajectory is less than or equal to the maximum curvature of the vehicle, where the maximum curvature of the vehicle is the maximum curvature allowed by the vehicle hardware when the vehicle is turning; The rate of change of the continuous steering trajectory is within the allowable rate of change when the vehicle steering actuator performs continuous steering; The distance between the target path's trajectory and the obstacle is greater than a preset safe distance threshold.

6. The parking emergency dispatch method based on multi-source triggering according to claim 1, characterized in that, When controlling the vehicle to leave the parking lot via the driving control command, the method further includes: Acquire real-time external environment data collected by vehicle-mounted sensors, and perform obstacle detection based on the real-time external environment data; If the obstacle detection result indicates that there is an obstacle that will collide with the reference trajectory within a certain period of time in the future, then a local path to the safety exit in the reference trajectory is replanned based on the vehicle's current position. The driving control command is updated according to the local path, and the vehicle is controlled to continue leaving the parking lot based on the updated driving control command.

7. The parking emergency dispatch method based on multi-source triggering according to claim 2, characterized in that, After controlling the vehicle to leave the parking lot via the driving control command, the method further includes: When the aforementioned emergency risk is the presence of vital signs in the cabin, a Level 1 warning is triggered. When the aforementioned emergency risk is a fire outside the vehicle, a Level 2 warning is triggered; When the aforementioned emergency risk is a risk of ground collapse where the vehicle is located, a Level 3 warning is triggered; The warning information of the first-level warning, the second-level warning, and the third-level warning is sent to the target system. The target system includes the online management system of the parking lot, the emergency response system linked to the parking lot, and the application terminal system corresponding to the vehicle owner. The warning priority of the second-level warning is higher than that of the third-level warning, but lower than that of the first-level warning.

8. The parking emergency dispatch method based on multi-source triggering according to claim 1, characterized in that, After controlling the vehicle to leave the parking lot via the driving control command, the method further includes: When any one of the target events is triggered, the sudden risk is determined to be resolved; The target event includes: The vehicle has left the parking lot via the aforementioned driving control command; The vehicle receives a risk clearance signal from the outside world; The vehicle owner sends a reset command to the vehicle after confirming that it is safe.

9. A parking emergency dispatching device based on multi-source triggering, characterized in that, include: The parking initialization module is used to respond to parking command requests from vehicles to autonomously park in parking lots, determine the digital map corresponding to the parking lot, and acquire onboard sensor data. A multi-source anomaly perception module is used to detect anomalies based on the vehicle-mounted sensor data and obtain a risk status signal; The dynamic switching module is used to interrupt the parking command request and switch the vehicle's scheduling priority from responding to the parking command request to responding to the emergency response request when the risk status signal indicates that there is a sudden risk in the current parking. An emergency route planning module is used to respond to the emergency response request by retrieving a reference trajectory for the vehicle to leave the parking lot from the digital map based on the vehicle's current initial position. The vehicle driving control module is used to generate driving control commands for the vehicle based on the reference motion trajectory, and to control the vehicle to leave the parking lot through the driving control commands.

10. A vehicle, characterized in that, It includes vehicle-mounted sensors, vehicle-mounted equipment, and electronic devices, wherein the electronic devices, vehicle-mounted sensors, and vehicle-mounted equipment all communicate via a vehicle network. The electronic device includes a memory and a processor, wherein the memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps of the parking emergency dispatch method based on multi-source triggering as described in any one of claims 1 to 7. The vehicle-mounted sensor is used to collect vehicle-mounted sensor data and store the sensor data in the memory of the electronic device; The on-board device is used to obtain driving control commands for the vehicle after the processor of the electronic device executes the program stored in the memory, and to control the vehicle to leave the parking lot through the driving control commands.

Images