A control method, system and vehicle for a seat

Abstract

This invention provides a seat control method, system, and vehicle, relating to the field of seat adjustment technology. Upon receiving a seat flattening command, it acquires vital sign data of a target cabin area collected by a bio-radar sensor; based on the vital sign data, it performs vital sign detection to obtain a vital sign detection result within the target cabin area; when the vital sign detection result indicates that no living being exists in the target cabin area, it controls the seat to perform a flattening operation. This proactive and accurate detection of living beings in the target area before seat flattening solves the problems of misjudgment and safety hazards caused by manual confirmation or weight sensors in existing technologies, thereby improving the active safety performance of vehicle seat systems.

Description

Technical Field

[0001] This invention relates to the field of seat adjustment technology, and more specifically, to a seat control method, system, and vehicle. Background Technology

[0002] With the development of automotive intelligent technology, the function of automatically flattening vehicle seats into a resting bed has become an important application of intelligent cockpits. Typically, in response to user trigger signals, the front and rear seats are automatically adjusted to form a bed, and the air mattress is automatically inflated, thereby reducing manual operation by the user. However, the existing technology has the following drawbacks: during the automatic flattening process of the seat, the safety judgment of the existing solution mainly relies on manual confirmation by the user or weight sensors, which makes it difficult to distinguish between static living beings (such as sleeping children or pets) and non-living objects (such as backpacks or blankets), and it is even more impossible to actively detect whether there are living beings in the flattening area of ​​the seat. Therefore, it is impossible to actively and accurately determine whether there are living beings in the area before the flattening action is performed, which poses a serious risk of pinching or squeezing. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide a seat control method, system and vehicle to improve the active safety capabilities of the vehicle seat system.

[0004] In a first aspect, this application provides a method for controlling a seat, including: Upon receiving a command to flatten the seat, acquire vital sign data of the target cabin area collected by the bio-radar sensor; Based on vital sign data, life detection is performed to obtain the life detection results within the target cockpit area; When the life detection results indicate that there are no living beings in the target cabin area, the control seat will perform a flattening operation.

[0005] Optionally, life detection is performed based on vital sign data to obtain life detection results within the target cabin area, including: Perform a fast Fourier transform on the vital signs data to extract the frequency domain peak values ​​corresponding to respiratory rate and / or heart rate; If the frequency domain peak value is within the preset vital sign frequency range and continues to exceed the preset time threshold, then the detection result of the vital sign in the target cockpit area is determined to be the presence of a vital sign. If the frequency domain peak value exceeds the preset vital sign frequency range, then the detection result of the vital signs in the target cockpit area is determined to be that there are no living beings.

[0006] Optionally, before controlling the seat to perform the flattening operation, the following is also included: Acquire depth image data of the target cockpit area collected by the visual sensor; Based on the depth image data, static obstacles on the moving path of the seat are detected; Based on the detection results of static obstacles, an obstacle avoidance movement path for the seat is generated.

[0007] Optionally, the process of controlling the seat to perform the flattening operation based on the detection results of living beings in the target cockpit area also includes: Real-time acquisition of operating data of the adjustment motors of each seat in the target cockpit area; If the running data exceeds the dynamic threshold, the cabin environment image of the target cabin area collected by the vision sensor and the pressure distribution data of the seat collected by the pressure distribution sensor are acquired. Obstacle detection is performed based on operational data, cabin environment images, and pressure distribution data to determine the type of obstacle. Perform the appropriate stop or rollback operation based on the type of obstacle.

[0008] Optionally, depending on the type of obstacle, perform the corresponding stop or rollback operation, including: When the obstacle is a hard obstacle, the control seat stops moving; When the obstacle is a flexible object, the control seat will reverse and retract to the first preset distance. When the obstacle is a living being, the control seat reverses to a second preset distance and triggers an alarm; wherein the second preset distance is greater than the first preset distance.

[0009] Optionally, during the process of controlling the seat to perform the flattening operation, the following are included: Acquire the vehicle's static tilt angle data collected by the inertial measurement unit; Based on static tilt angle data, determine the current tilt angle of the seat relative to the horizontal plane; Based on the difference between the current tilt angle and the target horizontal tilt angle, the surface of the control seat is kept horizontal relative to the horizontal plane.

[0010] Optionally, after the seat is folded out, the following steps are also included: Acquire point cloud data of the seat surface collected by a visual sensor, and construct a digital elevation model of the seat surface based on the point cloud data; The elevation difference of the seat surface is determined based on the digital elevation model; If the height difference is greater than the preset flatness threshold, the seat surface will be adjusted to a flat surface.

[0011] Optionally, the seat control method provided in this application further includes: When the life detection result indicates the presence of a life form in the target cabin area, the flattening operation is prohibited, and a safety alarm and / or voice prompt is triggered.

[0012] Secondly, this application provides a seat control system, including: The data receiving module is used to acquire vital sign data of the target cabin area collected by the bio-radar sensor when a seat flattening command is received. The data analysis module is used to detect living beings based on vital sign data and obtain the detection results of living beings in the target cockpit area; The control execution module is used to control the seat to perform a flattening operation when the life detection result indicates that there is no life in the target cabin area.

[0013] Thirdly, this application provides a vehicle including the aforementioned seat control system.

[0014] This application provides a seat control method, system, and vehicle. Upon receiving a seat flattening command, the system acquires vital sign data of a target cabin area collected by a bio-radar sensor; performs vital sign detection based on the vital sign data to obtain a vital sign detection result within the target cabin area; and controls the seat to perform a flattening operation when the vital sign detection result indicates that no living being exists in the target cabin area. This proactive and accurate detection of living beings in the target area before seat flattening solves the problems of misjudgment and safety hazards caused by manual confirmation or weight sensors in existing technologies, thereby improving the active safety performance of the vehicle seat system.

[0015] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A flowchart of a seat control method provided by an embodiment of the present invention is shown; Figure 2 A schematic diagram of a seat control system provided in an embodiment of the present invention is shown. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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 some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0019] This application provides a seat control method, see below. Figure 1 As shown, the seat control method provided in this application includes at least the following steps: Step 110: Upon receiving the seat flattening command, acquire vital sign data of the target cabin area collected by the bio-radar sensor.

[0020] In this embodiment of the application, upon receiving a seat flattening command, the bio-radar sensor deployed above and / or around the target cabin area is first activated. The bio-radar sensor uses millimeter-wave continuous wave radar technology to transmit low-power millimeter-wave signals to the target cabin area in a non-contact manner and receives echo signals reflected by the weak mechanical vibrations of the human or animal's chest cavity surface caused by breathing and heartbeat. By demodulating and filtering the echo signals, displacement change information of chest cavity vibration is obtained, and then vital sign data is obtained, including respiratory rate and / or heart rate data.

[0021] Furthermore, the bio-radar sensor is fixedly installed in the center of the vehicle cabin ceiling, specifically inside the headliner panel in the area above the second and third rows of seats, or around the reading lights or sunroof control module in the cabin ceiling. This ensures the antenna radiation surface of the bio-radar sensor faces vertically towards the cabin floor, thus covering the entire seating and flattening area between the second and third rows of seats. The bio-radar sensor's acquisition range is configured to encompass the entire target cabin area involved in the seat flattening operation, including: the seat cushions and backrests of the second and third rows, and the transition space formed between the second and third rows after the second and third rows of seats slide forward. Vertically, the acquisition range extends downwards from the ceiling mounting location to the seat cushion surface, covering low-lying spaces where children or pets may be present. Horizontally, the acquisition range covers the central aisle area between the two seats and the edges of the headliner panels near the doors. Additionally, the bio-radar sensor's beam angle is set to adequately cover the target cabin area without significantly exceeding the vehicle's interior boundaries, avoiding interference signals from unrelated moving objects outside the vehicle.

[0022] By configuring the installation location and acquisition range of the bio-radar sensor, a non-contact vital sign scan can be performed on the entire area to be flattened before the seat is moved or reclined, ensuring that living beings are included in the detection range regardless of whether they are sitting, lying down, or at rest. Compared to the passive judgment methods of traditional solutions that rely on user visual observation or weight sensors, this application uses a bio-radar sensor that can detect the breathing and heart rhythm of children or pets that are stationary or even asleep. At the same time, the bio-radar sensor can penetrate obstructions such as blankets and clothing to effectively collect signals of living beings in covered or concealed states, and can also detect reflected echoes from non-living objects. Compared to visual sensors that rely on optical images, bio-radar is not affected by light, obstructions, or blind spots, and can more comprehensively perceive the presence of living beings in various postures within the cabin. In addition, the acquisition process does not require any active cooperation from the occupants and will not pose any radiation safety risks or psychological pressure to the occupants in the cabin. The obtained respiratory rate and / or heart rate data will be used as the raw signal input for subsequent frequency domain analysis to accurately distinguish between living and non-living objects.

[0023] Step 120: Perform life detection based on vital sign data to obtain the life detection results within the target cabin area.

[0024] In this embodiment, frequency domain analysis can be performed on vital sign data to determine the detection result of a living being in the target cabin area, including: performing a fast Fourier transform on the vital sign data to extract the frequency domain peak corresponding to the respiratory rate and / or heart rate; if the frequency domain peak is located within a preset vital sign frequency range and continues to exceed a preset time threshold, then the detection result of a living being in the target cabin area is determined to be the presence of a living being.

[0025] Since the raw vital sign data collected by the bio-radar sensor is a continuous waveform signal with time as the horizontal axis, this waveform signal is mixed with components such as chest cavity fluctuations caused by breathing, weak vibrations caused by heartbeats, and various noise components in the vehicle environment. In order to effectively separate vital sign features from the mixed waveform signal, a Fast Fourier Transform is first performed on the vital sign data to convert the time domain signal into a frequency domain signal. That is, the time domain signal is decomposed into a superposition of different frequency components, and the energy intensity corresponding to each frequency component is calculated, thereby obtaining the frequency domain peak value corresponding to the respiratory rate and / or heart rate. Then, the frequency domain peak value corresponding to the respiratory rate and / or heart rate is compared with a preset vital sign frequency range. If the frequency domain peak value is located within the preset vital sign frequency range, and the frequency domain peak value state persists for more than a preset time threshold, then the detection result of the vital sign in the target cabin area is determined to be the presence of a living being.

[0026] In this embodiment, the time-domain signal is converted to the frequency domain for analysis using Fast Fourier Transform (FFT). This makes weak periodic signals (such as respiratory rate and heart rate data) that are difficult to extract directly in the time domain clearly discernible, thereby effectively enhancing the sensitivity of vital sign detection and improving the signal-to-noise ratio. By setting a duration threshold, the system can effectively distinguish between genuine vital sign signals and transient interference signals. For example, the instantaneous vibration generated when a vehicle travels over a bumpy road may produce a peak in the frequency spectrum similar to the respiratory rate, but this peak cannot last beyond the time threshold, thus preventing false alarms. The dual judgment mechanism consisting of frequency domain interval judgment and duration judgment provides both high sensitivity and high specificity in vital sign detection results. It will not miss children or pets in a stationary state, nor will it frequently trigger false alarms due to environmental interference, thus providing an accurate and reliable decision-making basis for whether to allow the seat flattening operation. Step 130: When the life detection result indicates that there is no life in the target cabin area, control the seat to perform a flattening operation.

[0027] In this embodiment, if the life detection result indicates the presence of a life form, the seat flattening operation is prohibited, and a high-priority safety alarm and / or voice prompt is triggered; if the life detection result indicates the absence of a life form, the seat flattening operation is permitted, and the seat is controlled to perform the flattening operation.

[0028] Furthermore, when the detection result indicates the presence of a living being, the seat flattening operation is prohibited, and a safety alarm is triggered. In this embodiment, the safety alarm is configured to have a higher interruption response priority than the seat motion control commands, enabling it to suspend any executing or preparing seat motion commands the moment a living being is detected; wherein, the triggering method of the safety alarm includes one or more combinations of the following: The vehicle's audio system can issue voice prompts, such as broadcasting a warning message like "Life form detected, stop seat flattening"; A red warning interface and text prompt information will be displayed on the central control screen or the rear entertainment screen; The seat provides tactile feedback to the user through a built-in vibration motor; The vehicle provides a visual cue by flashing a warning light on the roof.

[0029] The aforementioned alarm methods can be used individually or in combination to create a multimodal warning effect. Voice prompts can deliver danger information without distracting the user's visual attention, while tactile feedback can effectively attract the user's attention in noisy environments. Upon triggering the alarm, the seat flattening operation is marked as disabled, and the seat's current position and posture remain unchanged, awaiting further user commands. When the life detection result indicates that there is no life, the seat flattening operation is allowed, and the seat is controlled to move and recline according to the preset flattening program.

[0030] It should be noted that if the user re-initiates the seat flattening command or receives a new seat flattening command during the seat flattening operation, steps 110 to 130 above will be re-executed to ensure that the presence of living beings in the target cabin area is reconfirmed before any further action is taken by the seat.

[0031] In this embodiment, the detection results of living beings are directly used as the condition for allowing seat flattening operations, realizing a shift from user-responsible safety confirmation to proactive safety decision-making. When children, pets, or other living beings are present in the target cabin area, the seat flattening operation is actively refused, preventing accidents such as pinching and squeezing from the source. Compared with traditional solutions that rely on user visual confirmation or passive response from weight sensors, the technical solution provided in this application can complete the safety judgment before the seat has started any movement, possessing true proactive prevention capabilities. By setting an alarm mechanism with a higher priority than seat motion control commands, it is ensured that safety decisions are not overridden or bypassed by other functions or user commands, thereby guaranteeing the effectiveness of the safety mechanism. Through multimodal alarm methods such as voice prompts, visual displays, and tactile feedback, users can quickly learn the reason for prohibiting flattening and remove living beings in the target cabin area or adjust their positions according to the prompts, and then re-initiate the flattening command. The alarm mechanism not only ensures the safety of living beings but also provides users with clear operation guidance, significantly improving the intelligence level and user experience of the vehicle seat control system.

[0032] In one alternative embodiment, before controlling the seat to perform the flattening operation, the method further includes: acquiring depth image data of the target cabin area collected by a visual sensor; and detecting static obstacles on the movement path of the seat based on the depth image data. Based on the detection results of static obstacles, an obstacle avoidance movement path for the seat is generated.

[0033] In this embodiment, the vision sensor is a depth camera fixedly installed in the roof or side panel of the vehicle interior. The depth camera can be a structured light camera or a stereo vision camera, capable of acquiring real-time three-dimensional depth image data of the target cabin area. Each pixel in the depth image data contains spatial distance information from that pixel to the sensor. Furthermore, the vision sensor can also be an RGB-D camera, and the depth image data will also include color information. Based on the depth image data, the three-dimensional contours of the environment surrounding the seats in the cabin can be accurately reconstructed.

[0034] Furthermore, after acquiring depth image data, an image segmentation algorithm is first used to segment different objects in the depth image into independent image regions. The segmented target regions may include, but are not limited to, seat backs, seat cushions, floors, water cups, backpacks, and children's toys. Secondly, a pre-trained object recognition model is used to classify and identify each segmented region to determine whether it belongs to a static obstacle. In this embodiment, a static obstacle refers to an object that does not have the ability to move actively and is not expected to be moved actively in a short time, including a water cup left on the seat rail, a backpack placed on the folding path, and clothing or cables tangled in the mechanical mechanism. Then, after identifying the static obstacle, it is determined whether the static obstacle is located on the movement path that the seat takes from its current state to the target bed state. The movement path is a three-dimensional spatial trajectory, which includes the rotation arc of the seat back, the forward and backward translational line of the seat cushion, and the extension trajectory of the leg rest. If the spatial coordinates of a static obstacle fall within the area traversed by any of the aforementioned movement trajectories, then a static obstacle that may collide with the obstacle is identified on the movement path. At this point, the path planning algorithm is triggered. Based on the mechanical motion constraints of the seat, the size and position of the obstacle, and the flattened position of the target, the path planning algorithm uses collision detection and path search methods (such as the A* algorithm or the RRT algorithm) to calculate a collision-free movement trajectory that avoids all static obstacles, i.e., the obstacle avoidance movement path. For example, the obstacle avoidance movement path can be a multi-stage composite movement, such as the backrest first leaning back and then going around, or the seat cushion first shifting outward and then moving forward, so that the seat can reach a completely flat state. Finally, control commands are generated based on the obstacle avoidance movement path and sent to the adjustment motors of each seat to drive the electromechanical actuators to accurately follow the obstacle avoidance movement path.

[0035] In this embodiment, by proactively acquiring a depth image of the cabin and performing static obstacle detection before the flattening operation begins, collisions between the seat and clutter can be avoided from the outset, thereby protecting the seat's mechanical structure, motor, and the collided objects from damage, thus improving the smoothness of operation and user experience. For detected static obstacles, intelligent avoidance is prioritized over simple stopping, improving the success rate and flexibility of spatial transformation. By generating an obstacle avoidance path and moving according to the path, efficient utilization of cabin space is achieved. Even in everyday use scenarios with a lot of clutter, the seat flattening operation can be completed as much as possible, truly demonstrating the adaptive capabilities of the intelligent cockpit.

[0036] In one optional embodiment, during the process of controlling the seat to perform the flattening operation based on the detection results of living beings in the target cabin area, the method further includes: acquiring real-time operating data of the adjustment motors of each seat in the target cabin area; if the operating data exceeds a dynamic threshold, acquiring cabin environment images of the target cabin area collected by a visual sensor and pressure distribution data of the seats collected by a pressure distribution sensor; performing obstacle detection based on the operating data, cabin environment images, and pressure distribution data to obtain the obstacle type; and performing a corresponding stop or retraction operation according to the obstacle type. Further, performing the corresponding stop or retraction operation according to the obstacle type includes: when the obstacle type is a hard obstacle, controlling the seat to stop moving; when the obstacle type is a flexible object, controlling the seat to retract to a first preset distance; when the obstacle type is a living being, controlling the seat to retract to a second preset distance and triggering an alarm; wherein the second preset distance is greater than the first preset distance.

[0037] In this embodiment, after the seat flattening operation is permitted, the seats in the target cabin area are sequentially controlled to perform at least one of the following actions according to a preset action sequence: sliding, backrest reclining, and leg rest extension. The preset action sequence is a set of ordered instructions pre-planned based on the vehicle's interior space geometry, seat mechanical structure, and target bed configuration. For example, for a vehicle with a second and third row of seats, the preset action sequence could be: first, controlling the second row of seats to slide forward to their extreme position near the back of the first row of seats; then, controlling the backrest of the third row of seats to recline from an upright seating position forward or backward to a horizontal position; finally, controlling the backrest of the second row of seats to recline from an upright position to a horizontal position, so that the back of the second row of seat back and the front of the third row of seat back are horizontally aligned, forming a continuous sleeping plane. For seats with leg rests, the preset action sequence could also include raising the leg rests to a horizontally extended state simultaneously with or after the backrest reclines. The entire sequence of actions is initiated by the cockpit domain controller, which sends commands to the motor drivers of each seat in a sequential order to ensure that each moving part performs its actions in the predetermined sequence and speed curve, thus avoiding mechanical interference between different seats. During the seat flattening process, real-time current or torque feedback data of the adjustment motors driving each seat movement is acquired. The motor current and the motor's output driving torque are positively correlated. When the seat's moving parts encounter additional resistance during movement, the motor load increases, and the driving current rises accordingly. The real-time current or torque value of the motor is continuously monitored through a current sampling circuit or a torque sensor. Specifically, the current sensor is connected in series in the motor's power supply circuit to collect the motor's operating current value in real time; the torque sensor is installed on the motor's output shaft or the input end of the transmission mechanism to directly measure the motor's output driving torque. For example, when the seat's moving parts slide on the rails, the backrest rotates, or the leg rest extends, if there is an obstacle in the movement path, the motor load will suddenly increase, manifested as a sharp rise in the current value or a sudden jump in the torque value.

[0038] In this embodiment, the dynamic threshold can be adaptively adjusted based on the current movement stage, movement speed, and historical load baseline. Specifically, when the seat begins to move, current or torque data during the initial movement process is first collected to establish a normal load baseline. The load baseline reflects the basic resistance the seat experiences when moving in an unobstructed state, including factors such as mechanical friction, transmission loss, and the seat's own weight. During normal movement, the load will fluctuate within a certain range around this baseline. As long as the fluctuation amplitude does not exceed the allowable range, it is considered a normal state. The threshold curve for triggering emergency stop is dynamically adjusted based on the changing trends of movement speed, movement direction, and current load. For example, during the seat's acceleration phase, the current will have a normal upward spike, which may be mistakenly identified as a collision. During the uniform movement phase, the sensitivity to sudden load increases will increase accordingly. When the real-time monitored feedback data exceeds the dynamic threshold, an abnormal resistance event is determined to have occurred, i.e., there is an obstacle or clamping situation on the seat's movement path. Once a sudden load increase is detected, an emergency stop command is immediately triggered, cutting off the motor's power output or activating the braking mechanism to stop the seat's moving parts in the shortest possible time, thereby preventing injury to the occupant or damage to the seat's mechanical structure. By determining the dynamic threshold through a dynamic adaptive threshold mechanism, the sensitivity of triggering emergency stop can be flexibly adjusted according to different stages of seat movement and different working conditions. This ensures the reliability of the anti-pinch function and avoids frequent false triggers caused by normal motion load fluctuations.

[0039] Furthermore, after detecting that the feedback data exceeds the dynamic threshold, the obstacle type is determined based on the feedback data, cabin environment image, and seat pressure distribution data. Obstacle types include rigid obstacles, flexible objects, and living beings. Specifically, based on the feedback data, the waveform characteristics of the motor current or torque signal are analyzed. When the seat moving parts encounter a rigid obstacle during movement, since rigid obstacles are almost incompressible, a rigid-body collision occurs between the moving parts and the obstacle. The resistance increases suddenly and drastically at the moment of contact, so the current or torque signal will show a spike-like waveform with an extremely short rise time and a large amplitude variation. When the seat moving parts compress a flexible object, since the flexible object has a certain degree of compressibility, the resistance gradually increases with the depth of compression. Therefore, the current or torque signal will show a slowly rising arc-shaped waveform with a relatively gentle rise and a longer duration. By using waveform characteristic parameters such as waveform slope, rise time, and peak amplitude, it is preliminarily determined whether the obstacle is rigid or flexible. Target recognition and contour analysis are performed based on cockpit environment images. These images contain the shape contours, surface textures, and geometric features of obstacles. For rigid obstacles, such as water bottles, children's toys, metal objects, or backpacks left on seats, their shapes are clear, sharp, and exhibit regular geometric shapes, such as cylinders, cuboids, or spheres. Their shapes do not change significantly during compression, exhibiting typical rigid body characteristics. For flexible objects, such as clothing, blankets, soft cushions, or backpack straps, their shapes are relatively blurred, with irregular or gradually changing edges. They undergo significant deformation when compressed, and the images show dynamic changes in contour as the compression process occurs. For living organisms, such as children's hands and feet or pet limbs, image recognition can detect biologically characteristic morphologies, such as fingers, soles, and animal paws. Living organisms will actively avoid or struggle when compressed, and the image sequence shows the movement changes of the compressed area. A deep learning model is used to perform semantic segmentation and target classification on the cockpit environment images, outputting the obstacle category confidence score. Based on seat pressure distribution data, different types of obstacles produce different pressure distribution characteristics when they come into contact with the seat surface. Specifically, for rigid obstacles, due to the small contact area and concentrated pressure when the rigid surface contacts the seat, the pressure distribution image will show a tiny, extremely high-pressure spot, with the pressure in the surrounding area rapidly decreasing to zero. For flexible objects, because the flexible material deforms under pressure to conform to the seat surface, the contact area is larger and the pressure distribution is more uniform, the pressure distribution image will show a large area of ​​low pressure with a lower peak value, and the pressure will gradually transition from the center to the edge. For living organisms, such as children sitting or lying on the seat, the pressure distribution image will show a pressure distribution pattern with biological morphological characteristics, such as two symmetrical pressure areas on the buttocks, pressure lines along the spine on the back, and pressure transitions at the junctions of the limbs and torso. Furthermore, due to breathing and minute voluntary movements, the pressure distribution image of living organisms will show periodic minute fluctuations in the time series, forming pressure signal characteristics unique to living organisms. The feedback data, cabin environment images, and seat pressure distribution data are fused to determine the obstacle type. Specifically, weighted voting or Bayesian fusion methods can be used to comprehensively consider the waveform type of the feedback data, the recognition results of the cabin environment images, the feature classification of the pressure distribution images, and the vital sign detection results to output the final obstacle type label. When the judgment results from multiple sources are consistent, the result is directly adopted; when there are inconsistencies, the following priority rules are followed for adjudication: The first priority is the identification of living organisms. If the detection result indicates the presence of living organisms, or if periodic small fluctuations appear in the pressure distribution image, or if a morphology with biological characteristics is identified in the visual image, it will be prioritized as a type of living organism. The second priority is distinguishing between rigid obstacles and flexible objects. After ruling out the possibility of living organisms, a comprehensive judgment is made based on waveform characteristics and pressure distribution characteristics; if the waveform shows a peak shape and local high-pressure bright spots appear in the pressure distribution image, it is judged as a rigid obstacle; if the waveform shows an arc shape and a large area of ​​low pressure appears in the pressure distribution image, it is judged as a flexible object.

[0040] By cross-validating four dimensions—adjusting motor operating data, cabin environment images, seat pressure distribution data, and bio-radar detection results—the system can accurately distinguish between hard obstacles, flexible objects, and living beings, overcoming the limitations of single sensors in obstacle classification. For example, when the visual sensor is obstructed by the seat back and cannot directly see the obstacle, the adjusting motor operating data and pressure distribution data can still provide reliable judgment; when the pressure distribution sensor does not cover the obstacle contact area, the visual image can fill this blind spot. By introducing vital sign detection results as the highest priority judgment criterion, the system can achieve the fastest response when a living being is present, further improving the timeliness and reliability of safety protection. Through multimodal fusion, highly reliable and robust judgment of obstacle types is achieved, providing accurate input for subsequent execution of differentiated stop and rollback operations.

[0041] Furthermore, based on the identified obstacle type, the system executes corresponding stop or reversal operations: When the obstacle is a hard obstacle, the control seat immediately stops its current movement, ceasing both forward and reverse movement. Hard obstacles, such as water bottles, metal objects, or children's toys left on the seat rails, cannot be compressed, and continued movement could damage the seat's mechanical structure or the obstacle itself. Additionally, the physical position of a hard obstacle may have shifted or tilted after being compressed; reversing could cause the obstacle to roll further or fall, creating unpredictable risks. Therefore, only a stop operation is executed, waiting for the user to manually remove the hard obstacle before restarting the flattening process. When the obstacle is a flexible object, the control seat retracts to a first preset distance. Flexible objects, such as clothing, blankets, backpack straps, or soft interior trim, will deform when compressed but will not cause serious damage. The short first preset distance of retraction is sufficient to release the compressive pressure on the flexible object, preventing the fabric from being excessively stretched or deformed. At the same time, the user can easily pull the flexible object out of the seat's movement path. The first preset distance can be set to a short retraction amount, for example, between 10 mm and 30 mm. When the obstacle is a living being, the control seat retracts to a second preset distance, simultaneously triggering the highest priority safety alarm. Living beings, such as a child's hands, feet, body parts, or a pet's limbs, pose a risk of injury when squeezed, thus requiring a greater retraction distance to ensure the trapped part is completely freed. The second preset distance is set to a value greater than the first preset distance, for example, a retraction amount between 50 mm and 100 mm, to ensure the living being has sufficient space to escape. At the same time, a voice alarm is issued through the car audio system, a warning message is displayed on the central control screen, and tactile feedback is provided through seat vibration, alerting the user to the occurrence of a trapping event and informing them that the safety of the living being needs to be checked.

[0042] In this embodiment, by controlling the seat movement sequentially according to a preset action sequence, the system can achieve coordinated movement of multiple seats and multiple moving parts in complex spaces, avoiding mechanical interference and motion conflicts, and improving the reliability and efficiency of the flattening process. By monitoring motor current or torque feedback data in real time, it can detect abnormalities at the first moment of abnormal resistance, with a response speed faster than solutions that rely solely on vision or pressure sensors. By comprehensively analyzing four different types of data—feedback data, cabin environment images, seat pressure distribution data, and bio-radar data—the system determines the type of obstacle, overcoming the limitations of a single sensor in identifying the nature of obstacles. By fusing data from the cabin environment, seat pressure distribution, and bio-radar, the system can accurately distinguish between hard obstacles, flexible objects, and living beings, thereby adopting differentiated safety response strategies. Furthermore, the tiered response strategy demonstrates a fine balance between safety, reliability, and user experience: for hard obstacles, it simply stops without reversing, avoiding unnecessary secondary risks; for flexible objects, it reverses a short distance, releasing pressure without excessively interrupting the leveling process; and for living beings, it reverses a longer distance and triggers an alarm, maximizing personal safety and enhancing the intelligence and safety protection capabilities of the automatic seat leveling function.

[0043] In one alternative embodiment, controlling the seat to perform a flattening operation includes: acquiring static tilt angle data of the vehicle collected by an inertial measurement unit; determining the current tilt angle of the seat relative to a horizontal plane based on the static tilt angle data; and controlling the surface of the seat to remain horizontal relative to the horizontal plane according to the difference between the current tilt angle and a target horizontal tilt angle.

[0044] In this embodiment, the inertial measurement unit (IMU) is fixedly mounted on the vehicle chassis or seat frame. It integrates a three-axis accelerometer and a three-axis gyroscope, enabling it to measure the vehicle's spatial attitude angles when the vehicle is stationary and use these angles as static tilt angle data. In this embodiment, the static tilt angle data refers to the longitudinal tilt angle and lateral tilt angle output by the IMU when the vehicle is parked on a slope, shoulder, or inclined ground. Since the seats are installed inside the vehicle and the seat frame is rigidly connected to the vehicle body, the vehicle's static tilt angle directly determines the degree of inclination of the seat surface relative to the absolute horizontal plane in the unleveled state. After acquiring the static tilt angle data, it is transmitted to the seat controller. The seat controller first uses a coordinate transformation algorithm to map the vehicle's tilt angle data to the seat's own coordinate system, thereby accurately calculating the current tilt angle of the seat surface relative to the horizontal plane. The current tilt angle includes a fixed tilt angle component introduced by the non-horizontal parking ground, and an additional tilt component introduced by the seat's backrest or cushion adjustment mechanism not being fully flattened. During the leveling process, the seat is first adjusted to a fully level initial state, meaning the seat back, cushion, and leg rest are all horizontally laid out. At this point, the vehicle tilt component becomes the main factor affecting the flatness of the sleeping surface. The current tilt angle is compared with a preset target horizontal tilt angle, which is set to zero degrees (completely horizontal). The difference between the two is calculated, reflecting the magnitude and direction of the tilt angle that needs to be compensated. Based on this difference, a first adjustment command is generated, which includes the direction, amplitude, and speed parameters of the leveling actuator. The leveling actuator can be any of the following: multiple independently controlled electric push rods installed at the connection between the seat frame and the slide rail, or lifting columns with lead screw mechanisms installed at the four support points of the seat. According to the first adjustment command, the seat leveling actuator is controlled to move. For example, if there is a tilt angle in the front-rear direction, the lifting column of the front support point is controlled to rise or retract, and the rear support point does the opposite. If there is a tilt angle in the left-right direction, the left and right support points are controlled to move differentially. Through coordinated adjustment of each fulcrum, the entire flat surface of the seat is made to be consistent with the absolute horizontal plane in both the front-back and left-right directions. During the entire adjustment process, the inertial measurement unit continuously feeds back the current attitude, forming a closed-loop control until the difference between the current tilt angle and the target horizontal tilt angle converges to within the allowable error range.

[0045] In this embodiment, by introducing an inertial measurement unit to collect static tilt angle data and actively leveling the vehicle, the problem of seat tilting after being flattened when the vehicle is parked on uneven ground can be effectively solved. This keeps the sleeping surface inside the vehicle basically level, significantly improving the comfort and safety of passengers lying down and preventing body slippage or discomfort during sleep caused by tilting. By combining the leveling actuator with the seat flattening action, a fully automatic and seamless connection from the original folded state to the final level bed surface is achieved, eliminating the need for users to manually adjust pads or find a flat parking space, thus lowering the barrier to entry. By adopting a closed-loop control method to continuously correct the tilt angle difference, it can adapt to different slopes and even dynamically changing ground conditions, ensuring the accuracy and reliability of leveling.

[0046] In one optional embodiment, during the control of the seat to perform a flattening operation, this embodiment of the application acquires spatial contour and image information of the target cabin area collected by a visual sensor as a cabin environment image, used to identify static obstacles and / or living beings in the target cabin area; acquires pressure distribution images of the seat surface pressure distribution in the target cabin area collected by a pressure distribution sensor as seat pressure distribution data, used to determine whether living beings and / or non-living beings exist in the target cabin area; acquires the current angle and position information of the seat in the target cabin area collected by a Hall sensor or rotary encoder as seat posture data, used to determine whether the seat is ready for flattening; and performs fusion analysis based on the cabin environment image, seat pressure distribution data, and seat posture data to evaluate the safety of the seat during the flattening operation.

[0047] Furthermore, the vision sensor is installed in the cabin ceiling or near the rearview mirror to capture images of the cabin environment in the target cabin area. The cabin environment image contains two-dimensional color or grayscale information of the target cabin area, as well as three-dimensional spatial depth information obtained through stereo vision or structured light technology. By performing image processing and pattern recognition on the cabin environment image, static obstacles within the target cabin area can be identified, such as water bottles, backpacks, clothing, cables, etc., left on the seat surface or floor. It can also assist in identifying living beings, such as detecting the outline of a child or pet. Furthermore, the vision sensor is fixedly installed in the central area of ​​the vehicle cabin ceiling, specifically inside the headliner above the second and third rows of seats, or integrated near the rearview mirror base, reading light module, or sunroof control panel. This installation position allows the vision sensor's optical lens to be perpendicular to the cabin floor or tilted at a certain downward angle, thereby obtaining a complete field of view covering the second and third rows of seats and the transition area between them. The visual sensor's acquisition range covers the entire cabin width between the interior panels of both doors in the horizontal direction, and extends longitudinally from behind the front seats to the edge of the trunk entrance, ensuring that it can acquire images and perform three-dimensional spatial contour scanning of all target cabin areas involved in the seat flattening operation. Pressure distribution sensors are integrated in an array on the surface of the seat cushion and backrest, consisting of multiple densely arranged pressure sensing units. When an occupant or object is present on the seat surface within the target cabin area, the pressure sensing units detect pressure values ​​of varying magnitudes, i.e., seat pressure distribution data. By analyzing this data, it is possible to determine whether a living or non-living entity exists within the target cabin area. Specifically, living entities such as children or pets, due to their irregular body shapes and the ability to make small, voluntary or involuntary movements, will appear as areas with specific area and shape characteristics that change over time in the seat pressure distribution data. In contrast, non-living entities such as backpacks or water bottles typically exhibit clearly defined boundaries, concentrated pressure values, and a long-term stable position. Furthermore, the pressure distribution sensors are integrated into the seat assembly in the form of a flexible thin-film array, specifically laid above the seat cushion foam layer of the second and third row seats, below the seat cover, and on the front area of ​​the seat back. The pressure distribution sensor consists of multiple densely arranged pressure sensing units in a matrix of rows and columns. Each sensing unit can independently detect the pressure value at that location. When an occupant or an object sits on the seat surface, the pressure sensing units on the cushion and backrest collect the pressure distribution from different contact points such as the buttocks and back, forming complete seat pressure distribution data. Hall effect sensors or rotary encoders are installed at key locations in the moving mechanisms, such as the seat rails, backrest hinges, and leg rest hinges. Hall effect sensors detect changes in the magnetic field to sense the position and distance traveled by moving parts, while rotary encoders accurately measure the rotation angle by detecting pulse signals generated by shaft rotation. By collecting the current angle and position information of the seat through Hall effect sensors or rotary encoders, seat posture data is generated, specifically including the seat's fore-and-aft position coordinates on the rails, the backrest's tilt angle relative to the seat cushion, and the leg rest's extension angle relative to the seat cushion. Analyzing this seat posture data allows for determination of whether the seat currently possesses the mechanical conditions to be flattened, such as whether the second-row seats have slid forward enough to allow the third-row seats to fold down, whether the backrest angle is within the allowable adjustment range, and whether there is any risk of mechanical interference. Furthermore, Hall sensors are installed at the relative movement positions between the fixed and movable tracks of the seat slide rails, specifically at the output shaft end of the slide rail drive motor or at the detection position cooperating with the lead screw and nut mechanism. At least one Hall sensor is installed on each seat slide rail to detect the sliding displacement and current position of the seat in the fore-and-aft direction. The rotary encoder is installed on the angle adjuster shaft at the connection between the seat back and the seat cushion, and on the extension shaft at the connection between the leg rest and the seat cushion. The main body of the rotary encoder is fixed to the stationary part of the seat frame, and its rotation shaft is coaxially linked with the rotation shaft of the backrest or leg rest. When the angle of the backrest or leg rest changes, the rotary encoder outputs an electrical pulse signal proportional to the rotation angle. By detecting the slide rail position through Hall sensors and detecting the backrest angle and leg rest angle through rotary encoders, the system can obtain the precise posture of each moving part of the seat in real time.

[0048] In this embodiment of the application, after acquiring cabin environment images, seat pressure distribution data and seat posture data, the multimodal data composed of cabin environment images, seat pressure distribution data and seat posture data is subjected to time synchronization and coordinate system alignment data preprocessing to ensure that data from different sensors can be fused and analyzed under the same time reference and the same spatial reference.

[0049] Specifically, for time synchronization, since the vision sensor, pressure distribution sensor, Hall sensor, and rotary encoder each have different data acquisition frequencies and response delay characteristics, it is first necessary to unify the multiple data collected by these sensors onto the same time base to achieve high-precision time alignment using a hardware-triggered synchronization method. Specifically, the cockpit domain controller or gateway acts as the master clock device, simultaneously sending a hardware trigger signal to all sensors requiring synchronization according to a preset synchronization cycle. This trigger signal is transmitted in parallel to the vision sensor, pressure distribution sensor, Hall sensor, and rotary encoder via dedicated physical wires. When the trigger signal reaches each sensor, all sensors acquire a frame of data at the same time. Specifically, the vision sensor exposes and captures a frame of cockpit environment image at that moment; the pressure distribution sensor scans and records the current pressure values ​​of all pressure sensing units to form seat pressure distribution data; and the Hall sensor and rotary encoder latch the current position count and angle count values ​​to form seat attitude data. Through this hardware-triggered method, the time difference between the data collected by each sensor is controlled within the microsecond level, thus achieving precise synchronization of multi-sensor data in the time dimension. In this application, in addition to hardware-triggered synchronization, a timestamp based on the master clock can be added to each frame of data for subsequent data verification and interpolation. For sensor configurations that do not support hardware-triggered synchronization, a software timestamp method can be used, whereby each sensor records its local clock time when acquiring data, and then performs time alignment compensation on the data based on the offset and drift rate between the clocks. For coordinate system alignment, since the vision sensor, pressure distribution sensor, Hall sensor and rotary encoder are installed in different positions in the cockpit and each uses a different local coordinate system, it is necessary to convert the data from different sensors into a unified vehicle coordinate system. Specifically, a vehicle coordinate system with the vehicle center as the origin is established in advance, in which the vehicle's forward direction is the positive X-axis, the vehicle's width direction is the positive Y-axis, and the vehicle's height direction is the positive Z-axis. For cockpit environment images acquired by the vision sensor, the spatial coordinates of each pixel in the image are transformed from the camera coordinate system to the vehicle coordinate system based on the installation position parameters of the vision sensor (including its three-dimensional coordinates in the vehicle coordinate system and the pitch, yaw, and roll angles of the optical axis) and the camera's intrinsic parameter matrix. For seat pressure distribution data acquired by the pressure distribution sensor, the physical position of each pressure sensing unit on the seat surface is pre-calibrated, and this physical position is transformed into the vehicle coordinate system based on the current seat posture data. The seat posture data itself is provided by the Hall sensor and rotary encoder, including the fore-and-aft displacement of the seat rail and the rotation angles of the backrest and leg rest. Based on the seat posture data and the kinematic model of the seat frame, the current position of any point on the seat in the vehicle coordinate system is calculated in real time. Through the above coordinate transformation, the obstacle spatial contour of the vision sensor, the contact pressure position of the pressure distribution sensor, and the seat component positions of the Hall sensor and rotary encoder are all unified under the same vehicle coordinate system. By fusing and analyzing the perceived information from three dimensions—cabin environment images, seat pressure distribution data, and seat posture data—under the same temporal and spatial reference, the safety of seat flattening operations can be comprehensively assessed. This avoids misjudgments or omissions caused by incomplete information from a single sensor. The three judgment dimensions are independent yet complementary, together constituting the safety access conditions for seat flattening operations, providing a reliable decision-making basis for subsequent flattening execution.

[0050] In one optional embodiment, after the seat is controlled to perform a flattening operation, the method further includes: acquiring point cloud data of the seat surface collected by a visual sensor, and constructing a digital elevation model of the seat surface based on the point cloud data; determining the height difference of the seat surface based on the digital elevation model; and adjusting the seat surface to a flat surface if the height difference is greater than a preset flatness threshold.

[0051] In this embodiment, a partitioned inflatable mattress is pre-installed in the vehicle cabin. This partitioned inflatable mattress is a retractable flexible pad that is laid on top of the flattened seat surface during use and can be folded and stored in the vehicle trunk or under the seats when not in use. The mattress contains multiple independently controlled air chambers, each of which can be individually inflated or deflated to adjust the height of different areas of the mattress surface. The position of each air chamber spatially corresponds one-to-one with the detection areas of the pressure distribution sensor. This correspondence is established through a preset calibration mapping table, enabling precise correlation between flatness detection results and inflation compensation actions.

[0052] After the seat flattening operation is completed, the process also includes flatness detection and automatic leveling of the flattened seat surface. Specifically, point cloud data of the flattened seat surface within the target cabin area is acquired by a vision sensor. This is achieved by using a vision sensor installed in the cabin ceiling, employing structured light or stereo vision technology, to project a structured light pattern onto the seat surface or to acquire the three-dimensional spatial coordinates of a large number of sampling points on the seat surface through binocular stereo matching. The set of sampling points constitutes the point cloud data, and each point contains the lateral, longitudinal, and height coordinates of the sampling point in the vehicle coordinate system. The density of point cloud data determines the spatial resolution of flatness detection; higher density allows for the detection of finer unevenness. A digital elevation model (DEM) is constructed based on the acquired point cloud data. The specific construction process is as follows: First, the horizontal range of the seat surface is determined, and the seat surface is divided into multiple uniformly distributed grid cells. Then, for each grid cell, all point cloud data falling within that grid cell are collected, and the average or median value of the height coordinates of all point cloud data within that grid cell is calculated as the elevation value of that grid cell. For grid cells without any point cloud data falling within them, neighbor interpolation is used to fill them, and the elevation value of that grid cell is calculated based on the filled point cloud data. After the above processing, the three-dimensional shape of the seat surface is converted into a two-dimensional grid array, with each grid corresponding to an elevation value, thus forming a DEM. The elevation difference of the seat surface is determined based on a digital elevation model (DEM). The elevation difference is calculated by iterating through the elevation values ​​of all grid cells in the DEM, identifying the maximum and minimum values, and then calculating the difference between them. This difference represents the vertical distance between the highest and lowest points on the seat surface. A smaller elevation difference indicates a smoother seat surface; a larger elevation difference indicates more significant unevenness, steps, or gaps on the seat surface. After determining the height difference, it is compared with a preset flatness threshold. The preset flatness threshold is a pre-stored target value, which is the minimum acceptable standard for the flatness of the sleeping surface. When the height difference is less than or equal to the preset flatness threshold, it indicates that the flattened seat surface has met the flatness requirements and no leveling operation is needed. When the height difference is greater than the preset flatness threshold, it indicates that the flattened seat surface still has unevenness beyond the allowable range, and a seat leveling operation needs to be performed. The specific method for performing the seat leveling operation is to adjust the pressure of each independent air chamber in the zoned air mattress. This process creates a flat sleeping surface on the seat. Based on the deviation between the elevation values ​​of each grid cell in the digital elevation model and the target flat surface, inflation / deflation commands are generated for each air chamber. For concave areas where the elevation value is lower than the target flat surface, the corresponding air chamber is inflated, raising the mattress height in that area. For convex areas where the elevation value is higher than the target flat surface, the corresponding air chamber is deflated, lowering the mattress height in that area. All air chambers can be adjusted simultaneously, or sequentially from low to high or from high to low. During adjustment, the steps of point cloud data acquisition, digital elevation model construction, and elevation difference calculation can be repeated multiple times to form a closed-loop feedback control until the elevation difference is reduced to within a preset flatness threshold.

[0053] Furthermore, in this embodiment, after the user lies on the inflatable mattress, the mattress surface can be fine-tuned a second time based on the pressure distribution data collected by the pressure distribution sensor. Specifically, when the user lies on the inflatable mattress, the pressure distribution sensor collects pressure distribution data on the area where the user's body contacts the mattress, and identifies the pressure distribution of various parts of the user's body (such as the head, shoulders, waist, hips, and legs) based on the pressure distribution data. For areas with excessively high pressure values, it indicates that the mattress support in that area is too firm, and the corresponding air chamber is appropriately deflated to reduce the pressure; for areas with excessively low pressure values, it indicates that the mattress support in that area is insufficient, and the corresponding air chamber is appropriately inflated to increase support. Through this secondary fine-tuning, personalized firmness can be achieved according to the user's individual body shape and sleeping posture preferences, further improving sleep comfort.

[0054] In this embodiment, point cloud data is collected by a visual sensor and a digital elevation model is constructed, enabling precise three-dimensional quantitative assessment of the flatness of the unfolded seat surface. This overcomes the limitations of relying solely on visual observation or simple touch to quantify the flatness. Using elevation difference as a flatness assessment indicator is intuitive and easy to compare with preset thresholds, facilitating the determination of whether leveling is necessary. Pressure adjustment of each independent air chamber in the partitioned air mattress allows for refined compensation of local unevenness on the unfolded seat surface, rather than simply inflating the entire surface. The position of each air chamber corresponds spatially to the detection area of ​​the pressure distribution sensor, ensuring that the leveling command accurately applies to the local area requiring compensation, avoiding over-compensation or under-compensation. A closed-loop feedback control mechanism ensures the convergence of the leveling operation, gradually approaching the target flatness through multiple iterations. Secondary fine-tuning after the user lies down allows for personalized adaptation based on individual user body shape and sleeping posture, achieving a complete leveling function from static flatness to dynamic adaptation, significantly improving the user's comfort experience while resting in the vehicle.

[0055] This application provides a seat control system, see below. Figure 2 As shown in the embodiment of this application, the seat control system includes: The data receiving module 210 is used to acquire vital sign data of the target cabin area collected by the bio-radar sensor when it receives the seat flattening command. Data analysis module 220 is used to detect living beings based on vital sign data and obtain the detection results of living beings in the target cockpit area; The control execution module 230 is used to control the seat to perform a flattening operation when the life detection result indicates that there is no life in the target cabin area.

[0056] This application provides a vehicle including the aforementioned seat control system.

[0057] It should be noted that the principle of the seat control system provided in this application embodiment to solve the technical problem is similar to that of the seat control method provided in this application embodiment. Therefore, the implementation of the seat control system provided in this application embodiment can refer to the implementation of the seat control method provided in this application embodiment, and repeated details will not be repeated.

[0058] The computer-readable storage medium provided in the embodiments of this application is described below. The computer-readable storage medium provided in the embodiments of this application stores computer instructions, which, when executed by a processor, implement the seat control method provided in the embodiments of this application. Specifically, the computer instructions can be built into or installed in a processor, so that the processor can implement the seat control method provided in the embodiments of this application by executing the built-in or installed computer instructions.

[0059] In addition, the seat control method provided in this application embodiment can also be implemented as a computer program product, which includes program code. The program code implements the seat control method provided in this application embodiment when it is run on a processor.

[0060] The computer program product provided in this application embodiment may employ one or more computer-readable storage media, which may be, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination thereof. Specifically, more specific examples (a non-exhaustive list) of computer-readable storage media include electrical connections with one or more wires, portable disks, hard disks, RAM, ROM, erasable programmable read-only memory (EPROM), optical fibers, portable compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0061] The computer program product provided in this application embodiment can be a CD-ROM and include program code, and can also run on electronic devices such as computers. However, the computer program product provided in this application embodiment is not limited thereto. In this application embodiment, the computer-readable storage medium can be any tangible medium that contains or stores program code, which can be used by or in conjunction with an instruction execution system, device, or apparatus.

[0062] It should be noted that although several units or sub-units of the device have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of this application, the features and functions of two or more units described above can be embodied in one unit. Conversely, the features and functions of one unit described above can be further divided and embodied by multiple units.

[0063] Furthermore, although the operations of the method of this application are described in a specific order in the accompanying drawings, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.

[0064] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0065] Obviously, those skilled in the art can make various modifications and variations to the embodiments of this application without departing from the spirit and scope of the embodiments of this application. Therefore, if these modifications and variations to the embodiments of this application fall within the scope of the claims of this application and their equivalents, this application also intends to include these modifications and variations.

Claims

1. A seat control method, characterized in that, include: Upon receiving a command to flatten the seat, acquire vital sign data of the target cabin area collected by the bio-radar sensor; Based on the vital signs data, life detection is performed to obtain the life detection results within the target cabin area; When the life detection result indicates that there is no life in the target cabin area, the control seat performs a flattening operation.

2. The seat control method according to claim 1, characterized in that, Based on the vital signs data, a life form detection is performed to obtain the life form detection results within the target cabin area, including: Perform a fast Fourier transform on the vital signs data to extract the frequency domain peak values ​​corresponding to respiratory rate and / or heart rate; If the frequency domain peak value is located within the preset vital sign frequency range and continues to exceed the preset time threshold, then the detection result of the vital sign in the target cockpit area is determined to be the presence of a vital sign. If the frequency domain peak value exceeds the preset vital sign frequency range, then the detection result of the vital signs in the target cockpit area is determined to be that there are no living beings.

3. The seat control method according to claim 1, characterized in that, Before the control seat performs the flattening operation, it also includes: Acquire depth image data of the target cockpit area collected by the visual sensor; Based on the depth image data, static obstacles on the moving path of the seat are detected; Based on the detection results of static obstacles, an obstacle avoidance movement path for the seat is generated.

4. The seat control method according to claim 1, characterized in that, The process of controlling the seat to flatten out also includes: Real-time acquisition of operating data of the adjustment motors of each seat in the target cockpit area; If the operating data exceeds the dynamic threshold, the cabin environment image of the target cabin area collected by the visual sensor and the pressure distribution data of the seat collected by the pressure distribution sensor are acquired. Obstacle detection is performed based on the operational data, the cabin environment image, and the pressure distribution data to determine the obstacle type; Perform the corresponding stop or rollback operation based on the type of obstacle.

5. The seat control method according to claim 4, characterized in that, Perform corresponding stop or rollback operations based on the type of obstacle, including: When the obstacle is a hard obstacle, the seat is controlled to stop moving; When the obstacle is a flexible object, the seat is controlled to retract in the opposite direction to a first preset distance. When the obstacle is a living being, the seat is controlled to reverse to a second preset distance and an alarm is triggered; wherein the second preset distance is greater than the first preset distance.

6. The seat control method according to claim 1, characterized in that, The process of controlling the seat to perform the flattening operation includes: Acquire the vehicle's static tilt angle data collected by the inertial measurement unit; Based on the static tilt angle data, determine the current tilt angle of the seat relative to the horizontal plane; Based on the difference between the current tilt angle and the target horizontal tilt angle, the surface of the seat is controlled to remain horizontal relative to the horizontal plane.

7. The seat control method according to claim 1, characterized in that, After the control seat performs the flattening operation, it also includes: The point cloud data of the seat surface collected by the visual sensor is acquired, and a digital elevation model of the seat surface is constructed based on the point cloud data; Based on the digital elevation model, the height difference of the seat surface is determined; If the height difference is greater than the preset flatness threshold, the seat surface will be adjusted to a flat surface.

8. The seat control method according to claim 1, characterized in that, Also includes: When the life detection result indicates the presence of a life form in the target cabin area, the flattening operation is prohibited, and a safety alarm and / or voice prompt is triggered.

9. A seat control system, characterized in that, include: The data receiving module is used to acquire vital sign data of the target cabin area collected by the bio-radar sensor when a seat flattening command is received. The data analysis module is used to detect living beings based on the vital signs data and obtain the detection results of living beings in the target cockpit area; The control execution module is used to control the seat to perform a flattening operation when the life detection result indicates that there is no life in the target cabin area.

10. A vehicle, characterized in that, Includes the seat control system as described in claim 9.

Images