Intelligent airport trolley direction self-adaptive control system

Abstract

The invention discloses an airport intelligent trolley direction self-adaptive control system, and relates to the technical field of intelligent control. A current three-dimensional attitude model of a trolley is obtained; sensing the thrust direction of the user and the movement track change of the cart, and constructing a target intention vector; calculating an expected deflection angle based on attitude and intention vectors, and establishing an environment disturbance model in combination with parameters such as people flow density, channel width, passing inertia direction and transverse wind speed; coupling the deflection angle with the disturbance model through an adaptive adjustment function, and outputting an actual correction angle to drive a steering mechanism to complete direction adjustment; when it is detected that the directional oscillation frequency exceeds a set threshold value, an oscillation buffering strategy is executed, the control frequency and the disturbance weight are dynamically adjusted, and the adjustment function is re-fitted to recover the control stability; according to the invention, the direction of the airport trolley in a complex dynamic environment is accurately regulated and controlled, good environmental adaptability and user interactivity are achieved, and the advancing stability and safety are improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology, specifically to an airport intelligent trolley direction adaptive control system. Background Technology

[0002] Airports, as large transportation hubs with frequent passenger flow and complex spaces, have a significant demand for the use of trolleys to transport carry-on luggage. Traditional trolleys are mostly fixed swivel wheels, requiring users to manually control their direction of travel. Especially during peak hours or in narrow passages, they are prone to drifting, collisions, or even blockages, seriously affecting traffic efficiency and user experience.

[0003] Some airports have already begun introducing "smart trolleys" with electronically assisted systems, attempting to improve safety and convenience through motor-assisted movement and obstacle avoidance. However, in actual use, the following prominent technical problems still exist: When unattended or moving at low speeds, trolleys are prone to directional drift, especially on ramps, around turns, or in areas affected by crosswinds, where the trolley's direction deviates significantly from the user's intended path. Furthermore, existing systems often employ periodic sampling and preset path correction strategies, resulting in response lag and difficulty in dynamically correcting for environmental disturbances in real time. In confined spaces such as boarding gate passages and escalator entrances, insufficient precision in trolley directional control can easily lead to jams and collisions, impacting passenger flow efficiency. Summary of the Invention

[0004] The purpose of this invention is to provide an adaptive direction control system for intelligent airport trolleys to address the shortcomings of the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an airport intelligent trolley direction adaptive control system, comprising: Acquisition module: Acquires the current pose information of the smart trolley, including the center coordinates of the trolley body, forward speed, acceleration, yaw angle, roll angle and pitch angle, and establishes a three-dimensional spatial pose model; Sensing module: Real-time acquisition of changes in the trolley's trajectory per unit time, as well as the angle between the trajectory and the direction of the push force applied by the user. Construct the target intent vector ; Deflection calculation module: Based on the 3D spatial pose model and the target intent vector Calculate the desired deflection angle The desired deflection angle Used to indicate the direction and angle at which the cart needs to be turned; Environmental perception module: Acquires airport environmental characteristic parameters, including the current area's pedestrian density ρ, aisle width W, and direction of inertial movement. In addition to the lateral disturbance wind speed Vw, an environmental disturbance model was established. ; Control module: Determines the desired deflection angle Environmental disturbance model Couple the functions and construct an adaptive adjustment function for the direction. Output the actual correction angle ; Execution driver module: based on actual correction angle The front or rear wheel is driven to adjust its angle and correct the direction of the cart's movement. Update module: If the directional oscillation frequency is detected to exceed the set stability threshold during the correction process. This reduces the control frequency and triggers the oscillation buffer strategy, updating the function. To adapt to the current perturbation characteristics.

[0006] Preferably, the establishment of the three-dimensional spatial attitude model includes: Acquire triaxial linear acceleration and angular velocity data; Based on the quaternion attitude fusion algorithm and combined with the initial reference coordinate system, the yaw angle, pitch angle and roll angle of the trolley are calculated. The three-axis angles are fused with the spatial coordinates of the vehicle center to construct the pose model of the current frame.

[0007] Preferably, the construction of the target intent vector ,include: Get the thrust vector applied by the user to the cart And the thrust data is filtered for mean within a time window; Based on the position coordinate changes in the posture model, the actual change in motion trajectory ΔP per unit time is calculated to form the displacement vector of the current frame. The thrust vector is obtained using the formula for calculating the angle between vectors. The angle between the displacement vector ΔP and the displacement vector ΔP ; At the angle Less than the set angle threshold At that time, confirm that the current thrust direction is valid, and with Construct a normalized target intent vector based on the directional reference. .

[0008] Preferably, based on the current attitude model With the target intent vector Calculate the desired deflection angle ,include: Based on the constructed current 3D pose model, extract the unit vector of the trolley's forward direction. The is the unit vector along the forward axis in the attitude matrix; target intent vector Unit vector relative to the direction of the cart's movement Substituting the vectors into the formula for calculating the angle between them, we can obtain the initial deflection angle. , used to characterize the current directional deviation; Based on the current linear velocity v of the trolley, a sensitivity factor λ is constructed to adjust the deflection angle response intensity; Initial deflection angle By weighting and coupling with the sensitivity factor λ, the final desired deflection angle is obtained. The desired deflection angle Used to indicate the direction and angle at which the cart needs to be turned.

[0009] Preferably, the establishment of the environmental disturbance model ,include: Acquire images of the environment surrounding the cart and extract the current area's pedestrian density ρ using a pedestrian density recognition algorithm; Identify the current location channel boundary and calculate the spatial channel width W; Based on the average movement direction of adjacent pedestrians over multiple historical time points, the regional inertial direction of travel is extracted. ; Obtain the lateral disturbance wind speed Vw, and combine the parameters ρ, W, Combined with Vw, a set of perturbation impact vectors is constructed and normalized to form the current frame environment perturbation model. .

[0010] Preferably, the desired deflection angle is... Environmental disturbance model Couple the functions and construct an adaptive adjustment function for the direction. Output the actual correction angle ,include: Desired deflection angle Perturbation model of the current frame environment The perturbation parameter set is input into the multi-factor weighted adjustment function framework; Based on the sensitivity of each disturbance factor to directional control, weighting coefficients are set, and adaptive correction coefficients μ are constructed. Desired deflection angle By functionally coupling with the correction coefficient μ, a direction adaptive adjustment function is formed. Used to dynamically adjust the control output intensity; Actual correction angle after output coupling .

[0011] Preferably, the correction angle based on the actual angle Adjusting the angle of the front or rear wheels to correct the direction of the stroller's movement includes: Receive actual correction angle Based on the current vehicle attitude model's direction of travel and speed information, the actual correction angle will be... The signal is converted into a motor control signal, which drives the rudder to adjust the steering wheel via pulse width modulation. The adjustment rate is set jointly based on the angle change rate and vehicle speed. The steering execution status feedback signal is monitored; if the steering angle deviation exceeds a preset deviation threshold, the control signal is automatically corrected to ensure wheel assembly response is consistent with... They reached a consensus and ultimately corrected their course.

[0012] Preferably, the update function To adapt to the current perturbation features, including: During the steering adjustment process, the actual steering angle change sequence is collected in real time, and the directional oscillation frequency per unit time is calculated based on the short-time Fourier transform algorithm. ; oscillation frequency With the set stability threshold If a comparison is made, Greater than The current directional control is determined to have high-frequency oscillations. In high-frequency oscillation mode, the update frequency of direction control commands is reduced and the adjustment weights of non-critical disturbance factors are frozen to form an oscillation buffer strategy. Refit the adaptive adjustment function based on the current perturbation characteristics. Adjust the sensitivity coefficients of each environmental factor.

[0013] The technical effects and advantages provided by the present invention in the above technical solution are as follows: 1. This invention constructs a direction adaptive control method based on attitude perception, user intent parsing, and environmental disturbance coupling, achieving dynamic and precise adjustment of the travel direction of intelligent airport trolleys. The system can not only identify the difference between the direction of the user-applied thrust and the trolley's trajectory in real time, but also incorporate environmental factors such as pedestrian density, spatial constraints, inertial direction, and lateral wind speed to perform multi-factor fusion correction of the desired deflection angle, significantly improving the trolley's directional control accuracy and stability in high-traffic, complex spatial environments.

[0014] 2. This invention also introduces an oscillation detection and adaptive buffering mechanism, which can dynamically adjust the control frequency and adjustment function weights when high-frequency directional oscillations are detected, thereby effectively suppressing overcompensation and directional fluctuations, and improving the system's robustness and anti-interference capability under extreme disturbance scenarios. The overall solution does not rely on an external control center or trajectory preset, and has the technical advantages of high autonomy, flexible deployment, and user-friendly experience, making it suitable for intelligent hand-pushing auxiliary equipment in high-density traffic locations such as airports and train stations. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0016] Figure 1 This is a flowchart of the system modules of the present invention. Detailed Implementation

[0017] 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, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] For examples, please refer to Figure 1 As shown in this embodiment, an airport intelligent trolley direction adaptive control system includes: Acquisition Module: Acquires the current pose information of the smart trolley, including the center coordinates of the trolley body, forward speed, acceleration, yaw angle, roll angle and pitch angle, and establishes a three-dimensional spatial pose model.

[0019] The real-time position coordinates (x, y, z) of the trolley in three-dimensional space are obtained by a positioning device (such as a UWB module or a high-precision GPS module) installed below the center of the trolley body; at the same time, combined with the acceleration data output by the wheel speedometer and the inertial measurement unit (IMU), the forward velocity v and linear acceleration a of the trolley per unit time are calculated to describe the motion state of the trolley; the yaw angle, pitch angle and roll angle are calculated by the subsequent attitude fusion algorithm.

[0020] The IMU unit includes a three-axis accelerometer and a three-axis gyroscope, which output linear accelerations (Ax, Ay, Az) and angular velocities (Gx, Gy, Gz) along the x, y, and z directions, respectively. To improve data stability, the raw data is first preprocessed using a sliding window averaging filter to remove random high-frequency jitter, and the sensor data is continuously recorded at a sampling frequency of 100Hz as the basic input for attitude calculation.

[0021] Attitude calculation employs a quaternion attitude fusion algorithm to avoid the gimbal lock problem inherent in Euler angles. During initialization, the ground gravity direction is used as a reference, and the initial quaternion q = (q0, q1, q2, q3). During operation, the minute rotational increment dq of attitude change is calculated using the three-axis angular velocities (Gx, Gy, Gz), and attitude updates are performed through quaternion multiplication. The conversion relationship between quaternions and Euler angles is as follows: Yaw angle ; Pitch angle ; Roll angle .

[0022] The obtained angle values ​​are updated in each frame and synchronized with the timestamp.

[0023] The acquired three-axis angles (Yaw, Pitch, Roll) are integrated with the center coordinates (x, y, z) to construct the current three-dimensional spatial attitude model. Specifically, a 4×4 homogeneous transformation matrix T is constructed, containing a rotation matrix R and a displacement vector t, where R is the attitude rotation matrix derived from the three-axis angles. Let the position vector be the attitude model. Represented as: The posture model is updated every 10ms and is provided as a posture reference benchmark for the subsequent user intent vector construction module.

[0024] Sensing module: Real-time acquisition of changes in the trolley's trajectory per unit time, as well as the angle between the trajectory and the direction of the push force applied by the user. Construct the target intent vector .

[0025] In this embodiment of the invention, to accurately identify the user's actual pushing intention, a target intention vector is constructed by acquiring thrust information based on the perception module and combining it with pose change data. Its implementation process includes: A three-axis force sensor is integrated inside the trolley handle to collect the thrust vector applied by the user to the trolley in real time. It contains components in the X, Y, and Z directions, denoted as To improve data stability, a sliding time window mean filtering algorithm was used to process the raw thrust data. The filter window size was set to 200ms, and the update frequency was 10Hz. This filtering strategy can effectively smooth thrust disturbances caused by slight hand tremors and extract stable thrust direction features.

[0026] Using position coordinate information from the attitude model of the previous cycle and the current cycle and The change in the trajectory ΔP per unit time is calculated using the following formula: The vector ΔP represents the actual direction and distance the cart moves within a unit time Δt, serving as a spatial representation of the cart's self-motion state. This vector will be used to determine the consistency between the user's intended direction and the cart's actual direction of movement.

[0027] The obtained thrust vector Substitute the calculated displacement vector ΔP into the formula for calculating the angle between the vectors: Where "·" represents the vector dot product, and arccos represents the inverse cosine function. The obtained angle... This characterizes the degree of difference between the direction of force applied by the user and the actual direction of movement of the cart. To prevent division by zero errors during the angle calculation process, the system determines the current frame data as an invalid frame and excludes it from subsequent calculations when it detects that the magnitude of any vector is less than the threshold ε (e.g., 0.05 N or 0.05 m).

[0028] When the included angle Less than the set angle threshold At this point, it is assumed that the direction of the thrust is consistent with the actual direction of motion, indicating that the user's control over the cart is clear and effective. At this time, the original thrust vector is... Normalization is performed to obtain the unit target intent vector. The specific calculation formula is as follows: ;in For the thrust vector magnitude, It is a three-dimensional unit vector used in subsequent deflection angles. In the calculation, it is used as a reference input for the target direction. Angle threshold This is a system adjustable parameter, and the default setting is [value]. It can be dynamically optimized and adjusted according to different airport passage widths, usage scenarios, or user operating habits.

[0029] Deflection calculation module: Based on the 3D spatial pose model and the target intent vector Calculate the desired deflection angle The desired deflection angle Used to indicate the direction and angle at which the cart needs to be turned.

[0030] In this embodiment of the invention, to achieve real-time dynamic correction of the directional deviation of the intelligent handcart, the desired deflection angle is calculated through the vector relationship between posture and user intent. The specific process is as follows: A 3D posture model has been constructed in the acquisition module. The forward direction of the trolley is represented by the first column vector of the rotation matrix R, denoted as: ;in, These represent the rotational components along the X, Y, and Z axes in the attitude matrix, respectively. To ensure calculation accuracy, this vector needs to be normalized, i.e.: Finally, the unit vector of the cart's current direction of movement is obtained. It is used to compare the direction with the user intent vector.

[0031] Call the target intent vector output by the perception module , and obtained Perform angle calculation to obtain the initial deflection angle of the current frame. The specific formula is: ;because and All values ​​have been normalized, therefore the formula can be simplified to: Where "·" represents the vector dot product, and the resulting angle is... It represents the spatial deviation between the current direction of the trolley's movement and the direction expected by the user.

[0032] This invention introduces a velocity-dependent sensitivity factor λ to dynamically suppress the initial deflection angle. The value of this factor λ ranges from 0 to 1, and its specific calculation method is as follows: Where: v is the current linear velocity of the cart; The maximum stable operating speed is set (e.g., 1.5 m / s); as v increases, λ decreases to limit the deflection amplitude and improve stability; the lower limit of 0.2 is used to prevent complete suppression of the response.

[0033] Finally, the initial deflection angle The result is linearly weighted with a sensitivity factor λ, and the output is the desired deflection angle used by the subsequent control module. The calculation method is as follows: ; the result The target directional deflection angle required for the trolley is represented by radians or angles, depending on the control module settings. It serves as a direct input parameter for directional adjustment, driving the execution module to perform steering corrections.

[0034] Environmental perception module: Acquires airport environmental characteristic parameters, including the current area's pedestrian density ρ, aisle width W, and direction of inertial movement. In addition to the lateral disturbance wind speed Vw, an environmental disturbance model was established. .

[0035] A wide-angle camera is installed at the front of the cart to collect image data of the area in front in real time. A pre-trained convolutional neural network (such as YOLOv5) is used for target detection to identify the number N of human targets in the image. Combined with the actual area of ​​the camera's field of view S, the population density per unit area is calculated, defined as the population density ρ, and its calculation formula is: Where S is in square meters and N is the number of pedestrians detected. The ρ value is updated every 100ms to adapt to dynamic changes in airport pedestrian flow.

[0036] To obtain the spatial constraints of the current travel area of ​​the trolley, geometric modeling of the current passage is performed by fusing LiDAR point cloud data and high-precision map information. The system uses a grid map projection method to identify the left and right boundary lines and calculates the shortest horizontal distance between the boundaries, defined as the passage width W, in meters. If the boundaries are irregular, the minimum passage width within a 1-second time window is taken as the current frame W value to improve obstacle avoidance capability in narrow areas.

[0037] The motion trajectories of multiple adjacent pedestrians are identified in the perceived image, and their movement vectors are extracted using optical flow tracing or continuous frame target matching methods. Within a time window T (e.g., 2 seconds), calculate the average direction of all vectors, which serves as the regional travel inertial direction. Right now: ; Characterizing the overall movement trend of a crowd, directional stability can be used to help determine the adaptability of the stroller's direction of travel.

[0038] The lateral disturbance wind speed Vw (in m / s) is acquired in real time, perpendicular to the direction of the trolley's movement. Finally, the four disturbance parameters are combined into a disturbance vector set: To unify the scale, the system uses min-max normalization to map each parameter to the interval [0,1], thus constructing the final normalized environmental perturbation model: ; The normalized pedestrian density ρ, spatial passage width W, and direction of inertia are... And the lateral disturbance wind speed Vw.

[0039] Control module: Determines the desired deflection angle Environmental disturbance model Couple the functions and construct an adaptive adjustment function for the direction. Output the actual correction angle .

[0040] First, the desired deflection angle output by the deflection calculation module is... The current frame perturbation model output by the environment perception module Simultaneously input the adjustment function to construct the unit. Perturbation model. A perturbation vector set consisting of four normalized environmental factors: ;in: : Normalized population density in the current area; The reciprocal of the normalized channel width indicates the degree of space constraint; Normalized travel inertia direction modulus reflects the stability of crowd movement trends; : Normalized value of lateral disturbance wind speed.

[0041] To reflect the influence of each disturbance factor on the direction deflection control, a weighting coefficient is set for each disturbance parameter. , respectively corresponding to These weights can be fitted using historical data or set manually; the default configuration is as follows: (Personnel density weight): 0.3; (Channel contraction weight): 0.25; (Population flow trend weight): 0.2; (Wind speed disturbance weight): 0.25; Based on this, the disturbance correction coefficient μ for the current frame is calculated, with the specific expression as follows: The value of μ ranges approximately from 0 to 1, representing the value of μ. The degree of dynamic gain adjustment.

[0042] By constructing an adaptive adjustment function The desired deflection angle It is coupled with the perturbation correction coefficient μ. The coupling method is a linear enhancement model, expressed as: This function dynamically amplifies or suppresses the original deflection angle based on the complexity of the environment: when environmental disturbances are significant (such as high pedestrian density or confined space), μ increases. The angle is amplified to enhance the correction capability; when the environment is stable, μ approaches 0, and the system maintains the original deflection angle. Keep it unchanged, and avoid excessive control.

[0043] The results of the coupling calculation As an input angle command to control the electronically controlled steering system (such as servo-driven casters), it enables real-time fine-tuning of the trolley's direction of travel.

[0044] Output value The unit is angle or radian. The automatic matching format is set according to the drive control unit, and the output value of each frame is recorded for abnormal fluctuation detection and directional stability analysis.

[0045] Execution driver module: based on actual correction angle The front or rear wheel is driven to adjust its angle and correct the direction of the pushcart's movement.

[0046] First, the actual correction angle calculated by the receiving control module is obtained. This angle characterizes the angle at which the cart should turn. To ensure safety and structural rationality, the controller... Boundary restriction processing is performed, and the maximum adjustment range is set to ±30°.

[0047] Then, the current travel direction vector in the attitude model is read. The system determines the steering priority based on the trolley's linear velocity v, prioritizing either front wheel steering or tail wheel fine-tuning to adapt to different scenarios (such as turning in place or slow following). For example, if v > 0.5 m / s, the front wheel steering mechanism is driven first; if v ≤ 0.5 m / s, or in a confined space, the tail wheel participates in steering coordination control.

[0048] Will Converted to the control target displacement angle of the rudder motor (Unit: degrees), and calculate the target adjustment rate in conjunction with the current attitude update frequency. (Unit: degrees per second).

[0049] Then, the motor driver outputs a control signal via PWM (Pulse Width Modulation) to adjust the duty cycle and precisely control the motor's response angle. The control signal is updated every 10 ms to ensure real-time performance. Motor adjustment rate... It can be given by the following formula: ;in: The angle change between the current frame and the previous frame; v is the current linear velocity of the cart; k is an empirical coefficient (e.g., 20) used to adjust the rate sensitivity; the upper limit of the rate is limited by hardware to avoid directional oscillation caused by excessively fast rotation.

[0050] The rudder motor is equipped with an angle encoder module for real-time feedback of the current operating angle. , and the set angle By comparison, the deviation value is obtained. .like If the deviation exceeds a preset deviation threshold ε (e.g., 2.5°), it is determined that the action deviates from the target, triggering an error compensation mechanism to automatically adjust the PWM signal output to correct the deviation.

[0051] By using feedback correction closed-loop control, the actual rotation angle of the final steering mechanism is ensured. With the control module settings Maintain consistency (error less than 2°). Once the control accuracy meets the requirements, the execution module marks the current state as "stable" and sends a success flag to the upper control module for subsequent determination of whether to continue adjustment.

[0052] Ultimately, the accurate adjustment of the wheel angle enabled continuous and stable correction of the trolley's direction of travel, making it more in line with the user's intentions.

[0053] Update module: If the directional oscillation frequency exceeds the set threshold during the correction process... This reduces the control frequency and triggers the oscillation buffer strategy, updating the function. To adapt to the current perturbation characteristics.

[0054] Control angle for each output direction Then, the actual steering feedback angle sequence is acquired through the encoder of the rudder motor. A continuous time series window is formed. The sampling period is 10 ms, the window length is 2 seconds, and a total of 200 sampling points are acquired.

[0055] The angle sequence was analyzed in the frequency domain using Short-Time Fourier Transform (STFT) to extract its dominant frequency component, thus obtaining the dominant oscillation frequency of the direction control output per unit time. This frequency reflects whether the steering control exhibits rapid back-and-forth movements or small high-frequency fluctuations, and is an important basis for determining control instability. A stability threshold is set. This serves as the upper limit for judging oscillating behavior, with a typical value of 2 Hz. If oscillation behavior is detected... This means that if the direction control is found to have high-frequency oscillations, possibly due to overcompensation, excessive disturbance, or excessively fast control frequency, the oscillation buffer mechanism will be triggered. The oscillation determination condition is: if... (like If the Hz value is 2 Hz, then a stability intervention strategy will be implemented.

[0056] To quickly suppress high-frequency oscillations, the following two-level buffering strategy is implemented: Update frequency suppression: The update frequency of the direction control command is reduced from the normal 100 Hz to 25 Hz, which reduces the response speed of the control system and avoids oscillation caused by rapid command switching.

[0057] Weight freezing mechanism: This applies to the adjustment function. Non-critical disturbance factors (such as wind speed) Traffic inertia The weight of ) is set to zero, and only key factors (such as crowd density) are retained. The regulatory effect of [the mechanism] makes the control strategy more stable. This strategy is executed for 3 seconds, or until the oscillation frequency [is reached]. It has fallen back below the stable range.

[0058] After the oscillation buffering strategy is implemented, the environmental disturbance model is re-evaluated. The changing trends of each factor were analyzed, and the weight coefficients in the adjustment function were adjusted using an adaptive fitting method with a penalty factor. Adjustments were made. The adjusted weights. It can be set as follows: ;in: The original default weights; The frequency deviation exceeding the oscillation threshold; η is the dynamic penalty factor (recommended value is 0.3); when The larger the value, the more severe the oscillation, and the greater the corresponding reduction in weight.

[0059] After completing the weight adjustment, reconstruct the adjustment function: ; where μnew is the recalculated disturbance correction coefficient, ensuring that the adjustment function adapts to the current micro-disturbance environment, and achieving smooth transition control by gradually restoring the normal control frequency.

[0060] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. An airport intelligent trolley direction adaptive control system characterized by: Comprise: The acquisition module: acquire the current pose information of the intelligent trolley, including the trolley body center coordinates, forward speed, acceleration, yaw angle, roll angle and pitch angle, and establish a three-dimensional space attitude model; The perception module: real-time acquisition of the change amount of the trolley motion trajectory in a unit time, and the included angle with the pushing force direction applied by the user , construct the target intention vector ; deflection calculation module: according to three-dimensional space posture model and target intention vector calculate the expected deflection angle , the expected deflection angle for representing the direction and angle of the trolley to be turned An environmental perception module: obtain airport environmental characteristic parameters, including current area people flow density , space channel width W, traffic inertia direction , and transverse disturbance wind speed Vw, establish an environmental disturbance model ; Control module: desired deflection angle with the environmental disturbance model is coupled to build a direction adaptive adjustment function , output actual correction angle ; Performing driving module: based on actual correction angle , drive the front wheel or tail wheel to adjust the angle, correct the direction of the cart Update module: if during the correction process the frequency of the directional oscillations is detected to exceed a set stability threshold then the control frequency is reduced and an oscillation damping strategy is triggered, the update function is adapted to the current perturbation characteristics.

2. The airport intelligent trolley direction adaptive control system according to claim 1, characterized in that: The three-dimensional space attitude model comprises: Acquire three-axis linear acceleration and angular velocity data; Based on the quaternion attitude fusion algorithm, combined with the initial reference coordinate system, the yaw angle, pitch angle and roll angle of the trolley are calculated; The three-axis angle and the trolley body center space coordinates are fused by matrix to construct the current frame attitude model.

3. The airport intelligent trolley direction adaptive control system according to claim 1, wherein: The build target intent vector comprises: Acquiring a push force vector exerted by a user on a stroller and performing a time windowed mean filtering on the push force data; Based on the position coordinate change in the attitude model, the actual motion trajectory change quantity ΔP in unit time is calculated to form the displacement vector of the current frame; The thrust vector is calculated by using the vector angle calculation formula the angle between the displacement vector ΔP ; At the included angle less than a set angle threshold , it is confirmed that the current thrust direction is valid, and a normalized target intention vector is constructed with as the direction reference .

4. The airport intelligent trolley direction adaptive control system according to claim 3, wherein: According to the current pose model With the target intention vector Calculate the expected deflection angle , comprising: Based on the constructed current three-dimensional pose model, a trolley advancing direction unit vector is extracted , the is a unit vector in the advancing axis direction of the pose matrix; The target intention vector The trolley advancing direction unit vector Substitute the vector included angle calculation formula to obtain the initial deflection angle , for representing the current direction deviation; Combined with the current linear speed v of the trolley, the sensitivity factor λ is constructed to adjust the response strength of the deflection angle; The initial deflection angle is coupled with a weighting factor λ to obtain a final desired deflection angle , which is coupled with the sensitivity factor λ to obtain a final desired deflection angle is used to characterize the direction and angle in which the cart needs to be turned.

5. The airport intelligent trolley direction adaptive control system according to claim 4, wherein: The environment disturbance model is established , comprising: Acquire the image of the environment around the trolley, and extract the current area people flow density ρ through the people flow density recognition algorithm; Identify the current location channel boundary and calculate the space channel width W; Based on the average moving direction of adjacent pedestrians in multiple historical moments, the regional traffic inertia direction is extracted ; Obtain the lateral disturbance wind speed Vw, and parameters p, W, Construct the disturbance influence vector group with Vw, and normalize to form the current frame environment disturbance model .

6. The airport intelligent trolley direction adaptive control system according to claim 1, wherein: The desired deflection angle With the environmental disturbance model Coupling is carried out, and the direction adaptive adjustment function is constructed , Output actual correction angle , Including: The desired deflection angle The disturbance parameters in the current frame environment disturbance model are input to a multi-factor weighted adjustment function framework; According to the sensitivity of each disturbance factor to the direction control, set the weight coefficient to construct the adaptive correction coefficient μ; The desired deflection angle The function coupling with the correction coefficient μ forms a direction adaptive adjustment function For dynamically adjusting the control output intensity output coupling actual correction angle .

7. The airport intelligent trolley direction adaptive control system according to claim 1, wherein: The actual correction angle The method for driving the front wheel or the tail wheel to adjust the angle and correct the travel direction of the stroller comprises: Receiving actual correction angle According to the travel direction and vehicle speed information in the current vehicle body posture model, the actual correction angle is converted into a motor control signal, and the steering adjustment is performed by driving the rudder motor in a pulse width modulation manner. The adjustment rate is set according to the angle change rate and the vehicle speed. The steering execution state feedback signal is monitored. If it is detected that the steering angle deviation exceeds the preset deviation threshold, the control signal is automatically corrected to ensure that the wheel set response is consistent with the travel direction correction is finally completed.

8. The airport intelligent trolley direction adaptive control system according to claim 1, wherein: The update function to adapt to current perturbation characteristics, including: In the process of executing the steering adjustment, the actual steering angle change sequence is collected in real time, and the direction shock frequency per unit time is calculated based on a short-time Fourier transform algorithm ; The oscillation frequency is compared with a set stability threshold , and if is greater than , it is determined that there is high-frequency oscillation in the current directional control. In the high frequency oscillation state, reduce the update frequency of the direction control instruction, and freeze the adjustment weight of the non-key disturbance factor, form the oscillation buffer strategy; Re-fitting adaptive adjustment function based on current perturbation characteristics adjusting sensitivity coefficients of various environmental factors.

Images