A method and system for active anti-rollover control of forklifts based on multi-signal fusion

Abstract

This invention discloses a forklift active anti-rollover control method and system based on multi-signal fusion. It collects multi-source signals from the forklift in real time, including the forklift's lateral tilt angle, real-time load, fork lifting height, speed, and steering angle. An initial safe tilt angle threshold is calculated based on the real-time load and fork lifting height. This initial safe tilt angle threshold is then corrected by considering the speed and steering angle to obtain the final dynamic safe tilt angle threshold under the current operating conditions. A risk fusion calculation is performed based on the forklift's current lateral tilt angle and the final dynamic safe tilt angle threshold to obtain a comprehensive risk score. The risk level is assessed according to the interval of the comprehensive risk score, and corresponding dynamic intervention actions are executed accordingly. This invention ensures the reliability of the decision-making and triggering channels through an independent hardware architecture, and designs a physically independent, independently controlled, and independently computed safety controller. Through a graded safety response sequence, it actively intervenes in the accident chain to eliminate secondary injuries.

Description

Technical Field

[0001] This invention belongs to the field of engineering machinery safety control technology, and in particular relates to a method and system for active anti-rollover control of forklifts based on multi-signal fusion. Background Technology

[0002] Forklifts, as high-center-of-gravity and highly dynamic material handling equipment, pose an extremely high risk of tipping over under complex working conditions, making them a major cause of serious personal injury and property damage. Currently, the industry primarily relies on operator experience and passive safety devices such as seat belts and overhead guards, lacking proactive safety technologies that can automatically sense, assess, and intervene before the tipping risk reaches a critical point. Existing warning devices suffer from issues such as susceptible signal interference, simplistic decision-making logic, and inability to coordinate with actuators, thus failing to effectively prevent accidents. Therefore, a tipping-over prevention system capable of reliable sensing, intelligent decision-making, and proactive control is needed to address this problem.

[0003] The existing technology, with application number CN202211164080.5 and titled "A Method for Constructing a Digital Twin of a Forklift Based on a Flight Control System," constructs a digital twin of the forklift, runs a digital twin application, generates a result file, and obtains recommended travel speed values ​​for the forklift's current lifting capacity and predicts rollover based on the evaluation report and result file. This can prevent dangerous accidents without reducing forklift transportation efficiency. However, using a fixed threshold for judgment makes it susceptible to vibration or electromagnetic interference, prone to false alarms or missed alarms; moreover, it can only reflect the current attitude, with a serious perception lag, and cannot predict rollover trends, failing to achieve an advanced early warning effect. Summary of the Invention

[0004] The purpose of this invention is to provide a forklift active anti-rollover control method and system based on multi-signal fusion, which can realize a closed-loop control system with accurate perception, intelligent decision-making, reliable triggering and active handling, changing the forklift safety concept from mitigating injury after an accident to preventing accidents before they happen and intervening during accidents.

[0005] This invention adopts the following technical solution: a forklift active anti-rollover control method based on multi-signal fusion, comprising:

[0006] S1: Real-time acquisition of multi-source signals from the forklift, including vehicle lateral tilt angle, real-time load, fork lifting height, vehicle speed, and steering angle;

[0007] S2: Obtain the first set of correction parameters based on the forklift's speed and steering angle, and dynamically correct the initial speed correction coefficient and initial steering angle correction coefficient obtained from the actual vehicle calibration based on the first set of correction parameters to obtain the final speed correction coefficient and final steering angle correction coefficient.

[0008] S3: Obtain the initial safe tilt angle threshold based on the real-time load and fork lifting height. Combine the current vehicle speed and current steering angle, and correct the initial safe tilt angle threshold using the final correction coefficients for vehicle speed and steering angle to obtain the final dynamic safe tilt angle threshold under the current working conditions.

[0009] S4: Perform risk fusion calculation based on the current vehicle body lateral tilt angle and the final dynamic safety tilt angle threshold to obtain a comprehensive risk score; assess the risk level based on the comprehensive risk score and execute corresponding dynamic intervention actions according to the level.

[0010] Furthermore, the process for obtaining the final coefficients of vehicle speed correction and steering angle correction is as follows: A first set of correction parameters is obtained through amplitude-limited PID and online learning. The initial coefficients of vehicle speed correction and steering angle correction are corrected using the obtained first set of correction parameters to obtain the final coefficients of vehicle speed correction and steering angle correction. The first set of correction parameters includes the online learning correction amount of vehicle speed coefficient, the PID lag compensation correction amount of vehicle speed coefficient, the online learning correction amount of steering coefficient, and the PID lag compensation correction amount of steering coefficient.

[0011] Furthermore, the steps for obtaining the first set of modified parameters include:

[0012] S201: Based on the collected data of the forklift's body lateral tilt angle, speed, and steering angle, a real-time state observer is constructed to obtain a set of real-time state parameters for adaptively correcting the initial coefficients of the speed correction and steering angle correction; the set of real-time state parameters includes the body lateral tilt angle change rate, attitude prediction error, and vehicle response lag time.

[0013] S202: Based on the real-time state parameter set, establish dynamic mapping relationships between vehicle speed and steering angle and optimal correction coefficients respectively, and output the online learning correction amount of the correction coefficients, including the online learning correction amount of vehicle speed coefficient and the online learning correction amount of steering coefficient;

[0014] S203: Using the attitude prediction error and the rate of change of the vehicle body lateral tilt angle obtained from S201 as inputs, and the vehicle response lag time as the feedforward adjustment amount, an incremental PID compensation controller is constructed. The PID parameters are adaptively adjusted according to the lag time and the corresponding PID lag compensation correction amount is output, including the PID lag compensation correction amount of the vehicle speed coefficient and the PID lag compensation correction amount of the steering coefficient, and finally the first set of correction parameters is obtained.

[0015] Furthermore, in step S201, the first derivative of the vehicle body lateral tilt angle is performed to obtain the rate of change of the vehicle body lateral tilt angle, the difference between the real-time vehicle body lateral tilt angle and the initial safe tilt angle threshold is used to obtain the attitude prediction error, and the vehicle response lag time is calculated by the time phase difference between the steering angle input and the tilt angle response.

[0016] Furthermore, in step S202, a sample library of historical operating conditions and correction coefficients is constructed using a circular buffer. An effective sample library is obtained by filtering through a real-time state parameter set. The sample data is then fitted online using the least squares method to establish a dynamic mapping relationship between vehicle speed and steering angle and the optimal correction coefficient, thereby obtaining the online learning correction amount of the corresponding correction coefficient.

[0017] Furthermore, in step S203, the attitude prediction error obtained in S201 is used as the input to the proportional and integral components of the incremental PID controller, the tilt angle change rate is used as the differential term input, and the vehicle response lag time is used as the feedforward adjustment. The PID proportional, integral, and differential parameters are adaptively adjusted according to the real-time estimated vehicle response lag time, and the corresponding output vehicle speed coefficient PID lag compensation correction amount and steering coefficient PID lag compensation correction amount are output.

[0018] Furthermore, in S3, based on the currently collected real-time load and fork lifting height, the corresponding initial safe tilt angle threshold is obtained through the established forklift stability tilt angle mapping model.

[0019] Furthermore, S4 obtains a comprehensive risk score based on the current vehicle body lateral tilt angle and the final dynamic safety tilt angle threshold.

[0020] Furthermore, in step S4, risk level assessment is performed and corresponding dynamic intervention actions are executed in stages, including: risk levels are divided into Level 1 warning, Level 2 alert and Level 3 emergency.

[0021] If the risk score is less than or equal to the first threshold, the risk level is assessed as safe and no intervention action is taken.

[0022] If the risk score is greater than the first threshold and less than or equal to the second threshold, the risk level is assessed as a Level 1 warning, triggering the audible and visual alarm.

[0023] If the risk score is greater than the second threshold and less than or equal to the third threshold, the risk level is assessed as Level 2 warning, the driving power is limited to the preset upper limit, and the forks are controlled to descend at the first preset speed.

[0024] If the risk score exceeds the third threshold, the risk level is assessed as Level 3 Emergency, triggering active braking and simultaneously controlling the forks to descend at a second preset speed. The second preset speed is greater than the first preset speed.

[0025] A forklift active rollover prevention control system based on multi-signal fusion, applicable to the aforementioned forklift active rollover prevention control method, includes:

[0026] The perception layer includes an IMU for measuring the lateral tilt angle of the forklift body, a load sensor for acquiring the real-time load of the forklift, a speed sensor for acquiring the real-time speed of the forklift, a displacement sensor for acquiring the lifting height of the forklift forks, and a steering angle sensor for acquiring the steering angle of the forklift.

[0027] The control layer runs a risk fusion algorithm, including: calculating the initial safe tilt angle threshold based on the real-time load of the forks and the lifting height of the forks in the current state of the vehicle; dynamically correcting the initial safe tilt angle threshold according to the current vehicle speed and steering angle to obtain the final dynamic safe tilt angle threshold; and obtaining the risk assessment level based on the final dynamic safe tilt angle threshold.

[0028] The execution layer includes a motor driver for limiting the travel power to a preset upper limit, an electro-hydraulic valve for controlling the lifting of the forks, a braking module for active braking, and an audible and visual alarm for audible and visual alerts.

[0029] This invention achieves proactive and accurate risk perception through multi-source signal fusion and dynamic models, realizing a leap from measuring the state to predicting risk. The system maps continuous comprehensive risk scores to discrete risk levels, with each level corresponding to a set of preset action combinations. It actively consumes the kinetic and potential energy of the vehicle and cargo, maximizing vehicle stability before or during a rollover, preventing cargo from falling and vehicle from sliding, thereby directly cutting off the secondary injury chain. Attached Figure Description

[0030] To more clearly illustrate the technical solutions in the embodiments of the present invention or related technologies, the drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0031] Figure 1 This is a flowchart of the active anti-rollover control method for forklifts provided by the present invention;

[0032] Figure 2 This is a schematic diagram of the active anti-rollover control system for forklifts provided by the present invention. Detailed Implementation

[0033] To make the objectives, technical solutions, and advantages of the embodiments described herein clearer, the technical solutions in the embodiments will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments described herein, not all of them. All other embodiments obtained by those skilled in the art based on the embodiments described herein without creative effort are within the scope of protection of this document. It should be noted that, unless otherwise specified, the embodiments and features described herein can be arbitrarily combined with each other.

[0034] The following is in conjunction with the appendix Figure 1-2 The present invention will be described in detail with reference to the embodiments:

[0035] A forklift active rollover prevention control method based on multi-signal fusion includes:

[0036] S1: Real-time acquisition of multi-source signals from the forklift, including vehicle lateral tilt angle, real-time load, fork lifting height, vehicle speed, and steering angle;

[0037] S2: Obtain the first set of correction parameters based on the forklift's speed and steering angle, and dynamically correct the initial speed correction coefficient and initial steering angle correction coefficient obtained from the actual vehicle calibration based on the first set of correction parameters to obtain the final speed correction coefficient and final steering angle correction coefficient.

[0038] S3: Obtain the initial safe tilt angle threshold based on the real-time load and fork lifting height. Combine the current vehicle speed and current steering angle, and correct the initial safe tilt angle threshold using the final correction coefficients for vehicle speed and steering angle to obtain the final dynamic safe tilt angle threshold under the current working conditions.

[0039] S4: Perform risk fusion calculation based on the current vehicle body lateral tilt angle and the final dynamic safety tilt angle threshold to obtain a comprehensive risk score; assess the risk level based on the comprehensive risk score and execute corresponding dynamic intervention actions according to the level.

[0040] In this invention, the collected steering angle refers to the real-time deflection angle of the forklift steering wheel relative to the longitudinal centerline of the vehicle, which is used to characterize the current steering intensity of the vehicle.

[0041] In this invention, the initial correction coefficients are obtained through actual vehicle calibration, and these initial correction coefficients include the vehicle speed correction initial coefficients. and steering angle correction initial coefficient The initial coefficient for vehicle speed correction is corrected by limiting PID and online learning. and steering angle correction initial coefficient This yields online learning correction values ​​for vehicle speed coefficients. PID lag compensation correction amount for vehicle speed coefficient Online learning correction amount for steering coefficient and steering coefficient PID lag compensation correction amount The first set of correction parameters; the steps for obtaining the first set of correction parameters include:

[0042] S201: Construct a real-time state observer based on the collected forklift body lateral tilt angle, vehicle speed and steering angle data to obtain a set of real-time state parameters for adaptively correcting the initial coefficients of vehicle speed correction and steering angle correction.

[0043] S202: Based on the real-time state parameter set, establish dynamic mapping relationships between vehicle speed and steering angle and optimal correction coefficients respectively, and output the online learning correction amount of the corresponding correction coefficients, including the online learning correction amount of vehicle speed coefficient and the online learning correction amount of steering coefficient;

[0044] S203: Using the attitude prediction error and the rate of change of vehicle body lateral tilt angle obtained from S201 as inputs, and the vehicle response lag time as the feedforward adjustment amount, an incremental PID compensation controller is constructed. The PID parameters are adaptively adjusted according to the lag time, and the corresponding PID lag compensation correction amounts of vehicle speed coefficient and steering coefficient are output, and finally the first correction parameter set is obtained.

[0045] In step S201, the real-time state parameter set includes the vehicle body lateral tilt rate of change, attitude prediction error, and vehicle response lag time. Specifically, the vehicle body lateral tilt rate of change is obtained by performing the first derivative of the vehicle body lateral tilt angle, the attitude prediction error is obtained by subtracting the real-time vehicle body lateral tilt angle from the initial safe tilt angle threshold, and the vehicle response lag time is calculated by the time phase difference between the steering angle input and the tilt angle response.

[0046] In this embodiment, the dynamic instability trend of the forklift is observed in real time, providing accurate, real-time and reliable feedback for subsequent adaptive correction. This solves the problem that traditional fixed thresholds cannot reflect the true dangerous trend of the vehicle, enabling the correction action to follow the real-time changes in the vehicle's posture. The above-mentioned state variables together constitute the feedback basis for the adaptive adjustment of coefficients, which in real time characterizes the strength of the current rollover trend, the magnitude of the correction deviation and the dynamic response characteristics of the system. The rate of change and error are used to determine whether the current coefficient is too loose or too tight.

[0047] In step S202 of this invention, a sample library of historical operating conditions and correction coefficients is constructed using a ring buffer. An effective sample library is obtained by filtering through a real-time state parameter set. The sample data is fitted online using the least squares method to establish a dynamic mapping relationship between vehicle speed and steering angle and the optimal correction coefficient, and to obtain the online learning correction amount of the corresponding correction coefficient. The above process is not the main technical content discussed in this invention, and can be obtained through existing technology, so it will not be elaborated here.

[0048] In this embodiment, this step enables the system to have memory and iterative optimization capabilities, allowing it to autonomously optimize parameters from historical operating conditions, adapt to characteristic deviations caused by vehicle aging, tire wear, and load changes, and improve long-term stability.

[0049] In this invention, an incremental PID controller is used to compensate for the correction lag caused by the lag in vehicle dynamic response. In step S203, the attitude prediction error obtained in S201 is used as the input of the proportional and integral components of the incremental PID controller, the tilt angle change rate is used as the differential term input, and the vehicle response lag time is used as the feedforward adjustment. The PID proportional, integral and differential parameters are adaptively adjusted according to the real-time estimated vehicle response lag time, and the corresponding output vehicle speed coefficient PID lag compensation correction amount and steering coefficient PID lag compensation correction amount are output.

[0050] In this embodiment, step S203 is used to offset the response delay caused by the mechanical and hydraulic systems, so that the change of the correction coefficient is synchronized with the vehicle dynamics in real time, avoiding inaccurate prediction and delayed correction due to response delay, ensuring the real-time performance and effectiveness of active rollover prevention control, and ensuring that the correction action is synchronized with the dynamic changes of the vehicle.

[0051] The initial correction coefficient, the online learning correction amount, and the PID lag compensation correction amount are combined to obtain the final vehicle speed correction coefficient. and steering angle correction final coefficient :

[0052] )

[0053]

[0054] in, The initial coefficient for vehicle speed correction is obtained from actual vehicle calibration and has a value range of 0 to 1. ; The initial coefficient for steering angle correction is obtained from actual vehicle calibration and has a value range of 0 to 1. This is the online learning correction amount for the steering coefficient; This is the correction amount for PID lag compensation of the steering coefficient.

[0055] In this embodiment, all correction coefficients are constrained within the range of [0.2, 1.0], and the change in a single correction amount does not exceed ±0.05, ensuring a smooth and stable correction process.

[0056] In step S3 of this invention, the system obtains the corresponding initial safe tilt angle threshold based on the real-time load and fork lifting height of the forklift collected at present, through a preset forklift stability tilt angle mapping relationship model. The initial safe tilt angle threshold is the maximum safe tilt angle at which the forklift just does not tip over. The preset stability characteristic model is a forklift stability tilt angle mapping relationship model established in advance through forklift actual tipping test or multibody dynamics simulation.

[0057] In this embodiment, the specific implementation method is as follows: Under different load conditions and different fork lifting heights, the maximum safe tilt angle for the forklift to prevent tipping is determined in advance, and multiple sets of experimental data are fitted into a two-dimensional lookup table function or mathematical fitting formula.

[0058]

[0059] in, Indicates the initial safe tilt angle threshold A mapping function between real-time load and fork lifting height.

[0060] When the system is working, the independent safety controller obtains the initial safety tilt angle threshold that matches the current load value and the current fork lifting height value by looking up a table or substituting into a fitting formula, based on the real-time collected current load value and current fork lifting height value.

[0061] The threshold adaptively decreases as the load increases and the fork lifting height increases, and adaptively increases as the load decreases and the fork lifting height decreases, thus achieving dynamic matching between the threshold and real-time operating conditions.

[0062] In this invention, a final coefficient for vehicle speed correction is introduced. and steering angle correction final coefficient The initial safe tilt angle threshold is dynamically corrected through multi-dimensional coupling, ultimately yielding a uniquely matching and real-time adaptive final dynamic safe tilt angle threshold under the current operating conditions. When a sensor malfunctions or the system malfunctions, the adaptive correction process is automatically frozen, and the system switches to offline calibration of the baseline coefficients to ensure system failure safety.

[0063] Final dynamic safety tilt threshold Determined according to the following formula:

[0064]

[0065] in, This is the initial safe tilt angle threshold; The final coefficient is adjusted for vehicle speed; The final coefficient is used to correct the steering angle.

[0066] The final dynamic safety tilt angle threshold, which is obtained after real-time adaptive changes in load, lifting height, speed, and steering angle of the forklift forks, is used for subsequent risk fusion assessment.

[0067] In this invention, in step S3, a comprehensive risk score is calculated. Based on the current vehicle lateral camber angle and the final dynamic safety camber angle threshold The formula is as follows:

[0068]

[0069] in, This is the final dynamic safety tilt angle threshold.

[0070] In this invention, step S4 categorizes risk levels into Level 1 Early Warning, Level 2 Alert, and Level 3 Emergency; the rules for assessing risk levels and the corresponding intervention strategies are as follows:

[0071] If the risk score S is less than or equal to the first threshold, the risk level is assessed as safe and no intervention action is taken.

[0072] If the risk score S is greater than the first threshold and less than or equal to the second threshold, the risk level is assessed as a Level 1 warning, triggering the audible and visual alarm.

[0073] If the risk score S is greater than the second threshold and less than or equal to the third threshold, the risk level is assessed as Level 2 warning, the driving power is limited to the preset upper limit, and the forks are controlled to descend at the first preset speed.

[0074] If the risk score S is greater than the third threshold, the risk level is assessed as Level 3 Emergency, triggering active braking and simultaneously controlling the forks to descend at a second preset speed. The second preset speed is greater than the first preset speed.

[0075] In this embodiment, the first threshold, the second threshold, and the third threshold can be set to 70, 80, and 90, respectively.

[0076] This invention achieves forward-looking and accurate risk perception through multi-source signal fusion and dynamic models; it realizes the leap from measuring the state to predicting the risk; the system maps continuous comprehensive risk scores into discrete risk levels, each level corresponding to a set of preset handling actions; it actively consumes the kinetic and potential energy of the vehicle and cargo, and stabilizes the vehicle body to the maximum extent before or during the rollover, preventing cargo from falling and the vehicle from sliding, thereby directly cutting off the secondary injury chain.

[0077] An active rollover prevention control system for forklifts based on multi-signal fusion includes:

[0078] The perception layer includes an IMU for measuring the lateral tilt angle of the forklift body, a load sensor for acquiring the real-time load of the forklift, a speed sensor for acquiring the real-time speed of the forklift, a displacement sensor for acquiring the lifting height of the forklift forks, and a steering angle sensor for acquiring the steering angle of the forklift.

[0079] Control layer: Runs a risk fusion algorithm, including: calculating the initial safe tilt angle threshold for the vehicle's current state based on the real-time fork load and fork lifting height; and calculating the initial safe tilt angle threshold based on the current vehicle speed. The initial safe tilt angle threshold is dynamically corrected based on the steering angle to obtain the final dynamic safe tilt angle threshold; the risk assessment level is then obtained based on the final dynamic safe tilt angle threshold.

[0080] The execution layer includes a motor driver for limiting the travel power to a preset upper limit, an electro-hydraulic valve for controlling the lifting of the forks, a braking module for active braking, and an audible and visual alarm for audible and visual alerts.

[0081] This invention ensures the reliability of decision-making and triggering channels through an independent hardware architecture, designs a physically independent and independently controlled safety controller, and specifically implements risk fusion algorithms and safety logic; even if the vehicle's main controller fails due to collision or interference, the safety system can still function normally; through a graded safety response sequence, it proactively intervenes in the accident chain to eliminate secondary damage.

[0082] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A forklift active anti-rollover control method based on multi-signal fusion, characterized in that, Includes the following steps: S1: Real-time acquisition of multi-source signals from the forklift, including vehicle lateral tilt angle, real-time load, fork lifting height, vehicle speed, and steering angle; S2: Obtain the first set of correction parameters based on the forklift's speed and steering angle, and dynamically correct the initial speed correction coefficient and initial steering angle correction coefficient obtained from the actual vehicle calibration based on the first set of correction parameters to obtain the final speed correction coefficient and final steering angle correction coefficient. S3: Obtain the initial safe tilt angle threshold based on the real-time load and fork lifting height. Combine the current vehicle speed and current steering angle, and correct the initial safe tilt angle threshold using the final correction coefficients for vehicle speed and steering angle to obtain the final dynamic safe tilt angle threshold under the current working conditions. S4: A comprehensive risk score is obtained by performing risk fusion calculation based on the current vehicle body lateral tilt angle and the final dynamic safety tilt angle threshold; Risk levels are assessed based on comprehensive risk scores, and corresponding dynamic intervention actions are implemented according to the different levels.

2. The forklift active anti-tipping control method according to claim 1, characterized in that: The process of obtaining the final coefficients of vehicle speed correction and steering angle correction is as follows: the first set of correction parameters is obtained through amplitude-limited PID and online learning. The initial coefficients of vehicle speed correction and steering angle correction are corrected using the obtained first set of correction parameters to obtain the final coefficients of vehicle speed correction and steering angle correction. The first set of correction parameters includes the online learning correction amount for vehicle speed coefficient, the PID lag compensation correction amount for vehicle speed coefficient, the online learning correction amount for steering coefficient, and the PID lag compensation correction amount for steering coefficient.

3. The forklift active anti-tipping control method according to claim 1, characterized in that: The steps for obtaining the first set of corrected parameters include: S201: Based on the collected data of the forklift's body lateral tilt angle, speed, and steering angle, a real-time state observer is constructed to obtain a set of real-time state parameters for adaptively correcting the initial coefficients of the speed correction and steering angle correction; the set of real-time state parameters includes the body lateral tilt angle change rate, attitude prediction error, and vehicle response lag time. S202: Based on the real-time state parameter set, establish dynamic mapping relationships between vehicle speed and steering angle and optimal correction coefficients respectively, and output the online learning correction amount of the correction coefficients, including the online learning correction amount of vehicle speed coefficient and the online learning correction amount of steering coefficient; S203: Using the attitude prediction error and the rate of change of the vehicle body lateral tilt angle obtained from S201 as inputs, and the vehicle response lag time as the feedforward adjustment amount, an incremental PID compensation controller is constructed. The PID parameters are adaptively adjusted according to the lag time and the corresponding PID lag compensation correction amount is output, including the PID lag compensation correction amount of the vehicle speed coefficient and the PID lag compensation correction amount of the steering coefficient, and finally the first set of correction parameters is obtained.

4. The forklift active anti-tipping control method according to claim 3, characterized in that: In step S201, the first derivative of the vehicle body lateral tilt angle is used to obtain the rate of change of the vehicle body lateral tilt angle. The difference between the real-time vehicle body lateral tilt angle and the initial safe tilt angle threshold is used to obtain the attitude prediction error. The vehicle response lag time is calculated by the time phase difference between the steering angle input and the tilt angle response.

5. The forklift active anti-tipping control method according to claim 3, characterized in that: In step S202, a sample library of historical operating conditions and correction coefficients is constructed using a ring buffer. An effective sample library is obtained by filtering through a real-time state parameter set. The sample data is then fitted online using the least squares method to establish a dynamic mapping relationship between vehicle speed and steering angle and the optimal correction coefficient, thereby obtaining the online learning correction amount of the corresponding correction coefficient.

6. The forklift active anti-tipping control method according to claim 3, characterized in that: In step S203, the attitude prediction error obtained in S201 is used as the input to the proportional and integral components of the incremental PID controller, the tilt angle change rate is used as the differential term input, and the vehicle response lag time is used as the feedforward adjustment. The PID proportional, integral, and differential parameters are adaptively adjusted according to the real-time estimated vehicle response lag time, and the corresponding outputs are the PID lag compensation correction amount for vehicle speed coefficient and the PID lag compensation correction amount for steering coefficient.

7. The forklift active anti-tipping control method according to claim 1, characterized in that: In step S3, based on the currently collected real-time load and fork lifting height, the corresponding initial safe tilt angle threshold is obtained through the established forklift stability tilt angle mapping model.

8. The forklift active anti-tipping control method according to claim 1, characterized in that: The aforementioned S4 is based on the current vehicle body lateral tilt angle and the final dynamic safety tilt angle threshold to obtain a comprehensive risk score.

9. The forklift active anti-tipping control method according to claim 1, characterized in that: In step S4, risk levels are assessed and corresponding dynamic intervention actions are performed according to their classification, including: risk levels are divided into Level 1 warning, Level 2 alert, and Level 3 emergency. If the risk score is less than or equal to the first threshold, the risk level is assessed as safe and no intervention action is taken. If the risk score is greater than the first threshold and less than or equal to the second threshold, the risk level is assessed as a Level 1 warning, triggering the audible and visual alarm. If the risk score is greater than the second threshold and less than or equal to the third threshold, the risk level is assessed as Level 2 warning, the driving power is limited to the preset upper limit, and the forks are controlled to descend at the first preset speed. If the risk score exceeds the third threshold, the risk level is assessed as Level 3 Emergency, triggering active braking and simultaneously controlling the forks to descend at a second preset speed. The second preset speed is greater than the first preset speed.

10. A forklift active anti-rollover control system based on multi-signal fusion, applicable to the forklift active anti-rollover control method according to any one of claims 1-9, characterized in that, include: The perception layer includes an IMU for measuring the lateral tilt angle of the vehicle body, a load sensor for acquiring the real-time load of the forklift, a speed sensor for acquiring the real-time speed of the forklift, a displacement sensor for acquiring the lifting height of the forklift forks, and a steering angle sensor for acquiring the steering angle of the forklift. The control layer runs a risk fusion algorithm, including: calculating the initial safe tilt angle threshold based on the real-time load of the forks and the lifting height of the forks in the current state of the vehicle; dynamically correcting the initial safe tilt angle threshold according to the current vehicle speed and steering angle to obtain the final dynamic safe tilt angle threshold; and obtaining the risk assessment level based on the final dynamic safe tilt angle threshold. The execution layer includes a motor driver for limiting the travel power to a preset upper limit, an electro-hydraulic valve for controlling the lifting of the forks, a braking module for active braking, and an audible and visual alarm for audible and visual alerts.

Images