Robot chassis stable running control system and control method thereof

Abstract

The invention discloses a robot chassis stable running control system and a control method thereof, and relates to the technical field of intelligent robot control, the system collects inertial attitude data and operation data through an MEMS accelerometer, a gyroscope and a magnetic encoder, and the inertial attitude data and the operation data are processed to form an inertial data set and a running data set; dividing the global path into a plurality of path segments j, executing stability scoring analysis on each path segment, and judging ground disturbance; when the path has disturbance, attitude coupling analysis is carried out on the attitude and speed of the robot in the driving process according to the driving data set, attitude compensation correction is carried out on the PWM signal according to a coupling state judgment result, and the compensation correction amount delta PWM is calculated; and carrying out secondary analysis on the compensated attitude data, calculating a comprehensive stability index Sta, and carrying out driving stability evaluation. According to the system, through multi-source coupling analysis of an inertia attitude and a driving state, dynamic stable driving control of a robot chassis in a complex ground environment is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent robot control, in particular to a robot chassis stable driving control system and a control method thereof. BACKGROUND

[0002] With the popularity of mobile robots in intelligent warehousing, inspection services, logistics transportation and other scenarios, the stability of the chassis driving of the robot becomes a key factor affecting the operation efficiency and the service life of the equipment. When the robot chassis drives in a complex ground environment, it will be affected by many factors such as ground roughness, slope, friction coefficient change and mechanical structure self-vibration, resulting in problems such as attitude fluctuation, excessive vibration or driving deviation. The traditional chassis control system relies on single inertial sensor or wheel speed feedback for control, and lacks global perception ability for ground disturbance and attitude coupling. Therefore, it is of great significance to establish a stable driving control system with multi-source data fusion perception and dynamic attitude compensation for ensuring the stable operation of the robot chassis, prolonging the service life of the key components and maintaining the driving path accuracy.

[0003] The existing stable driving control scheme usually adopts PID closed-loop control or single-source inertial feedback algorithm to simply correct the acceleration or speed of the robot driving process. However, such control systems generally have two shortcomings: first, the inertial and speed information cannot be dynamically coupled and analyzed, making it difficult to judge the coordination between attitude change and speed fluctuation, resulting in a lag in compensation response under ground disturbance; second, the recognition ability of the ground state change is weak, relying only on acceleration change for judgment, which is easily disturbed by motor noise and tire slip, causing false correction. Especially in uneven ground or alternating areas with multiple friction coefficients, the correlation between the attitude change of the robot chassis and the driving current is not easy to analyze in real time, so that the system is not easy to make accurate attitude correction and energy output control. SUMMARY

[0004] In view of the deficiencies of the prior art, the present application provides a robot chassis stable driving control system and a control method thereof, which solves the problems in the background art.

[0005] To achieve the above purpose, the present application realizes the following technical scheme: a robot chassis stable driving control system, comprising a multi-source perception module, a path stability analysis module, a coupling control module and a comprehensive stability feedback module; The multi-source perception module is used to collect the inertial attitude data and running data of the robot, and transmit them to the driving control system for data processing to obtain the inertial data set and the driving data set; The path stability analysis module is configured to divide the global path into a plurality of path segments j, analyze the path stability of each path segment j according to the inertial data set, and generate a path stability evaluation according to the analysis result, and trigger the coupling control module when the path is disturbed; The coupling control module is configured to analyze the posture and speed of the robot during travel according to the travel data set, generate a coupling state evaluation according to the analysis result, and perform posture compensation correction on the robot when the robot travel has an unstable trend; The comprehensive stability feedback module is configured to analyze the stability of the robot after the PWM output intensity is adjusted, and generate a travel stability evaluation according to the analysis result.

[0006] Preferably, the multi-source perception module includes an inertial posture perception unit, a motion state acquisition unit, and a data processing unit; The inertial posture perception unit is configured to output the inertial posture data of the robot in real time through the internal MEMS accelerometer and gyroscope of the robot, including the vertical acceleration component az and the pitch angular velocity ωy; The motion state acquisition unit is configured to acquire the wheel speed vi of the robot according to the magnetic encoder arranged at the end of each drive wheel shaft, and connect the current detection pin at the output end of the motor drive chip to the ADC channel of the master control chip to read the instantaneous current Ii in real time.

[0007] Preferably, the data processing unit is configured to transmit the inertial posture data and the running data to the travel control system through the SPI bus for data processing to obtain the inertial data set and the travel data set; The data processing is configured to preprocess the inertial posture data and the running data, and then perform ground roughness analysis and center speed analysis; The preprocessing includes timestamp alignment, denoising, missing value filling, and Kalman filtering; The timestamp alignment performs interpolation and prediction completion on missing time points of the inertial posture data and the running data through linear interpolation resampling technology, the denoising retains the transient structure characteristics of the inertial posture data and the running data through wavelet denoising technology, the missing value filling completes the processing of the inertial posture data and the running data with sampling interruption and communication packet loss through missing value interpolation technology, and the Kalman filtering fuses the accelerometer and gyroscope data through Kalman filtering algorithm to suppress the drift of the inertial posture data; The ground roughness analysis is configured to fit the periodic fluctuation of the acceleration az(t) with the trend of the motor current Ii(t), and calculate the ground roughness σs through the sliding window correlation analysis method, σs=corr(|Ii(t)- <ii>|, | az(t)- <az>|), where corr(·) is the Pearson correlation function, representing the degree of linear coupling between the current fluctuation amplitude and the acceleration fluctuation amplitude, <ii>and <az>respectively represent the mean of current and the mean of vertical acceleration within the sliding window; The center velocity analysis is used to average the wheel speed vi obtained by all the wheel encoders of the robot to obtain the instantaneous linear velocity Vce of the center of the robot chassis; The inertial data set includes a vertical acceleration component az and a ground roughness σs; The driving data set includes a pitch angular velocity ωy and an instantaneous linear velocity Vce.

[0008] Preferably, the path smoothness analysis module includes a path segmentation unit, a smoothness analysis unit and a path evaluation unit; The path segmentation unit is used to structurally segment the pre-driving path of the robot according to the environment map established by the SLAM technology, combine the path identification information, divide the global path into a plurality of path segments j, identify the starting point coordinates and the end point coordinates in space, and establish a set of sampling points {t1, t2, …, tN} in each path segment j at equal time steps. N}; The smoothness analysis unit is used to fit the inertial data set to analyze the path smoothness of each path segment j and construct a path smoothness score Spa j , which is used to analyze the smoothness degree of the path segment j during the driving of the robot chassis, quantify the comprehensive influence of the ground vibration and the mechanical response on the driving smoothness, and specifically: , wherein N represents the total number of time sampling points in the path segment j, az j (t) represents the vertical acceleration component at time t in the path segment j, σs(t) represents the ground roughness σs at time t in the current path segment, and a max represents the maximum vertical acceleration that the robot chassis can tolerate.

[0009] Preferably, the path evaluation unit is used to calculate the mean of the path smoothness score Spa of the stable historical path according to the statistical method, preset the mean as a path smoothness critical threshold Sy, and perform path smoothness evaluation on the real-time obtained path smoothness score Spa, and the specific evaluation scheme is as follows: When the path smoothness score Spa is less than the path smoothness critical threshold Sy, it indicates that the path is disturbed, and at this time the posture coupling analysis is triggered; When the path smoothness score Spa is greater than or equal to the path smoothness critical threshold Sy, it indicates that the current path is stable, and at this time the current program is maintained to continue driving, and the path state is continuously monitored.

[0010] Preferably, the coupling control module includes a coupling deviation analysis unit, a deviation judgment unit and a compensation execution unit; The coupling deviation analysis unit is configured to perform pose coupling analysis on the pose and speed of the ground robot during the driving of the disturbed path according to the driving data set when the path smoothness evaluation indicates that there is disturbance in the path, and construct a pose-speed coupling coefficient Kzv for measuring the coordination between the pose change rate and the linear speed of the robot chassis during driving, specifically: wherein Kzv(t) represents the pose-speed coupling coefficient at time t; The deviation judgment unit is configured to calculate the mean and standard deviation of the pose-speed coupling coefficient Kzv of the historical robot pose fluctuation and speed matching according to the statistical method, and preset the sum of the mean and standard deviation as a pose-speed coordination threshold Yk, and then perform coupling state evaluation with the real-time obtained pose-speed coupling coefficient Kzv, and the specific evaluation scheme is as follows: When the pose-speed coupling coefficient Kzv is less than or equal to the pose-speed coordination threshold Yk, it indicates that the robot pose fluctuation and speed matching, and the robot driving is stable, at this time the current differential configuration and running program are maintained; When the pose-speed coupling coefficient Kzv is greater than the pose-speed coordination threshold Yk, it indicates that the robot pose fluctuation and speed do not match, and the robot driving has an unstable trend, at this time a compensation correction control instruction is triggered.

[0011] Preferably, the compensation execution unit is configured to, when the robot driving has an unstable trend, control the chip to fit the real-time path smoothness score Spa(t) and the pose-speed coupling coefficient Kzv of the path segment, adjust the PWM output intensity of the robot to perform pose compensation correction on the robot, and calculate a compensation correction amount ΔPWM, which represents the correction amplitude of the control system to the standard pulse width modulation PWM0 signal, specifically: wherein PWM0 represents the standard PWM duty cycle when maintaining constant speed driving on a stable and flat ground, Yk represents the pose-speed coordination threshold, and Spa(t) represents the path smoothness score at time t.

[0012] Preferably, the comprehensive smoothness feedback module comprises a comprehensive smoothness analysis unit and a judgment feedback unit. The comprehensive smoothness analysis unit is configured to record the adjusted inertial pose data and running data in real time after adjusting the PWM output intensity of the robot, and perform secondary analysis through the multi-source perception module, and extract the path smoothness score Spa(t) and the pose-speed coupling coefficient Kzv obtained by the secondary analysis for fitting, and perform smoothness analysis on the robot after adjusting the PWM output intensity, and calculate a comprehensive smoothness index Sta, specifically: .

[0013] Preferably, the determination feedback unit is used to extract the demarcation value of the stable driving state and the existence of the driving instability risk at each analysis in history, and calculate the mean value of all demarcation values by statistical method to preset the driving stability demarcation threshold Sy, and then evaluate the driving stability by the real-time obtained comprehensive stability index Sta, and the specific evaluation scheme is as follows: When the comprehensive stability index Sta is less than the driving stability demarcation threshold Sy, it indicates that the robot still exists the driving instability risk, and a secondary control trigger signal is immediately generated and output to the compensation execution unit for iterative compensation correction. When the comprehensive stability index Sta is greater than or equal to the driving stability demarcation threshold Sy, it indicates that the robot is in a stable driving state, and the current control strategy and driving parameters are maintained.

[0014] A robot chassis stable driving control method, comprising the following steps: S1, collecting the inertial attitude data and running data of the robot, and transmitting to the driving control system for data processing to obtain the inertial data set and the driving data set; S2, dividing the global path into a plurality of path segments j, and performing path stability analysis on each path segment j according to the inertial data set, and generating path stability evaluation according to the analysis result, and triggering S3 when the path exists disturbance; S3, performing attitude coupling analysis on the attitude and speed of the robot during driving according to the driving data set, and generating coupling state evaluation according to the analysis result, and performing attitude compensation correction on the robot when the robot driving exists instability trend; S4, performing stability analysis on the robot after adjusting the PWM output intensity, and generating driving stability evaluation according to the analysis result.

[0015] The present application provides a robot chassis stable driving control system and its control method. It has the following advantages: (1) The multi-source perception module of the system cooperates with MEMS accelerometer, gyroscope and magnetic encoder to collect inertial attitude data and running data, and obtains inertial data set and driving data set with time sequence continuity through data processing. This module enables the system to simultaneously monitor the vertical impact, pitch change and driving load fluctuation of the robot chassis, realizes dynamic perception and quantitative description of ground disturbance, and has higher resolution capability in sampling accuracy and response real-time of chassis dynamic information compared with the traditional detection method relying on single inertial sensor, which lays an accurate data foundation for subsequent joint analysis of path and attitude.

[0016] (2) The system path smoothness analysis module is based on the path map constructed by the SLAM technology, divides the global path into multiple path segments, fits the inertial data set for each path segment j, analyzes the path smoothness of each path segment, constructs the path smoothness score Spa, and performs time resolution analysis on the ground disturbance features by the sliding window correlation method. The module can determine whether there is a structural disturbance or ground fluctuation in the path segment during the robot driving process, determine the path smoothness critical threshold Sy by the statistical method, evaluate the path smoothness, and realize the quantitative comparison and trigger control of the path segment stability. Compared with the traditional fixed path modeling method, the invention introduces a dynamic smoothness scoring mechanism at the path segmentation level, which can distinguish the small disturbance differences of different terrain areas, automatically identify non-structural ground changes, and provide timely and accurate trigger signals for subsequent attitude coupling analysis and control, solving the problems of response delay and path judgment ambiguity of the existing control system.

[0017] (3) The system coupling control module analyzes the attitude and speed of the robot driving on the disturbed path ground through the driving data set to calculate the attitude speed coupling coefficient Kzv when the path smoothness evaluation is disturbed, and sets the attitude speed coordination threshold Yk based on the statistical method. When the attitude fluctuation and speed are detected to be uncoordinated, the compensation execution unit is automatically triggered, the compensation correction amount ΔPWM is calculated according to the path smoothness score Spa(t) and the attitude speed coupling coefficient Kzv, and real-time dynamic compensation of the driving signal is realized. The comprehensive smoothness feedback module establishes a smoothness reevaluation system after PWM adjustment through the comprehensive smoothness analysis unit and the judgment feedback unit, calculates the comprehensive smoothness index Sta, and determines the driving stability threshold Sy based on the historical demarcation value. When the comprehensive smoothness index Sta is lower than the driving stability demarcation threshold Sy, a secondary control trigger signal is automatically generated, and iterative compensation correction is performed until the driving state tends to be stable. Through the progressive control mechanism, the invention realizes the whole process of adaptive adjustment from path disturbance recognition to attitude compensation to driving smoothness closed-loop feedback, which has continuous adjustment capability and multi-dimensional coupling response characteristics compared with the traditional single PID control system. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 A flowchart of the robot chassis smooth driving control system of the invention; Figure 2 A flowchart of the robot chassis smooth driving control method of the invention; Figure 3 A robot chassis smooth driving control system operation principle block diagram of the invention. DETAILED DESCRIPTION

[0019] With reference to the accompanying drawings: clearly and completely describe the technical solutions in the embodiments of the present application, obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor are within the scope of the present application. Embodiment 1

[0020] Please refer to Figure 1 The present application provides a robot chassis smooth driving control system, in order to achieve the above object, the present application is realized by the following technical scheme: including multi-source perception module, path stability analysis module, coupling control module and comprehensive stability feedback module; The multi-source perception module is used for collecting the inertial attitude data and running data of the robot, and transmitting to the driving control system for data processing to obtain the inertial data set and the driving data set; The path stability analysis module is used for dividing the global path into several path segments j, and performing path stability analysis on each path segment j according to the inertial data set, and generating path stability evaluation according to the analysis result, and triggering the coupling control module when the path is disturbed; The coupling control module is used for performing attitude coupling analysis on the attitude and speed of the robot during driving according to the driving data set, and generating coupling state evaluation according to the analysis result, and performing attitude compensation correction on the robot when the robot driving exists unstable trend; The comprehensive stability feedback module is used for performing stability analysis on the robot after adjusting the PWM output intensity, and generating driving stability evaluation according to the analysis result.

[0021] In this embodiment, the multi-source perception module realizes multi-dimensional data acquisition of the chassis running state, including inertial attitude data and running data, and obtains high-quality inertial data set and driving data set through data processing. The module enables the robot to reflect the coupling characteristics of ground disturbance, structural tilt and driving resistance in real time, solves the problem that the traditional single sensing method cannot simultaneously reflect the attitude and motion state, provides comprehensive and detailed dynamic input basis for subsequent analysis, and improves the response capability and perception accuracy of the system to complex road conditions from the data level. The path stability analysis module takes the inertial data set as the core, models the robot driving path in a structured segmented manner, divides the global path into a plurality of path segments j, and analyzes the path stability of each path segment j through the inertial data set to realize real-time identification of ground undulation and structural change. When there is an abnormal disturbance in the path segment, the coupling control module analyzes the attitude and speed of the robot during driving according to the driving data set, generates a coupling state evaluation according to the analysis result, and establishes a dynamic linkage mechanism from ground disturbance to attitude response. The segmented and self-triggering design enables the system to distinguish the ground stability of different regions and avoid affecting the overall control judgment due to an abnormality in a single interval. Compared with the traditional path evaluation method relying on fixed threshold and experience model, the system has higher spatial resolution and adaptability, and completes the quantitative stability of the path and dynamic warning tasks. The coupling control module analyzes the attitude and speed of the ground robot during driving according to the driving data set when the path stability evaluation indicates that there is a disturbance in the path, determines the stable trend of the robot in real time, and calculates the PWM correction amount to realize attitude compensation when an unstable trend is detected; the comprehensive stability feedback module reanalyzes the stable state of the robot after compensation, calculates the comprehensive stability index Sta, verifies the compensation effect through the comprehensive stability index Sta, realizes dynamic correction and self-calibration. The progressive control structure overcomes the problems of control lag, attitude drift and secondary oscillation in the prior art, enabling the robot to maintain continuous stable driving state on uneven or strongly disturbed ground. The scheme forms a multi-layer adaptive mechanism driven by data in the perception layer, analysis layer and control layer, achieves the technical goal of high chassis driving stability, strong path passing ability and smooth attitude response of the robot, and provides a bottom control solution for autonomous mobile robots in complex environments. Embodiment 2

[0022] For reference Figure 3 , specifically: the multi-source perception module includes an inertial attitude perception unit, a motion state acquisition unit and a data processing unit; The inertial attitude perception unit is used to output the inertial attitude data of the robot in real time through the internal MEMS accelerometer and gyroscope of the robot, including the vertical acceleration component az and the pitch angular velocity ωy; The vertical acceleration component az represents the dynamic impact in the vertical direction of the robot chassis; The pitch angular velocity ωy reflects the instantaneous change trend of the robot chassis in the front-back tilting amplitude; The motion state acquisition unit is used to acquire the running data of the robot, the wheel speed vi of the robot is obtained according to the magnetic encoder arranged at the end of each drive wheel shaft, and the current detection pin at the output end of the motor drive chip is connected to the ADC channel of the main control chip, so as to read the instantaneous current Ii in real time, and reflect the load fluctuation of the motor under different ground friction conditions.

[0023] The data processing unit is used to transmit the inertial attitude data and the running data to the driving control system through the SPI bus for data processing, so as to obtain the inertial data set and the driving data set; The data processing is used to pre-process the inertial attitude data and the running data, and then perform ground roughness analysis and center speed analysis; The preprocessing includes timestamp alignment, denoising, missing value filling and Kalman filtering; The timestamp alignment performs interpolation and prediction filling on the missing time points of the inertial attitude data and the running data through the linear interpolation resampling technology, the denoising retains the transient structure characteristics of the inertial attitude data and the running data through the wavelet denoising technology, the missing value filling completes the processing of the inertial attitude data and the running data with sampling interruption and communication packet loss through the missing value interpolation technology, and the Kalman filtering fuses the accelerometer and gyroscope data through the Kalman filtering algorithm to suppress the drift of the inertial attitude data; The ground roughness analysis is used to fit the periodic fluctuation of the acceleration az(t) and the change trend of the motor current Ii(t), and the ground roughness σs is calculated through the sliding window correlation analysis method, σs=corr(|Ii(t)-Ii(t-1)|,|az(t)-az(t-1)|), wherein corr represents the correlation coefficient. <ii>|, | az(t)- <az>|), where corr(·) is the Pearson correlation function, representing the degree of linear coupling between the current fluctuation amplitude and the acceleration fluctuation amplitude, for measuring the linear correlation between the current and the vertical acceleration in a statistical sense, <ii>and <az>respectively represent the current mean and the vertical acceleration mean within the sliding window; The center velocity analysis is used to average the wheel speed vi obtained by all wheel encoders of the robot to obtain the instantaneous linear velocity Vce of the robot chassis center; The inertial data set includes the vertical acceleration component az and the ground roughness σs; The running data set includes the pitch angular velocity ωy and the instantaneous linear velocity Vce.

[0024] In this embodiment, the inertial attitude sensing unit utilizes the real-time output of the MEMS accelerometer and the gyroscope to obtain the vertical acceleration component az and the pitch angular velocity ωy, realizing the synchronous monitoring of the chassis vertical impact and the attitude tilt change; the motion state acquisition unit obtains the wheel speed vi and the instantaneous current Ii through the driven wheel end magnetic encoder and the current detection channel, reflecting the dynamic response characteristics of the robot under different ground friction conditions; the data processing unit integrates the inertial attitude data and the running data through the SPI bus, and realizes data consistency and noise suppression by using timestamp alignment, wavelet denoising, interpolation repair technology and Kalman filtering algorithm, and the Kalman filtering algorithm is typically implemented as: filtered angle =0.98*(prev angle +gyro*dt)+0.02*accel angle , so that the attitude angle error is controlled within ±0.5°, prev angle represents the angle estimation value at the last time, gyro represents the angular velocity output by the gyroscope, dt represents the sampling time interval, accel angle represents the attitude angle estimated by the accelerometer according to the direction of gravity, 0.98*(prev angle +gyro*dt) is the weight given to the predicted angle, which is biased to believe the short-term change provided by the gyroscope, indicating the short-term credibility of the robot relying on the gyroscope, 0.02*accel angle is the correction angle, providing a low-frequency, long-term reference, which is used to correct the drift trend of the gyroscope, and filtered angle For the final fusion angle, it is 98% of the fast but possibly drifting gyroscope trend, 2% of the stable but noisy acceleration reference, and the two are fused to form a stable and continuous attitude estimation; on this basis, the ground roughness σs is calculated through sliding window correlation analysis, which describes the coupling relationship between acceleration and current fluctuation, and the chassis center line velocity Vce is obtained through wheel speed average calculation, forming a double-layer data structure of inertial data set and driving data set. This module realizes the multi-dimensional perception of the robot to the ground disturbance, attitude change and driving load, and also makes the data have the characteristics of quantifiable, traceable and fusible. Compared with the traditional system which relies on single inertial measurement or speed feedback, the identification sensitivity and response real-time of the robot to the micro disturbance under complex terrain are improved, which provides a solid data foundation for subsequent path smoothness analysis and attitude coupling control, and improves the stability, path passability and overall control accuracy of the robot. Embodiment 3

[0025] Please refer to Figure 3 , specifically: the path smoothness analysis module includes a path segmentation unit, a smoothness analysis unit and a path evaluation unit; The path segmentation unit is used to segment the robot pre-driving path according to the environment map established by the SLAM technology, combine the path identification information, divide the global path into several path segments j, identify the start point coordinates and end point coordinates in space, and establish a sampling point set {t1, t2, …, t N} in each path segment with equal time steps; The smoothness analysis unit is used to fit the inertial data set to analyze the path smoothness of each path segment j, and construct the path smoothness score Spa j , which is used to analyze the smoothness of the path segment j in the robot chassis driving process, quantify the comprehensive influence of ground vibration and mechanical response on driving smoothness, specifically: , wherein N represents the total number of time sampling points contained in the path segment j, az j (t) represents the vertical acceleration component at time t in the path segment j, σs(t) represents the ground roughness σs at time t in the current path segment, and a max represents the maximum vertical acceleration that the robot chassis can tolerate.

[0026] The path evaluation unit is used to calculate the mean value of the path smoothness score Spa of the historical path stability according to the statistical method, and preset the mean value as the path smoothness critical threshold Sy, and then evaluate the path smoothness with the real-time obtained path smoothness score Spa, the specific evaluation scheme is as follows: When the path smoothness score Spa is less than the path smoothness critical threshold Sy, it indicates that the path is disturbed, and the attitude coupling analysis is triggered at this time; When the path smoothness score Spa≥ path smoothness critical threshold Sy, it indicates that the current path is stable, at which the current program is maintained to continue driving, and the path state is continuously monitored.

[0027] In this embodiment, the path segmentation unit establishes an environment map using SLAM technology, and divides the global path into structured path segments in combination with path identification information. Each path segment is accurately identified by spatial coordinates and heading angle, and equal time step sampling points are set within the segment to realize dual quantifiable modeling of path and time. The smoothness analysis unit jointly calculates the vertical acceleration azj(t) and the ground roughness σs(t) by fitting the inertial data set to construct the path smoothness score Spa j , which quantifies the combined effect between ground vibration and mechanical response and dynamically reflects the ride smoothness characteristics of the path segment. The formula derivation is based on the classic vibration response analysis and normalized stability function; the vibration response basic formula F z (t)=m×az(t), where Fz(t) is the instantaneous reaction force of the ground on the chassis, the vertical acceleration az(t) is the direct response quantity of the ground input vibration when the chassis drives on uneven ground, and the road input and acceleration transmission relationship is az(t)=H(ω)×y′(t), where y′(t) represents the vertical acceleration of the road input, H(ω) is the system transfer function, which reflects the amplification or attenuation effect of ground disturbance on the chassis, and if H(ω) is approximated as a function related to ground roughness, H(ω)≈1+σs(t) can be taken; and the normalized stability function in the control system evaluation, the stability is often expressed by the error normalization , which represents the proportion of the system state deviating from the ideal state, where x(t) is the actual deviation, x max is the upper limit. Since the amplification relationship between the ground roughness and the vertical acceleration is a eff (t)=|az(t)|×(1+σs(t)), where a eff (t) is the equivalent vertical vibration response amplitude, which reflects the superimposed effect of ground undulation and inertial impact, and a eff (t) is substituted into the normalized stability function to obtain , the robot is sampled N times at equal time steps within the path segment j, the time average of the instantaneous smoothness function is obtained to reflect the overall trend, and the path smoothness score Spa j . In this formula, the vertical acceleration component az and the maximum vertical acceleration a max have the same dimension, and the dimension influence is cancelled by taking the ratio, and the ground roughness σs is a dimensionless parameter, so the path smoothness score Spa j The overall formula operation has no dimension influence for non-dimensional value. The path evaluation unit extracts the score average of the historical stable path based on the statistical method as a path stability critical threshold Sy, and compares it with the real-time path stability score Spa. When the path stability score Spa is lower than the path stability critical threshold Sy, the posture coupling analysis is automatically triggered, realizing the real-time identification and control closed-loop adaptive start of the path disturbance. Through this module, the system realizes the quantifiable, predictable and adjustable driving stability at the path level. Not only can it perceive the unstable area in advance under complex terrain, reduce the posture fluctuation and speed mutation, but also enables the robot to have dynamic path self-optimization ability based on the actual ground characteristics, thereby improving the overall driving stability and path passing accuracy. Embodiment 4

[0028] Please refer to Figure 3 Specifically, the coupling control module comprises a coupling deviation analysis unit, a deviation judgment unit and a compensation execution unit. The coupling deviation analysis unit is configured to, when the path stability evaluation indicates that there is a disturbance in the path, perform posture coupling analysis on the posture and speed of the ground robot during driving on the disturbed path according to the driving data set, and construct a posture-speed coupling coefficient Kzv for measuring the coordination between the posture change rate and linear speed of the robot chassis during driving, specifically as follows: Wherein, Kzv(t) represents the posture-speed coupling coefficient at time t. The deviation judgment unit is configured to calculate the mean and standard deviation of the posture-speed coupling coefficient Kzv of the historical robot posture fluctuation and speed matching by statistical method, and preset the sum of the mean and standard deviation as a posture-speed coordination threshold Yk, and then perform coupling state evaluation on the real-time acquired posture-speed coupling coefficient Kzv. The specific evaluation scheme is as follows: When the posture-speed coupling coefficient Kzv is less than or equal to the posture-speed coordination threshold Yk, it indicates that the robot posture fluctuation and speed matching, and the robot driving is stable. At this time, the current differential configuration and running program are maintained. When the posture-speed coupling coefficient Kzv is greater than the posture-speed coordination threshold Yk, it indicates that the robot posture fluctuation and speed do not match, and the robot driving has an unstable trend. At this time, the compensation correction control instruction is triggered.

[0029] The compensation execution unit is configured to, when the robot driving has an unstable trend, control the chip to fit the real-time path stability score Spa(t) and the posture-speed coupling coefficient Kzv of the path segment, adjust the PWM output intensity of the robot to perform posture compensation correction on the robot, and calculate the compensation correction amount ΔPWM, which represents the correction amplitude of the control system to the standard pulse width modulation PWM0 signal, specifically as follows: Wherein, PWM0 represents the standard PWM duty ratio when maintaining constant speed driving under steady state and flat ground, Yk represents the attitude speed coordination threshold, Spa(t) represents the path smoothness score at time t.

[0030] In this embodiment, the coupling deviation analysis unit extracts the time sequence relationship between pitch angle velocity and linear velocity based on the driving data set, calculates the attitude speed coupling coefficient Kzv, which is used to represent the synchronization between the attitude change and the speed response of the chassis. The angular velocity is defined as Where θy is the pitch angle, which is defined by the basic kinematics; the arc length velocity of the mass point is defined as Where s is the arc length coordinate of the chassis along the driving path, Vce is the instantaneous linear speed of the chassis center, which is derived from the standard definition of curve motion and arc length parameterization, and the "time change rate" is replaced by the "change rate along the path" through the chain rule, The arc length derivative of the yaw angle ψ is commonly used in road geometry Indicates the planar curvature, and the pitch angle θy is isomorphic, Understood as the "spatial gradient of the pitch attitude". According to this, the absolute value of the formula is taken to measure the "attitude change intensity per unit distance", and the attitude speed coupling coefficient Kzv is finally obtained. The deviation judgment unit establishes the attitude speed coordination threshold Yk according to the statistical method, superimposes the mean and standard deviation of the attitude speed coupling coefficient Kzv under the historical stable working conditions to form an adaptive reference boundary, and judges the real-time coupling state. When it is detected that the attitude speed coupling coefficient Kzv exceeds the attitude speed coordination threshold Yk, the system determines that the attitude and speed of the robot are unbalanced, the compensation execution unit controls the chip to fit the real-time path smoothness score Spa(t) and the attitude speed coupling coefficient Kzv of the path segment, adjusts the PWM output intensity of the robot to compensate and correct the attitude of the robot, and calculates the compensation correction amount ΔPWM to compensate the amplitude of the motor control signal, so that the driving output is re-matched with the ground disturbance characteristics and the attitude change rate, and the attitude swing and driving instability caused by the terrain undulation or friction mutation are suppressed. Through the progressive control mode of the above coordination analysis, threshold judgment and adaptive compensation, the system realizes the closed-loop dynamic adjustment of the attitude, speed and driving signal, and achieves the smooth driving control goal of the robot in complex ground environment. The compensation correction amount ΔPWM formula is derived from the closed-loop correction idea in control theory and mechanical system dynamics, and its core reasoning comes from the proportional regulation and attitude and speed coordination control model. In classical automatic control theory, the output intensity of the PWM signal can be regarded as the control amount of the actuator, and its amplitude is usually proportional to the error term e(t) Δu(t)=K p ×e(t), where K p As a proportional gain, the formula embodies the basic principle of correcting by proportion to offset the attitude or speed error; in the robot chassis driving control, the attitude deviation and speed change are not a single error source, but a dynamically coupled variable, therefore, the error term e(t) is extended to the attitude and speed coupling term Kzv(t), and the path stability correction factor Spa(t) is introduced, and the compound adjustment function is constructed on the basis of the proportional control idea, and the PWM output is dynamically adjusted according to the attitude and speed matching degree, so that the robot maintains stable operation under different terrains, and the original proportional control is expanded to ΔPWM=PWM0×f(Kzv(t),Spa(t)), in order to ensure the adaptive characteristics of the system, the linear decreasing function (1-Kzv(t) / Yk) is used to represent the coordination degree of attitude and speed, and the path stability parameter Spa(t) is combined in the form of (1-Spa(t)) to reflect the inhibition of the control signal by the ground disturbance, so as to construct the compensation correction amount ΔPWM formula. In the formula, the attitude speed coupling coefficient Kzv and the attitude speed coordination threshold Yk have the same dimension, the ratio offset dimension is eliminated, the real-time path stability score Spa(t) is a dimensionless value, the standard PWM duty ratio PWM0 is a dimensionless value, and therefore the compensation correction amount ΔPWM is a dimensionless value, and the formula does not have dimension influence. Compared with the traditional PID regulation mode based on fixed parameters, the scheme is significantly optimized in real-time responsiveness, disturbance suppression ability and stable recovery speed, so that the robot has higher path adaptability and driving smoothness, and provides a high-precision and adaptive implementation means for chassis-level dynamic stability control. Embodiment 5

[0031] Please refer to Figure 3 , specifically: the comprehensive stability feedback module comprises a comprehensive stability analysis unit and a judgment feedback unit; The comprehensive stability analysis unit is used to record the adjusted inertial attitude data and running data in real time after adjusting the PWM output intensity of the robot, and to perform secondary analysis through the multi-source perception module, and to extract the path stability score Spa(t) and the attitude speed coupling coefficient Kzv obtained by the secondary analysis for fitting, and to analyze the stability of the robot after adjusting the PWM output intensity, and to calculate the comprehensive stability index Sta, specifically: .

[0032] The judgment feedback unit is used to extract the stable driving state and the dividing value of the risk of unstable driving at each analysis in history, and to calculate the preset driving stability dividing threshold Sy by statistical method, and to perform driving stability evaluation with the real-time obtained comprehensive stability index Sta, and the specific evaluation scheme is as follows: When the comprehensive stability index Sta is less than the driving stability decomposition threshold Sy, it indicates that the robot still has the risk of driving instability, and a secondary regulation trigger signal is immediately generated and output to the compensation execution unit for iterative compensation correction; When the comprehensive stability index Sta is greater than or equal to the driving stability decomposition threshold Sy, it indicates that the robot is in a stable driving state, and the current control strategy and driving parameters are maintained.

[0033] In this embodiment, after the dynamic adjustment of the PWM output intensity is completed, the multi-source perception module is used to collect and analyze the adjusted inertial attitude data and running data, and the comprehensive stability index Sta is calculated based on the fitting results of the path stability score Spa(t) and the attitude speed coupling coefficient Kzv to quantify the overall stability of the robot. The formula is the core expression of the comprehensive stability index calculation model in the robot chassis driving stability evaluation system, which is evolved from the proportional correction model in control theory and the dynamics coordination function, and aims to quantify the overall stability after attitude compensation. In classical control theory, the stability of system output can be expressed in proportional form as S(t) = S0 x (1 - η), where S0 represents the stable output in the ideal state, and η is the disturbance term or correction factor, representing the stability decay of the system under external disturbance. In robot driving control, system stability is not only affected by ground disturbance, but also affected by the coupling effect of attitude and speed matching, so the disturbance term is expanded to a composite function between the attitude and speed coupling coefficient and the ground stability , and the comprehensive stability index Sta is finally obtained. In this formula, the attitude speed coupling coefficient Kzv and the attitude speed coordination threshold Yk have the same dimension, and the dimension influence is offset by ratio. The real-time path stability score Spa(t) is dimensionless, so the comprehensive stability index Sta is dimensionless, and the formula does not have dimension influence. The determination feedback unit calculates the driving stability decomposition threshold Sy based on the dividing value between stable and unstable states in the historical running record, and compares and evaluates the current comprehensive stability index Sta in real time: when the comprehensive stability index Sta is lower than the driving stability decomposition threshold Sy, a secondary regulation signal is automatically generated to trigger the compensation execution unit for iterative correction; when the comprehensive stability index Sta is higher than the driving stability decomposition threshold Sy, the current control strategy is maintained. This module realizes closed-loop optimization from one-time compensation to multiple dynamic self-correction at the control level, so that the robot can continuously maintain a stable driving state under different ground friction, load disturbance or attitude changes, reduces control lag and attitude overshoot, and improves the driving continuity, attitude smoothness and control accuracy of the robot in complex terrain. Embodiment 6

[0034] Please refer to Figure 2 A robot chassis stable driving control method, comprising the following steps: S1, collect the inertial attitude data and running data of the robot, and transmit to the driving control system for data processing to obtain the inertial data set and the driving data set; S2, divide the global path into several path segments j, and analyze the path stability of each path segment j according to the inertial data set, and generate a path stability evaluation according to the analysis result, and trigger S3 when there is disturbance in the path; S3, analyze the attitude and speed of the robot when driving according to the driving data set, and generate a coupling state evaluation according to the analysis result, and make attitude compensation correction to the robot when there is an unstable trend in the robot driving; S4, analyze the stability of the robot after adjusting the PWM output intensity, and generate a driving stability evaluation according to the analysis result.

[0035] Although the embodiments of the present application have been shown and described, it can be understood by those skilled in the art that various changes, modifications, replacements and variations can be made to the embodiments without departing from the principles and spirits of the present application, and the scope of the present application is defined by the appended claims and their equivalents.< / az> < / ii> < / az> < / ii> < / az> < / ii> < / az> < / ii>

Claims

1. A robot chassis smooth driving control system, characterized in that: It includes a multi-source sensing module, a path stability analysis module, a coupled control module, and a comprehensive stability feedback module; The multi-source sensing module is used to collect the robot's inertial attitude data and operation data, and transmit them to the driving control system for data processing to obtain inertial datasets and driving datasets. The path stability analysis module is used to divide the global path into several path segments j, perform path stability analysis on each path segment j based on the inertial dataset, generate a path stability assessment based on the analysis results, and trigger the coupling control module when there is a disturbance in the path. The coupling control module is used to perform attitude coupling analysis on the robot's attitude and speed during driving based on the driving dataset, and generate a coupling state assessment based on the analysis results. When the robot's driving shows an unstable trend, it performs attitude compensation and correction. The integrated stability feedback module is used to perform stability analysis on the robot after adjusting the PWM output intensity, and generate a driving stability assessment based on the analysis results.

2. The robot chassis smooth driving control system according to claim 1, characterized in that: The multi-source sensing module includes an inertial attitude sensing unit, a motion state acquisition unit, and a data processing unit; The inertial attitude sensing unit is used to output the robot's inertial attitude data in real time through the robot's internal MEMS accelerometer and gyroscope, including the vertical acceleration component az and the pitch angular velocity ωy. The motion state acquisition unit is used to collect the robot's running data. It obtains the robot's wheel speed vi based on the magnetic encoders installed at the ends of each drive wheel shaft, and connects the current detection pin of the motor drive chip output to the ADC channel of the main control chip to read the instantaneous current Ii in real time.

3. The robot chassis smooth driving control system according to claim 2, characterized in that: The data processing unit is used to transmit inertial attitude data and running data to the driving control system via the SPI bus for data processing, and to obtain inertial dataset and driving dataset. The data processing is used to preprocess the inertial attitude data and operational data before performing ground roughness analysis and center velocity analysis. The preprocessing includes timestamp alignment, noise reduction, missing value filling, and Kalman filtering; The timestamp alignment uses linear interpolation resampling technology to interpolate and predict missing time points in inertial attitude data and operational data; the denoising uses wavelet denoising technology to preserve the transient structural features of inertial attitude data and operational data; the missing value filling uses missing value interpolation technology to complete inertial attitude data and operational data affected by sampling interruptions and communication packet loss; and the Kalman filtering uses the Kalman filtering algorithm to fuse accelerometer and gyroscope data to suppress drift in inertial attitude data. The ground roughness analysis is used to fit the periodic fluctuations of acceleration az(t) with the changing trend of motor current Ii(t), and the ground roughness σs is calculated using the sliding window correlation analysis method, σs=corr(|Ii(t)- <ii>|,|the(t)- <az>|), where corr(·) is the Pearson correlation function, representing the degree of linear coupling between the current fluctuation amplitude and the acceleration fluctuation amplitude. <ii>and <az> These represent the average current and average vertical acceleration within the sliding window, respectively.< / az> < / ii> < / az> < / ii> The center velocity analysis is used to average the wheel velocities vi obtained by the encoders of all the wheels of the robot to obtain the instantaneous linear velocity Vce of the center of the robot chassis. The inertial dataset includes the vertical acceleration component az and the ground roughness σs; The driving dataset includes pitch angular velocity ωy and instantaneous linear velocity Vce.

4. The robot chassis smooth driving control system according to claim 3, characterized in that: The path stability analysis module includes a path segmentation unit, a stability analysis unit, and a path evaluation unit; The path segmentation unit is used to perform structured segmentation of the robot's pre-driving path based on the environmental map established by SLAM technology and the path identification information. The global path is divided into several path segments j, spatially identified by their start and end coordinates, and a set of sampling points {t1, t2, ..., t...} is established within each path segment at equal time steps. N }; The stationarity analysis unit is used to fit the inertial dataset to perform path stationarity analysis on each path segment j and construct a path stationarity score Spa. j This is used to analyze the stability of path segment j during robot chassis movement, quantifying the combined impact of ground vibration and mechanical response on ride comfort. Specifically: Where N represents the total number of time sampling points contained in path segment j, az j (t) represents the vertical acceleration component at time t in path segment j, σs(t) represents the ground roughness σs of the current path segment at time t, and a max This indicates the maximum vertical acceleration that the robot chassis can tolerate.

5. The robot chassis smooth driving control system according to claim 4, characterized in that: The path evaluation unit is used to calculate the mean of the path stability score Spa when the historical path is stable according to the statistical method, and preset the mean as the path stability critical threshold Sy, and then perform path stability evaluation with the real-time obtained path stability score Spa. The specific evaluation scheme is as follows. When the path stability score Spa is less than the path stability critical threshold Sy, it indicates that there is a disturbance in the path, and attitude coupling analysis is triggered. When the path stability score Spa is greater than or equal to the path stability critical threshold Sy, it indicates that the current path is stable. At this time, the current program continues to run and the path status is continuously monitored.

6. The robot chassis smooth driving control system according to claim 5, characterized in that: The coupling control module includes a coupling deviation analysis unit, a deviation judgment unit, and a compensation execution unit; The coupling deviation analysis unit is used to perform attitude coupling analysis on the attitude and velocity of the ground robot traveling on the disturbed path based on the driving dataset when the path stability assessment indicates that there is a disturbance in the path. It also constructs an attitude-velocity coupling coefficient Kzv to measure the coordination between the robot chassis's attitude change rate and linear velocity during travel. Specifically: , where Kzv(t) represents the attitude-velocity coupling coefficient at time t; The deviation judgment unit is used to calculate the mean and standard deviation of the attitude-velocity coupling coefficient Kzv when the historical robot attitude fluctuation and speed are matched according to the statistical method, and preset the sum of the mean and standard deviation as the attitude-velocity coordination threshold Yk, and then evaluate the coupling state with the real-time acquired attitude-velocity coupling coefficient Kzv. The specific evaluation scheme is as follows. When the attitude-velocity coupling coefficient Kzv ≤ attitude-velocity coordination threshold Yk, it means that the robot's attitude fluctuations are matched with its speed, and the robot's movement is stable. At this time, the current differential speed configuration and running program are maintained. When the attitude-velocity coupling coefficient Kzv > attitude-velocity coordination threshold Yk, it indicates that the robot's attitude fluctuations and speeds are mismatched, and the robot's movement tends to be unstable. At this time, a compensation and correction control command is triggered.

7. A robot chassis smooth driving control system according to claim 6, characterized in that: The compensation execution unit is used to perform attitude compensation correction on the robot when the robot's movement shows an unstable trend. The control chip fits the real-time path stability score Spa(t) of the path segment with the attitude-velocity coupling coefficient Kzv, adjusts the robot's PWM output intensity, and calculates the compensation correction amount ΔPWM, which represents the correction magnitude made by the control system to the standard pulse width modulation (PWM0) signal. Specifically: Where PWM0 represents the standard PWM duty cycle when maintaining a constant speed under steady-state and flat ground conditions, Yk represents the attitude-velocity coordination threshold, and Spa(t) represents the path stability score at time t.

8. A robot chassis smooth driving control system according to claim 7, characterized in that: The integrated stability feedback module includes an integrated stability analysis unit and a judgment feedback unit; The comprehensive stability analysis unit is used to record the adjusted inertial attitude data and running data in real time after adjusting the robot's PWM output intensity. It then performs secondary analysis through a multi-source sensing module, extracts the path stability score Spa(t) and attitude-velocity coupling coefficient Kzv obtained from the secondary analysis, fits them, performs stability analysis on the robot after PWM output intensity adjustment, and calculates the comprehensive stability index Sta. Specifically: .

9. A robot chassis smooth driving control system according to claim 8, characterized in that: The judgment feedback unit is used to extract the boundary value between the stable driving state and the risk of driving instability in each historical analysis, and calculate the mean of all boundary values ​​by statistical method to set a preset driving stability boundary threshold Sy, and then compare it with the real-time acquired comprehensive stability index Sta to evaluate driving stability. The specific evaluation scheme is as follows. When the overall stability index Sta < driving stability decomposition threshold Sy, it indicates that the robot still has the risk of driving instability. A secondary control trigger signal is immediately generated and output to the compensation execution unit for iterative compensation and correction. When the overall stability index Sta is greater than or equal to the driving stability decomposition threshold Sy, it indicates that the robot is in a stable driving state and maintains the current control strategy and driving parameters.

10. A method for controlling the smooth driving of a robot chassis, applied to the robot chassis smooth driving control system according to any one of claims 1-9, characterized in that: Includes the following steps: S1. Collect the robot's inertial attitude data and operation data, and transmit them to the driving control system for data processing to obtain inertial datasets and driving datasets; S2. Divide the global path into several path segments j, and perform path stability analysis on each path segment j based on the inertial dataset. Generate a path stability assessment based on the analysis results. Trigger S3 when there is a disturbance in the path. S3. Perform attitude coupling analysis on the robot's attitude and speed during driving based on the driving dataset, and generate a coupling state assessment based on the analysis results. Perform attitude compensation and correction on the robot when there is an unstable trend in the robot's driving. S4. Perform a stability analysis on the robot after adjusting the PWM output intensity, and generate a driving stability assessment based on the analysis results.

Images