Sightseeing vehicle smoothness optimization method and system based on big data

Abstract

The invention relates to the technical field of smoothness optimization, in particular to a sightseeing vehicle smoothness optimization method and system based on big data, and the method comprises the steps: multi-dimensional data collection: obtaining sightseeing vehicle vibration acceleration data, front road surface image data and passenger heart rate variability data; performing multi-source data fusion processing: performing frequency domain feature extraction on the vibration acceleration data to obtain a vibration energy distribution spectrum, performing texture segmentation and pit recognition on the road surface image data to generate a road surface roughness matrix, and performing standardization processing on the passenger heart rate variability data to generate a somatosensory discomfort index; dynamic weight smoothness evaluation: inputting a preset fuzzy logic model, and outputting a comprehensive smoothness index; multi-objective optimization strategy generation: generating a multi-parameter collaborative adjustment instruction set; and closed-loop feedback dynamic optimization: carrying out correlation verification, and correcting a weight distribution rule of the fuzzy logic model. The comfort and stability of the sightseeing vehicle in the running process are improved, and the intelligence of the control system is enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of ride comfort optimization, and in particular to a method and system for optimizing the ride comfort of a sightseeing vehicle based on big data. Background Art

[0002] With the continuous advancement of the integration of culture and tourism and the construction of smart scenic spots, sightseeing buses, as an important tool for carrying tourists for short-distance travel, have put forward higher requirements for smoothness and comfort during the ride. The smoothness control of traditional sightseeing buses mainly relies on suspension structure design, path detour strategy or manual experience adjustment. It usually only responds passively based on the acceleration data of the vehicle itself, lacks the multi-dimensional perception ability of the driving environment, road conditions and passengers' subjective feelings, and it is difficult to achieve personalized and intelligent smoothness management; in addition, most existing systems lack feedback mechanisms and cannot dynamically optimize the control strategies according to the passengers' immediate physiological or behavioral reactions, resulting in unstable comfort performance of the vehicle in complex road conditions or variable passenger flow scenarios.

[0003] In recent years, with the gradual application of technologies such as big data perception, fuzzy control, and machine learning in the field of intelligent travel, researchers have begun to try to introduce multi-source heterogeneous data for vehicle comfort modeling, such as using cameras to identify potholes on the road, monitoring passengers' heart rate changes through wearable devices, and combining vibration signals for road condition and response analysis. However, existing methods generally have problems such as insufficient data fusion dimensions, fixed parameter weights, and lagging feedback mechanisms, making it difficult to achieve global optimization of the smoothness control of sightseeing vehicles. Summary of the Invention

[0004] The present invention provides a method and system for optimizing the ride comfort of sightseeing vehicles based on big data. This method integrates a big data-driven intelligent optimization method that integrates vehicle dynamic response, environmental image recognition, and passenger physiological state perception, and constructs a closed-loop adaptive ride comfort control system to improve the overall operating quality of intelligent sightseeing vehicles and the user's riding experience.

[0005] A method for optimizing the ride comfort of a sightseeing vehicle based on big data comprises the following steps: S1, multi-dimensional data collection: obtain the sightseeing car vibration acceleration data through the on-board sensor, the front road image data through the camera, and the passenger heart rate variability data through the wearable device; S2, multi-source data fusion processing: frequency domain feature extraction of vibration acceleration data to obtain the vibration energy distribution spectrum, texture segmentation and pothole identification of road surface image data to generate the road surface roughness matrix, and normalization of passenger heart rate variability data to generate the somatic discomfort index; S3, Dynamic Weighted Ride Comfort Evaluation: The vibration energy distribution spectrum, road roughness matrix, and somatosensory discomfort index are input into a preset fuzzy logic model. The weights of the parameters are dynamically assigned based on the current driving scenario, and a comprehensive ride comfort index is output. S4, multi-objective optimization strategy generation: Based on the comprehensive ride comfort index, the suspension damping coefficient, drive motor torque distribution coefficient, and path planning priority parameters are simultaneously optimized through a genetic algorithm to generate a multi-parameter coordinated adjustment instruction set; S5, closed-loop feedback dynamic optimization: The real-time collected passenger facial expression data is correlated with the somatosensory discomfort index, and the weight allocation rules of the fuzzy logic model are modified based on the verification results.

[0006] Optionally, the S1 includes: S11, collects vehicle vibration acceleration data in real time through a three-axis acceleration sensor installed at the center of the front and rear axles of the vehicle body; S12 uses a wide-angle camera installed at the front of the vehicle to continuously collect road image frames within a 5-meter range in front of the vehicle and uses image texture features and edge gradient change rate to identify rough road areas; S13 collects heart rate variability data through a photoplethysmography device worn on the passenger's wrist, using time domain indicators Calculate the somatosensory stimulation level.

[0007] Optionally, the S2 includes: S21, performing a fast Fourier transform on the vibration acceleration data to extract frequency domain features to obtain a vibration energy distribution spectrum; S22, performing texture segmentation and pothole recognition on the collected sequence of images of the road ahead to generate a road roughness matrix, wherein the roughness matrix is ​​constructed based on image gradients and local binary pattern texture features; S23, the heart rate variability index Perform standardization to generate a physical discomfort index .

[0008] Optionally, the S3 includes: S31, inputting the vibration energy distribution spectrum, road roughness matrix, and somatosensory discomfort index into input terminals of a fuzzy logic model respectively to construct a ternary input variable set; S32, dynamically assigning weights to the input variables based on the current driving scene information using the fuzzy logic rule base to form a normalized weight coefficient set; S33, calculates and outputs a comprehensive smoothness index based on the assigned weight coefficient and the ternary input variables.

[0009] Optionally, the S4 includes: S41, Optimization variable modeling: Based on comprehensive smoothness index , constructed with the suspension damping coefficient , motor torque distribution coefficient and path planning priority parameters Modeling multi-objective optimization of optimization variables; S42, fitness function construction: Based on multi-objective optimization modeling, a fitness function is defined as the optimization basis for measuring the quality of the ride comfort of each parameter combination; S43, Genetic Algorithm Iterative Optimization: Initialize the population and perform iterative search through selection, crossover and mutation operations to finally output the optimal parameter combination , generating a multi-parameter collaborative adjustment instruction set.

[0010] Optionally, the step S43 uses a genetic algorithm to perform global search optimization on the parameter vector, and the iterative process includes: S431, population initialization: set the population size to , randomly generated initial individuals, each individual is a set of control parameter combinations to be optimized ; S432, fitness evaluation: input individual parameters into the comprehensive ride comfort model and calculate the fitness value , used to measure the smooth performance of the current combination; S433, selection operation: based on the fitness level, prioritize the best individuals from the current population as the next generation parent; S434, crossover operation: Generate new individuals between parents through the crossover algorithm, expand the search space and transfer excellent features; S435, mutation operation: perturb the newly generated individuals to enhance the global search capability and avoid falling into the local optimum; S436, termination judgment: when the set number of iterations or fitness convergence condition is reached, the optimization process ends; S437, output result: extract the optimal parameter combination , generating an adjustment instruction set including suspension damping, motor torque distribution and path priority .

[0011] Optionally, the S5 includes: S51, Expression-Somatics Correlation Verification: Passenger facial image data is collected in real time via an onboard camera. A trained convolutional neural network model is used to extract expression features and convert them into emotion score sequences. These scores are then synchronized with the normalized somatosensory discomfort index collected by a wearable device. S52, fuzzy model weight correction: dynamically adjust the set value of the body perception weight in the fuzzy logic model according to the calculated correlation strength, while maintaining the normalization constraint that the total weight is 1.

[0012] Optionally, the S51 expression-body correlation verification uses an on-board camera to collect passenger facial image data in real time, and extracts facial expression feature vectors through a convolutional neural network model. , and combined with the somatic discomfort index , the Pearson correlation coefficient was used to calculate the correlation between the two .

[0013] Optionally, in the S52 fuzzy logic model weight correction, when the correlation coefficient Above the set threshold When the facial expression is considered as a supplementary indicator of body sensation, the exponential sliding update method is used to dynamically adjust the weight coefficient of body sensation discomfort in the fuzzy logic model. .

[0014] A sightseeing vehicle ride comfort optimization system based on big data is used to implement the above-mentioned sightseeing vehicle ride comfort optimization method based on big data, and includes the following modules: Multi-dimensional data acquisition module: acquires vibration acceleration data, front road image data, and passenger heart rate variability data during the operation of the sightseeing car; Multi-source data fusion processing module: This module extracts frequency domain features from vibration data to generate a vibration energy distribution spectrum, performs texture segmentation and pothole identification on image data to generate a road roughness matrix, and normalizes heart rate variability data to generate a somatic discomfort index. Ride Comfort Evaluation Module: This module inputs the vibration energy distribution spectrum, road roughness matrix, and somatosensory discomfort index into a fuzzy logic model, dynamically assigns parameter weights based on the current driving scenario, and outputs a comprehensive ride comfort index. Multi-objective optimization control module: Based on comprehensive ride comfort indicators, it uses a genetic algorithm to optimize the suspension damping coefficient, drive motor torque distribution coefficient, and path planning priority parameters, generating a multi-parameter coordinated adjustment instruction set. Closed-loop feedback adaptive module: collects passenger facial expression data, performs correlation analysis with the somatic discomfort index, and modifies the weight allocation rules in the fuzzy logic model based on the analysis results to achieve dynamic optimization.

[0015] Beneficial effects of the present invention: By integrating multi-dimensional heterogeneous data such as the vibration acceleration of the sightseeing car during operation, the image of the road ahead, and the physiological reactions of passengers, this invention has established for the first time a smoothness evaluation model based on the joint perception of the three elements of "vehicle-road-passenger", breaking through the limitations of traditional reliance on a single acceleration or fixed suspension adjustment; by utilizing a fuzzy logic dynamic weighting mechanism, it has achieved an adaptive fusion evaluation of vibration energy distribution, road roughness, and somatosensory discomfort index, effectively improving the objectivity of the comprehensive smoothness index and the individual sensitivity response capability.

[0016] The present invention uses a genetic algorithm to achieve multi-parameter collaborative optimization of suspension damping, drive torque distribution and path priority, constructs a closed-loop dynamic adjustment mechanism, and introduces relevant feedback from passengers' facial expressions and somatosensory data to perform real-time correction of fuzzy model weights, significantly improving the system's adaptability to changeable road conditions and individual reactions. Compared with the existing technology, the present invention not only improves the comfort and stability of the sightseeing car during operation, but also enhances the intelligence, accuracy and user perception consistency of the control system, and has good application prospects and promotion value. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 is a flow chart of a method according to an embodiment of the present invention; Figure 2 2 is a system module diagram of an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments. It is also noted that, to provide a more detailed description, the following embodiments are best and preferred embodiments, and those skilled in the art may employ alternative methods for implementing certain known technologies. Furthermore, the accompanying drawings are intended only to provide a more detailed description of the embodiments and are not intended to limit the present invention.

[0020] It should be noted that references in the specification to "one embodiment," "an embodiment," "exemplary embodiments," "some embodiments," etc. indicate that the described embodiments may include specific features, structures, or characteristics, but not necessarily every embodiment will include such specific features, structures, or characteristics. Furthermore, when specific features, structures, or characteristics are described in conjunction with an embodiment, it is within the knowledge of persons skilled in the relevant art to implement such features, structures, or characteristics in conjunction with other embodiments (whether or not explicitly described).

[0021] In general, terms can be understood, at least in part, from their use in context. For example, depending at least in part on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending at least in part on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0022] like Figure 1 As shown, a sightseeing car ride comfort optimization method based on big data includes the following steps: S1, multi-dimensional data collection: obtain the sightseeing car vibration acceleration data through the on-board sensor, the front road image data through the camera, and the passenger heart rate variability data through the wearable device; S2, multi-source data fusion processing: frequency domain feature extraction of vibration acceleration data to obtain the vibration energy distribution spectrum, texture segmentation and pothole identification of road surface image data to generate the road surface roughness matrix, and normalization of passenger heart rate variability data to generate the somatic discomfort index; S3, Dynamic Weighted Ride Comfort Evaluation: The vibration energy distribution spectrum, road roughness matrix, and somatosensory discomfort index are input into a preset fuzzy logic model. The weights of the parameters are dynamically assigned based on the current driving scenario, and a comprehensive ride comfort index is output. S4, multi-objective optimization strategy generation: Based on the comprehensive ride comfort index, the suspension damping coefficient, drive motor torque distribution coefficient, and path planning priority parameters are simultaneously optimized through a genetic algorithm to generate a multi-parameter coordinated adjustment instruction set; S5, closed-loop feedback dynamic optimization: The real-time collected passenger facial expression data is correlated with the somatosensory discomfort index, and the weight allocation rules of the fuzzy logic model are modified based on the verification results.

[0023] S1 includes: S11, the vehicle vibration acceleration data is collected in real time by a three-axis acceleration sensor installed at the center of the front and rear axles of the vehicle body. The vibration acceleration data is expressed in vector form, which is expressed as: ; in, Respectively represent the acceleration of the vehicle in the front and rear directions, left and right directions, and up and down directions, Indicates the timestamp, the acceleration signal is sampled at the frequency Keep records; S12, through the wide-angle camera installed at the front of the vehicle, continuously collects road image frame sequences within 5 meters in front, and uses image texture features and edge gradient change rate to identify road roughness areas. Expressed as: ; in, represents the grayscale intensity value of the image, Indicates that the image is at pixel points The gradient amplitude at , is the local texture complexity coefficient, which is obtained by calculating the gray-level co-occurrence matrix; S13 collects heart rate variability (HRV) data through a photoplethysmography device worn on the passenger's wrist, using time domain indicators Calculate the somatosensory stimulation level, expressed as: ; in, is the time interval between two consecutive heartbeats, For all The mean of the interval, For the sampling period Number of intervals, The interval reflects the length of time between each heartbeat. The smaller the value, the higher the passenger's current physiological stress level, which reflects the degree of physical discomfort.

[0024] S2 includes: S21, vibration acceleration data Perform fast Fourier transform (FFT) to extract frequency domain features to obtain vibration energy distribution spectrum , expressed as: ; in, Indicates the vertical acceleration signal Perform Fourier transform, is the frequency component, Indicates the frequency The vibration energy density at S22, collect the front road image sequence Perform texture segmentation and pothole identification to generate a road roughness matrix , the roughness matrix is ​​constructed based on the image gradient and the local binary pattern (LBP) texture feature, which is expressed as: ; in, Represents the image in pixels The gradient amplitude at , Indicates the local binary pattern texture response value of the local area of ​​the pixel, which is used to characterize the road surface texture roughness. is the experience weight factor, satisfying , used to adjust the fusion ratio of gradient features and texture features, The larger the value, the rougher the road surface or the more potholes there are in that area; S23, heart rate variability index Perform standardization to generate a physical discomfort index , the standardization process uses the Z-score method, which is expressed as: ; in, Represents the SDNN mean calculated based on historical samples, represents the SDNN standard deviation of historical samples, The larger the value, the more uncomfortable the passenger feels, reflecting a significant physiological stress state.

[0025] S3 includes: S31, the vibration energy distribution spectrum , road roughness matrix and discomfort index Input them into the input end of the fuzzy logic model respectively to construct a set of ternary input variables, which can be expressed as: ; ; in, Represents the integrated energy of the frequency band in the vibration energy distribution spectrum, represents the resonance frequency range of the vehicle body structure, , Represents the road roughness matrix The mean of is used to represent the overall road condition level, is the somatic discomfort index; S32, based on the current driving scene information (vehicle speed and path curvature ), the fuzzy logic rule base dynamically assigns the weights of each input variable to form a normalized weight coefficient set, which is expressed as: ; in, are the relative importance of vibration energy, road roughness and physical discomfort in the current scenario, The weight distribution is derived from fuzzy rules, which include: "If the speed is high and the road condition is bad, the vibration energy weight is high; if the speed is low and the facial stress is significant, the body sensation weight is high"; S33, based on the assigned weight coefficient and the ternary input variables, calculates and outputs the comprehensive smoothness index , expressed as: ; in, is a comprehensive smoothness index. They are respectively The result after normalization is in the range of , The larger the value, the more unstable the vehicle's current running state is.

[0026] S4 includes: S41, Optimization variable modeling: Based on comprehensive smoothness index , constructed with the suspension damping coefficient , motor torque distribution coefficient and path planning priority parameters Modeling multi-objective optimization of optimization variables, including: ; ; in, is the damping coefficient of the sightseeing car active suspension, and its domain is , used to adjust the vehicle body vibration response, is the driving motor torque distribution coefficient, and are the driving forces of the front and rear wheel motors of the vehicle, Indicates the distribution ratio of torque to the front wheels. is the path planning priority parameter, which represents the weight of the smoothness target in the planning process, and its domain is ,The larger the value, the more inclined to choose a route with a smoother road surface but a longer distance; S42, fitness function construction: based on multi-objective optimization modeling, define the fitness function , as the optimization basis for measuring the quality of the smoothness of each parameter combination, is expressed as: ; in, To optimize the fitness value, the larger the value, the better the smoothness. is the comprehensive smoothness index output by the fuzzy logic model, which represents the comprehensive evaluation of the smoothness of the current vehicle operation state. The larger the value, the less smooth it is. The goal is to maximize the fitness function , that is, minimizing the degree of roughness; S43, Genetic Algorithm Iterative Optimization: Initialize the population and perform iterative search through selection, crossover and mutation operations to finally output the optimal parameter combination , generating a multi-parameter collaborative adjustment instruction set.

[0027] S43 uses genetic algorithm to perform global search optimization on parameter vectors. The iterative process includes: S431, population initialization: set the population size to , randomly generated initial individuals, each individual is a set of control parameter combinations to be optimized ; S432, Fitness Assessment: Inputting Individual Parameters into the Comprehensive Ride Compliance Model , calculate the fitness value , used to measure the smooth performance of the current combination, expressed as: ; S433, selection operation: based on the fitness level, prioritize the best individuals from the current population as the next generation parent; S434, crossover operation: Generate new individuals between parents through the crossover algorithm, expand the search space and transfer excellent features, expressed as: ; in, is a crossover control factor, whose value is a function of the distance between the parent individuals to control the balance between local search and global exploration; S435, mutation operation: perturb the newly generated individuals to enhance the global search capability and avoid falling into the local optimum, expressed as: ; in, is a disturbance factor that obeys the normal distribution, is the allowed range of the parameter; S436, termination judgment: when the set number of iterations or fitness convergence condition is reached, the optimization process ends; S437, output result: extract the optimal parameter combination , generating an adjustment instruction set including suspension damping, motor torque distribution and path priority .

[0028] S5 includes: S51, Expression-Somatics Correlation Verification: Passenger facial image data is collected in real time via an onboard camera. A trained convolutional neural network model is used to extract expression features and convert them into emotion score sequences. These scores are then synchronized with the normalized somatosensory discomfort index collected by a wearable device. S52, fuzzy model weight correction: dynamically adjust the set value of the body perception weight in the fuzzy logic model according to the calculated correlation strength, while maintaining the normalization constraint that the total weight is 1.

[0029] S51 Expression-body correlation verification uses an on-board camera to collect passenger facial image data in real time and extracts facial expression feature vectors through a convolutional neural network (CNN) model. , , and combined with the somatic discomfort index , the Pearson correlation coefficient was used to calculate the correlation between the two , expressed as: ; in, Indicates the The discomfort index of the time slice, Indicates the The facial expression feature scores in the time slices, and They are and The mean of is the total number of sampling points, , the larger its absolute value is, the stronger the consistency between facial expression and somatosensory data is.

[0030] In the weight correction of S52 fuzzy logic model, when the correlation coefficient Above the set threshold When the facial expression is considered as a supplementary indicator of body sensation, the exponential sliding update method is used to dynamically adjust the weight coefficient of body sensation discomfort in the fuzzy logic model. , expressed as: ; in, and Represent the current and updated somatosensory weights respectively, is an estimate based on the impact of facial emotions, which can be normalized by the proportion of facial tension. is the weight update rate coefficient, which determines the response speed. .

[0031] like Figure 2 As shown, a sightseeing car ride comfort optimization system based on big data is used to implement the above-mentioned sightseeing car ride comfort optimization method based on big data, including the following modules: Multi-dimensional data acquisition module: acquires vibration acceleration data, front road image data, and passenger heart rate variability data during the operation of the sightseeing car; Multi-source data fusion processing module: This module extracts frequency domain features from vibration data to generate a vibration energy distribution spectrum, performs texture segmentation and pothole identification on image data to generate a road roughness matrix, and normalizes heart rate variability data to generate a somatic discomfort index. Ride Comfort Evaluation Module: This module inputs the vibration energy distribution spectrum, road roughness matrix, and somatosensory discomfort index into a fuzzy logic model, dynamically assigns parameter weights based on the current driving scenario, and outputs a comprehensive ride comfort index. Multi-objective optimization control module: Based on comprehensive ride comfort indicators, it uses a genetic algorithm to optimize the suspension damping coefficient, drive motor torque distribution coefficient, and path planning priority parameters, generating a multi-parameter coordinated adjustment instruction set. Closed-loop feedback adaptive module: collects passenger facial expression data, performs correlation analysis with the somatic discomfort index, and modifies the weight allocation rules in the fuzzy logic model based on the analysis results to achieve dynamic optimization.

[0032] The present invention encompasses any alternatives, modifications, equivalents, and solutions that fall within the spirit and scope of the present invention. To provide a thorough understanding of the present invention, specific details are described in detail below in connection with the preferred embodiments of the present invention, but those skilled in the art will be able to fully understand the present invention without these detailed descriptions. Furthermore, to avoid unnecessary confusion regarding the essence of the present invention, well-known methods, processes, procedures, components, and circuits have not been described in detail.

[0033] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A sightseeing car ride comfort optimization method based on big data, characterized in that: The following steps are involved: S1, multi-dimensional data collection: obtain the vibration acceleration data of the sightseeing car through the on-board sensor, collect the image data of the road ahead through the camera, and obtain the heart rate variability data of the passengers through the wearable device; S2, multi-source data fusion processing: frequency domain feature extraction of vibration acceleration data to obtain vibration energy distribution spectrum, texture segmentation and pothole recognition of road image data to generate road roughness matrix, and normalization of passenger heart rate variability data to generate somatic discomfort index; S3, dynamic weighted ride comfort evaluation: the vibration energy distribution spectrum, road roughness matrix and somatosensory discomfort index are input into the preset fuzzy logic model, the weight of each parameter is dynamically allocated according to the current driving scenario, and the comprehensive ride comfort index is output; S4, multi-objective optimization strategy generation: Based on the comprehensive smoothness index, the suspension damping coefficient, the drive motor torque distribution coefficient and the path planning priority parameter are optimized simultaneously through the genetic algorithm to generate a multi-parameter coordinated adjustment instruction set; S5, closed-loop feedback dynamic optimization: verify the correlation between the real-time collected passenger facial expression data and the somatosensory discomfort index, and modify the weight allocation rule of the fuzzy logic model according to the verification results.

2. The method for optimizing the ride comfort of a sightseeing vehicle based on big data according to claim 1, characterized in that: The S1 includes: S11, collecting vehicle vibration acceleration data in real time through a three-axis acceleration sensor installed at the center of the front and rear axles of the vehicle body; S12, using a wide-angle camera installed at the front of the vehicle to continuously collect road image frame sequences within a range of 5 meters ahead, and using image texture features and edge gradient change rate to identify road roughness areas; S13, collects heart rate variability data through a photoplethysmography device worn on the passenger's wrist, using time domain indicators Calculate the somatosensory stimulation level.

3. The method for optimizing the ride comfort of a sightseeing vehicle based on big data according to claim 2, characterized in that: The S2 includes: S21, performing fast Fourier transform on the vibration acceleration data to extract frequency domain features to obtain a vibration energy distribution spectrum; S22, performing texture segmentation and pothole recognition on the collected front road surface image sequence to generate a road surface roughness matrix, wherein the roughness matrix is ​​jointly constructed based on image gradient and local binary pattern texture features; S23, the heart rate variability index Standardize the process to generate a physical discomfort index .

4. The method for optimizing the ride comfort of a sightseeing vehicle based on big data according to claim 3, characterized in that: The S3 includes: S31, inputting the vibration energy distribution spectrum, road roughness matrix and body discomfort index into the input end of the fuzzy logic model respectively to construct a ternary input variable set; S32, dynamically allocating weights of input variables according to current driving scene information by a fuzzy logic rule base to form a normalized weight coefficient set; S33, calculating and outputting a comprehensive smoothness index based on the allocated weight coefficient and the ternary input variables.

5. The method for optimizing the ride comfort of a sightseeing vehicle based on big data according to claim 4, characterized in that: The S4 includes: S41, Optimization variable modeling: Based on comprehensive smoothness index , constructed with suspension damping coefficient , Motor torque distribution coefficient and path planning priority parameters Modeling multi-objective optimization of optimization variables; S42, fitness function construction: Based on multi-objective optimization modeling, a fitness function is defined as the optimization basis for measuring the smoothness of each parameter combination; S43, Genetic Algorithm Iterative Optimization: Initialize the population and perform iterative search through selection, crossover and mutation operations, and finally output the optimal parameter combination , generating a multi-parameter collaborative adjustment instruction set.

6. The method for optimizing the ride comfort of a sightseeing vehicle based on big data according to claim 5, characterized in that: S43 uses a genetic algorithm to perform global search optimization on the parameter vector, and the iterative process includes: S431, population initialization: set the population size to , randomly generated initial individuals, each of which is a set of control parameter combinations to be optimized ; S432, fitness evaluation: input individual parameters into the comprehensive ride comfort model and calculate the fitness value , used to measure the smooth performance of the current combination; S433, selection operation: based on the fitness level, prioritize the best individuals from the current population as the next generation parent; S434, crossover operation: Generate new individuals between parents through the crossover algorithm, expand the search space and transfer excellent features; S435, mutation operation: perturb the newly generated individuals to enhance the global search capability and avoid falling into the local optimum; S436, termination judgment: when the set number of iterations or fitness convergence condition is reached, the optimization process ends; S437, output result: extract the optimal parameter combination , generating a set of adjustment instructions including suspension damping, motor torque distribution and path priority .

7. The method for optimizing the ride comfort of a sightseeing vehicle based on big data according to claim 6, characterized in that: The S5 includes: S51, Expression-body feeling correlation verification: The passenger’s facial image data is collected in real time through the on-board camera, and the expression features are extracted using the trained convolutional neural network model. The expression features are converted into an emotion score sequence, which is synchronized with the body feeling discomfort index collected and standardized by the wearable device; S52, fuzzy model weight correction: dynamically adjust the set value of the body sense weight in the fuzzy logic model according to the calculated correlation strength, while maintaining the normalization constraint that the sum of the weights is 1.

8. The method for optimizing the ride comfort of a sightseeing vehicle based on big data according to claim 7, characterized in that: The S51 expression-body perception correlation verification uses the on-board camera to collect passenger facial image data in real time and extracts facial expression feature vectors through the convolutional neural network model. , combined with the physical discomfort index , the Pearson correlation coefficient was used to calculate the correlation between the two .

9. The method for optimizing the ride comfort of a sightseeing vehicle based on big data according to claim 7, characterized in that: In the weight correction of the S52 fuzzy logic model, when the correlation coefficient Above the set threshold When the facial expression is considered as a supplementary indicator of body sensation, the exponential sliding update method is used to dynamically adjust the weight coefficient of body sensation discomfort in the fuzzy logic model. .

10. A sightseeing car ride comfort optimization system based on big data, used to implement a sightseeing car ride comfort optimization method based on big data as claimed in any one of claims 1 to 9, characterized in that: Includes the following modules: Multi-dimensional data acquisition module: obtains vibration acceleration data, front road image data and passenger heart rate variability data during the operation of the sightseeing car; Multi-source data fusion processing module: extracts frequency domain features of vibration data to generate vibration energy distribution spectrum, performs texture segmentation and pothole recognition on image data to generate road roughness matrix, and standardizes heart rate variability data to generate somatic discomfort index; Ride comfort evaluation module: Input the vibration energy distribution spectrum, road roughness matrix and somatosensory discomfort index into the fuzzy logic model, dynamically assign parameter weights according to the current driving scenario, and output a comprehensive ride comfort index; Multi-objective optimization control module: Based on the comprehensive smoothness index, the suspension damping coefficient, the drive motor torque distribution coefficient and the path planning priority parameter are optimized through the genetic algorithm to generate a multi-parameter coordinated adjustment instruction set; Closed-loop feedback adaptive module: collects passenger facial expression data, conducts correlation analysis with the somatic discomfort index, and modifies the weight allocation rules in the fuzzy logic model based on the analysis results to achieve dynamic optimization.

Images