Design method and system based on deep learning

Abstract

The invention relates to the technical field of computers, in particular to a design method and system based on deep learning. The method comprises the following steps: sequentially carrying out feature extraction on static three-dimensional body shape data and dynamic riding posture data to obtain static three-dimensional body shape features and dynamic riding posture features; performing feature fusion on the static three-dimensional body feature and the dynamic riding posture feature according to a preset proportion to obtain a fused feature; inputting the fusion features into a preset deep learning model to obtain personalized customized bicycle parameters of the current user; determining a three-dimensional model of the personalized customized bicycle based on the personalized customized bicycle parameters; when the three-dimensional model of the personalized customized bicycle does not meet the personalized constraint condition of the bicycle, the step of obtaining the static three-dimensional body shape data and the dynamic riding posture data of the current user is executed again. Through the configuration mode, the generation quality of the personalized design scheme of the user can be accurately improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, and particularly relates to a design method and system based on deep learning. BACKGROUND

[0002] Traditional bicycle design and manufacturing usually adopts a standardized and large-scale production mode, which is difficult to meet the increasingly diversified personalized needs of consumers. Although there are some customized bicycle services in the market, most of them rely on manual experience and simple rule matching, which has low design efficiency and poor accuracy, and cannot fully tap the potential needs of users and generate high-quality personalized design schemes. With the development of deep learning technology, it has achieved remarkable results in image generation, pattern recognition and other fields, providing new ideas and methods for bicycle personalized customization design.

[0003] Therefore, the present application provides a design method and system based on deep learning to solve the above technical problems. SUMMARY

[0004] The present application describes a design method and system based on deep learning, which can accurately improve the generation quality of user personalized design scheme.

[0005] According to a first aspect, the present application provides a design method based on deep learning, which comprises:

[0006] Obtaining static three-dimensional body shape data and dynamic riding posture data of a current user;

[0007] Sequentially extracting features from the static three-dimensional body shape data and the dynamic riding posture data to obtain static three-dimensional body shape features and dynamic riding posture features;

[0008] Fusing the static three-dimensional body shape features and the dynamic riding posture features according to a preset proportion to obtain fusion features;

[0009] Inputting the fusion features into a preset deep learning model to obtain personalized customization bicycle parameters of the current user;

[0010] Determining a three-dimensional model of the personalized customization bicycle based on the personalized customization bicycle parameters;

[0011] When the three-dimensional model of the personalized customization bicycle does not meet the personalized constraint conditions of the bicycle, re-executing the step of obtaining the static three-dimensional body shape data and the dynamic riding posture data of the current user.

[0012] According to a second aspect, the present application provides a design system based on deep learning, which comprises:

[0013] An acquisition unit is configured to acquire static three-dimensional body shape data and dynamic riding posture data of a current user;

[0014] A first data processing unit is configured to sequentially perform feature extraction on the static three-dimensional body shape data and the dynamic riding posture data to obtain static three-dimensional body shape features and dynamic riding posture features;

[0015] A second data processing unit is configured to perform feature fusion on the static three-dimensional body shape features and the dynamic riding posture features according to a preset ratio to obtain fused features;

[0016] A third data processing unit is configured to input the fused features into a preset deep learning model to obtain personalized customized bicycle parameters of the current user;

[0017] A fourth data processing unit is configured to determine a three-dimensional model of a personalized customized bicycle based on the personalized customized bicycle parameters;

[0018] A fifth data processing unit is configured to, when the three-dimensional model of the personalized customized bicycle does not satisfy a personalized constraint condition of the bicycle, re-perform the step of acquiring the static three-dimensional body shape data and the dynamic riding posture data of the current user.

[0019] In a third aspect, an electronic device is provided, including a memory and a processor, the memory storing a computer program, and the processor implementing the method of any of the embodiments of the present specification when executing the computer program.

[0020] In a fourth aspect, a computer-readable storage medium is provided, storing a computer program, and the computer program, when executed in a computer, causes the computer to perform the method of any of the embodiments of the present specification.

[0021] According to the design method and system based on deep learning provided by the application, the static three-dimensional body data and dynamic riding posture data of the current user are collected in real time through a multi-source sensor array and a three-dimensional scanning technology. The static three-dimensional body data and dynamic riding posture data are processed respectively by using a feature extraction algorithm in the field of deep learning. Specifically, for the static three-dimensional body data, a fully connected neural network architecture is used to realize the extraction of static three-dimensional body features through multi-layer nonlinear transformation; for the dynamic riding posture data, a hybrid model of long short-term memory network (LSTM) and convolutional neural network (CNN) is used to effectively capture time series information and local features, and then obtain dynamic riding posture features. Based on the preset weight parameters, the static three-dimensional body features and dynamic riding posture features are weighted and fused to obtain a fusion feature vector containing ergonomics and motion characteristics. The fusion feature vector is fed into a preset deep learning regression model based on a multi-layer perceptron (MLP) as an input, and through the forward propagation operation of the model, the personalized customization bicycle parameters of the current user are output, which include key design elements such as geometric size and component specification. The three-dimensional digital model of the personalized customization bicycle is constructed according to the personalized customization bicycle parameters by using computer-aided design (CAD) software. After the construction is completed, the three-dimensional digital model is comprehensively evaluated based on a pre-set set of personalized constraint conditions. If the evaluation result shows that the three-dimensional digital model fails to meet any of the above-mentioned personalized constraint conditions, the iteration process of data collection and model construction is triggered, and the step of collecting the static three-dimensional body data and dynamic riding posture data of the current user through a multi-source sensor array and a three-dimensional scanning technology is re-executed until a three-dimensional digital model that meets all the personalized constraint conditions is generated. Through the above closed-loop design process, the generation quality and design accuracy of the user's personalized design scheme can be significantly improved, and the customized bicycle can achieve an optimal balance in terms of human adaptability, functional performance and manufacturing feasibility. BRIEF DESCRIPTION OF DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, brief descriptions will be given below for the drawings needed to be used in the embodiments or prior art descriptions. Obviously, the drawings described below are some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0023] Figure 1 A flowchart of a design method based on deep learning according to an embodiment is shown;

[0024] Figure 2 A schematic block diagram of a design system based on deep learning according to an embodiment is shown. DETAILED DESCRIPTION

[0025] The scheme provided by the present application is described below in combination with the drawings.

[0026] Figure 1 A flowchart of a deep learning-based design method according to an embodiment is shown. It can be understood that the method can be executed by any system, device, platform, device cluster having computing and processing capabilities. As shown, the method comprises: Figure 1

[0027] Step 100, acquiring static three-dimensional body shape data and dynamic riding posture data of a current user;

[0028] Step 102, sequentially performing feature extraction on the static three-dimensional body shape data and the dynamic riding posture data to obtain static three-dimensional body shape features and dynamic riding posture features;

[0029] Step 104, performing feature fusion on the static three-dimensional body shape features and the dynamic riding posture features according to a preset ratio to obtain fusion features;

[0030] Step 106, inputting the fusion features into a preset deep learning model to obtain personalized customized bicycle parameters of the current user;

[0031] Step 108, determining a three-dimensional model of the personalized customized bicycle based on the personalized customized bicycle parameters;

[0032] Step 110, when the three-dimensional model of the personalized customized bicycle does not satisfy the personalized constraint conditions of the bicycle, re-executing the step of acquiring the static three-dimensional body shape data and the dynamic riding posture data of the current user.

[0033] ​In the embodiment, the static three-dimensional body data and the dynamic riding posture data of the current user are collected in real time by a multi-source sensor array and a three-dimensional scanning technology. The static three-dimensional body data and the dynamic riding posture data are processed respectively by using a feature extraction algorithm in the field of deep learning. Specifically, for the static three-dimensional body data, a fully connected neural network architecture is used to realize the extraction of the static three-dimensional body feature through multi-layer nonlinear transformation; for the dynamic riding posture data, a hybrid model of long short-term memory network (LSTM) and convolutional neural network (CNN) is used to effectively capture the time series information and local features, and then the dynamic riding posture feature is obtained. Based on the preset weight parameters, the static three-dimensional body feature and the dynamic riding posture feature are weighted and fused to obtain a fusion feature vector containing ergonomics and motion characteristics. The fusion feature vector is fed into a preset deep learning regression model based on a multi-layer perceptron (MLP) as an input, and through the forward propagation operation of the model, the personalized customization bicycle parameters of the current user are output, which include key design elements such as geometric size and component specification. The three-dimensional digital model of the personalized customization bicycle is constructed according to the personalized customization bicycle parameters by using computer-aided design (CAD) software. After the construction is completed, the three-dimensional digital model is comprehensively evaluated based on a pre-set set of personalized constraint conditions. If the evaluation result shows that the three-dimensional digital model fails to meet any of the above-mentioned personalized constraint conditions, the iteration process of data collection and model construction is triggered, and the step of collecting the static three-dimensional body data and the dynamic riding posture data of the current user by the multi-source sensor array and the three-dimensional scanning technology is re-executed until the three-dimensional digital model meeting all the personalized constraint conditions is generated. Through the above closed-loop design process, the generation quality and design accuracy of the user's personalized design scheme can be significantly improved, and the customized bicycle can achieve an optimal balance in terms of human adaptability, functional performance and manufacturing feasibility.

[0034] In an embodiment of the present application, the personalized constraint conditions include a frame stiffness constraint and a joint angle constraint.

[0035] In an embodiment of the present application, the joint angle constraint is constructed by the following formula:

[0036] |θ min ≤θ knee ≤θ max |

[0037]

[0038] In the formula, θ knee is the knee joint angle, θ min is the lower limit of the knee joint angle, θ max is the lower limit of the knee joint angle, L leg is the user's leg length, and Hseat is the seat tube height, t peldl is the pedal thickness, t shoe is the sole thickness, L crank is the crank length, and a is the angle of the crank relative to the horizontal direction, is the rate of change of the crank angle, is the acceleration of the crank angle, t is time, and f(a(t), t) is a fatigue influence function related to the crank angle and time.

[0039] In this embodiment, by constructing a joint angle constraint based on biomechanics, the generation quality and design accuracy of the user's personalized design scheme can be effectively improved. Based on the principles of ergonomics, the range of motion and motion trajectory of key joints such as the knee joint, hip joint, and shoulder joint during cycling are comprehensively considered, and by establishing an accurate mathematical model and parameterized constraint condition, the joint angle is limited within a comfortable and efficient range. In the specific implementation process, first, the user's static three-dimensional body data and dynamic cycling posture data are analyzed in depth, and the key features related to the joint angle are extracted. Then, these features are integrated into the personalized design process as important constraint conditions, and the design parameters such as the geometric size of the bicycle and the layout of the components are optimized and adjusted. In this way, it ensures that the customized bicycle can minimize the risk of fatigue and injury during cycling in terms of human adaptability; in terms of functional performance, it ensures that the controllability and power transmission efficiency of the bicycle are in the best state; in terms of manufacturing feasibility, it takes into account the machinability and cost-effectiveness of the design scheme, thereby achieving an optimal balance in terms of human adaptability, functional performance, and manufacturing feasibility.

[0040] In one embodiment of the present application, the frame stiffness constraint is constructed by the following formula:

[0041]

[0042] In the formula, K torsion is the torsional stiffness of the frame, t is the integral variable, T(t) is the time-varying torsional moment, is the time-varying torsional angle, c torsion is the torsional damping coefficient, is the rate of change of the torsional angle, I torsion is the torsional moment of inertia, is the acceleration of the torsional angle, k dir is the direction coefficient, K min is the minimum torsional stiffness requirement.

[0043] In this embodiment, by constructing a multi-dimensional frame stiffness constraint system, the generation quality and design accuracy of the user's personalized design scheme can be significantly improved. Based on the principles of material mechanics and structural dynamics, the system establishes accurate mathematical constraints and parameterized threshold standards from three aspects of static stiffness, dynamic stiffness and fatigue stiffness. In the specific implementation process, first, the user's personalized requirements such as riding habits, weight distribution, and riding conditions are quantitatively analyzed, and the required stiffness indicators of the frame under different working conditions are determined in combination with the use scenarios of the bicycle (such as road racing, mountain off-road, urban commuting, etc.). In the design process, the finite element analysis (FEA) technology is used to simulate the mechanics of the frame, and the stiffness of the frame is verified in real time to meet the constraint requirements. By continuously adjusting the geometric size, material distribution and structural details of the frame, the lightweight design of the frame is optimized while ensuring that the frame has sufficient load-bearing capacity and anti-deformation ability. In this way, in terms of human adaptability, the frame can provide stable and comfortable support for the user during riding, reducing the riding fatigue caused by frame deformation; in terms of functional performance, the response speed and power transmission efficiency of the bicycle are improved to meet the needs of different riding scenarios; in terms of manufacturing feasibility, the design scheme is ensured to meet the production process requirements, reducing the manufacturing cost and production difficulty, and finally realizing the optimal balance of the customized bicycle in terms of human adaptability, functional performance and manufacturing feasibility.

[0044] In an embodiment of the present application, the static three-dimensional body data includes height, weight, leg length, and ischial tuberosity spacing, and the dynamic riding posture data includes pedal frequency, center of gravity distribution, and joint activity angle; the personalized customized bicycle parameters include the sizes of the bicycle components.

[0045] In this embodiment, the static three-dimensional body data acquisition system covers key anthropometric parameters such as height, weight, leg length, and ischial tuberosity spacing, and achieves millimeter-level precision data acquisition through high-precision three-dimensional scanning equipment; the dynamic riding posture data includes motion biomechanics parameters such as pedal frequency, center of gravity distribution, and joint activity angle, which are monitored in real time by multi-axis inertial sensors and motion capture systems. The generated personalized customized bicycle parameters focus on the accurate size specifications of the bicycle components, including but not limited to frame geometry (such as head tube length, top tube angle), handlebar width, seat tube height, crank length, and other core component parameters, providing quantitative design basis for high-precision bicycle personalized customization.

[0046] In an embodiment of the present application, after determining the three-dimensional model of the personalized customized bicycle based on the personalized customized bicycle parameters, the method further comprises:

[0047] Obtaining heart rate data of the user's historical riding;

[0048] determine a heart rate fluctuation curve of the historical cycling of the user based on heart rate data of the historical cycling of the user;

[0049] determine a seat comfort adjustment parameter based on the heart rate fluctuation curve;

[0050] adjust the seat comfort of the three-dimensional model of the personalized customized bicycle based on the seat comfort adjustment parameter.

[0051] In the embodiment, after the three-dimensional model is constructed based on the personalized customized bicycle parameters, the design optimization process is further expanded. Specifically, the heart rate data of the historical cycling of the user is collected through the wearable device and the built-in sensor of the bicycle, which covers the real-time change information of the heart rate during the whole cycling. Then, the time series analysis method is used to process the collected heart rate data, and through the preprocessing steps such as data smoothing and outlier elimination, the heart rate fluctuation curve of the historical cycling of the user is constructed, which intuitively presents the dynamic change trend of the heart rate with time, cycling intensity and other factors during the cycling. Based on the heart rate fluctuation curve, combined with the fatigue evaluation model and comfort quantification standard in the biomedical field, the characteristic parameters such as the heart rate fluctuation amplitude and the proportion of high-frequency fluctuation are analyzed in depth, and the mapping relationship between the heart rate fluctuation characteristics and the seat comfort is established through regression analysis or machine learning algorithm, and then the seat comfort adjustment parameters are determined, including but not limited to the seat shape curvature parameters, the filling hardness coefficient, the surface material friction coefficient and other key indicators. Finally, the determined seat comfort adjustment parameters are fed back to the three-dimensional model design link of the personalized customized bicycle, and the parametric modeling function of the computer aided design (CAD) software is used to make targeted adjustment and optimization on the seat geometry, material properties and other parameters in the three-dimensional model, so as to ensure that the customized bicycle can provide the user with a cycling experience that meets the physiological characteristics and comfort needs in the actual cycling process, and realize the complete closed loop from parameter design to user experience optimization.

[0052] In one embodiment of the present application, after the three-dimensional model of the personalized customized bicycle is determined based on the personalized customized bicycle parameters, the following steps are further included:

[0053] acquire route data of the historical cycling of the user;

[0054] determine a friction adjustment coefficient of the wheel based on the route data of the historical cycling of the user;

[0055] adjust the pattern of the tire in the three-dimensional model of the personalized customized bicycle based on the friction adjustment coefficient of the wheel.

[0056] In the present embodiment, after the three-dimensional model is constructed according to the personalized customization of the bicycle parameters, a further deepening design optimization process is carried out. Specifically, through a riding record device and a geographic information system (GIS) data interface, route data of the user's historical riding is obtained, which contains multi-dimensional information such as route terrain characteristics (slope, roughness), road surface material (asphalt, cement, gravel, etc.), environmental temperature and humidity, etc. Subsequently, based on the historical route data, a road surface friction characteristic analysis model is used, combined with the empirical formula of the friction coefficient under different road surface materials and environmental conditions, to construct the route-friction coefficient mapping relationship. Through data mining and machine learning algorithms, the correlation between historical route characteristics and actual wheel friction requirements is analyzed, and the wheel friction adjustment coefficient suitable for the user's riding habits and commonly encountered road conditions is determined, which comprehensively reflects the performance indicators such as the required tire grip, wear resistance, etc. Finally, the determined wheel friction adjustment coefficient is fed back to the three-dimensional model design link, and with the aid of computer-aided design (CAD) and simulation analysis software, the tire pattern in the three-dimensional model of the personalized customized bicycle is parameterized adjusted. According to the friction adjustment coefficient, the pattern depth, shape (block, longitudinal, mixed pattern) and arrangement density are optimized to ensure that the tire can provide sufficient grip to ensure riding safety and also take into account the rolling resistance and service life under the user's commonly ridden road conditions, realizing the precise matching of the performance of the customized bicycle and the actual riding needs of the user.

[0057] The above describes specific embodiments of the present application. Other embodiments are within the scope of the appended claims. In some cases, the acts or steps recited in the claims can be performed in an order other than that in which they are recited in the embodiments and still achieve desirable results. Also, the processes depicted in the figures do not necessarily require the particular order shown, or sequential order, to achieve the desired results. In certain implementations, multitasking and parallel processing can be advantageous.

[0058] According to another aspect, embodiments of the present application provide a deep learning based design system. Figure 2 A schematic block diagram of a deep learning based design system according to one embodiment is shown. It can be appreciated that the system can be implemented by any system, device, platform and cluster of devices having computing, processing capabilities. As shown, the system includes an acquisition unit 200, a first data processing unit 202, a second data processing unit 204, a third data processing unit 206, a fourth data processing unit 208 and a fifth data processing unit 210. The main functions of each component unit are as follows: Figure 2 The acquisition unit is configured to acquire static three-dimensional body shape data and dynamic riding posture data of a current user;

[0059]

[0060] ​The first data processing unit is configured to sequentially perform feature extraction on the static three-dimensional body shape data and the dynamic riding posture data to obtain static three-dimensional body shape features and dynamic riding posture features.

[0061] The second data processing unit is configured to perform feature fusion on the static three-dimensional body shape features and the dynamic riding posture features according to a preset ratio to obtain fusion features.

[0062] The third data processing unit is configured to input the fusion features into a preset deep learning model to obtain personalized customized bicycle parameters of the current user.

[0063] The fourth data processing unit is configured to determine a three-dimensional model of the personalized customized bicycle based on the personalized customized bicycle parameters.

[0064] The fifth data processing unit is configured to re-perform the step of "obtaining static three-dimensional body shape data and dynamic riding posture data of the current user" when the three-dimensional model of the personalized customized bicycle does not satisfy the personalized constraint condition of the bicycle.

[0065] In an embodiment of the present application, the personalized constraint condition includes a frame stiffness constraint and a joint angle constraint.

[0066] In an embodiment of the present application, the joint angle constraint is constructed by the following formula:

[0067] |θ min ≤θ knee ≤θ max |

[0068]

[0069] In the formula, θ knee is a knee joint angle, θ min is a lower limit of the knee joint angle, θ max is a lower limit of the knee joint angle, L leg is a user leg length, H seat is a seat tube height, t peldl is a pedal thickness, t shoe is a shoe sole thickness, L crank is a crank length, and α is an angle of the crank relative to the horizontal direction. is a change rate of the crank angle, is an acceleration of the crank angle, and t is time.

[0070] In an embodiment of the present application, the frame stiffness constraint is constructed by the following formula:

[0071]

[0072] wherein K torsion is the torsional stiffness of the frame, τ is the integral variable, T(τ) is the time-varying torsional moment, is the time-varying torsional angle, c torsion is the torsional damping coefficient, is the rate of change of the torsional angle, I torsion is the torsional moment of inertia, is the acceleration of the torsional angle, k dir is the directional coefficient, K min is the minimum torsional stiffness requirement.

[0073] In one embodiment of the present application, the static three-dimensional body data includes height, weight, leg length, and ischial tuberosity distance, the dynamic riding posture data includes pedaling frequency, center of gravity distribution, and joint activity angle, and the personalized customized bicycle parameters include sizes of bicycle components.

[0074] In one embodiment of the present application, after the three-dimensional model of the personalized customized bicycle is determined based on the personalized customized bicycle parameters, the method further comprises:

[0075] obtaining heart rate data of historical riding of the user;

[0076] determining a heart rate fluctuation curve of the historical riding of the user based on the heart rate data of the historical riding of the user;

[0077] determining a saddle comfort adjustment parameter based on the heart rate fluctuation curve;

[0078] adjusting the saddle comfort in the three-dimensional model of the personalized customized bicycle based on the saddle comfort adjustment parameter.

[0079] In one embodiment of the present application, after the three-dimensional model of the personalized customized bicycle is determined based on the personalized customized bicycle parameters, the method further comprises:

[0080] obtaining route data of historical riding of the user;

[0081] determining a friction adjustment coefficient of the wheel based on the route data of the historical riding of the user;

[0082] adjusting a tread pattern of a tire in the three-dimensional model of the personalized customized bicycle based on the friction adjustment coefficient of the wheel.

[0083] According to another aspect, embodiments also provide a computer readable storage medium having stored thereon a computer program which, when executed in a computer, cause the computer to perform the method described in connection with Figure 1 the above.

[0084] According to an embodiment of still another aspect, an electronic device is also provided, including a memory and a processor, the memory having stored therein executable code which, when executed by the processor, implements the method as Figure 1 described above.

[0085] Each of the embodiments of the present application is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other. Each of the embodiments mainly describes the difference from other embodiments. In particular, the system embodiment is described simply because it is basically similar to the method embodiment, and the relevant part can be referred to the description of the method embodiment.

[0086] Those skilled in the art should understand that, in one or more examples described above, the functions described in the present application can be implemented by hardware, software, firmware or any combination thereof. When implemented by software, the functions can be stored in a computer readable medium or transmitted as one or more instructions or codes on a computer readable medium.

[0087] The above detailed description sets forth the purpose, technical solutions and beneficial effects of the present application. It should be understood that the above description is only a specific embodiment of the present application, and is not used to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. made on the basis of the technical solutions of the present application shall be included in the protection scope of the present application.

Claims

1. A deep learning-based design method, characterized by, The method comprises: acquiring static three-dimensional body shape data and dynamic riding posture data of a current user; sequentially extracting features from the static three-dimensional body shape data and the dynamic riding posture data to obtain static three-dimensional body shape features and dynamic riding posture features; performing feature fusion on the static three-dimensional body shape features and the dynamic riding posture features according to a preset ratio to obtain fused features; inputting the fused features into a preset deep learning model to obtain personalized customized bicycle parameters of the current user; determining a three-dimensional model of a personalized customized bicycle based on the personalized customized bicycle parameters; when the three-dimensional model of the personalized customized bicycle does not satisfy a personalized constraint condition of the bicycle, re-executing the step of acquiring the static three-dimensional body shape data and the dynamic riding posture data of the current user.

2. The method of claim 1, wherein, The personalized constraint condition comprises a frame stiffness constraint and a joint angle constraint.

3. The method of claim 2, wherein, The joint angle constraint is constructed by the following formula: | θ min ≤ θ knee ≤ θ max | where θ knee is the knee angle, θ min is the lower limit of the knee angle, θ max is the lower limit of the knee angle, L leg is the user leg length, H seat is the seat tube height, t peldl is the pedal thickness, t shoe is the shoe sole thickness, L crank is the crank length, α is the crank angle relative to the horizontal direction, is the rate of change of the crank angle, is the acceleration of the crank angle, t is time, and f(α(τ),τ) is a fatigue influence function related to the crank angle and time.

4. The method of claim 2, wherein, The frame stiffness constraint is constructed by the following formula: where K torsion is the torsional stiffness of the frame, τ is the integral variable, T(τ) is the time-varying torsional moment, is the time-varying torsional angle, c torsion is the torsional damping coefficient, is the rate of change of the torsional angle, I torsion is the torsional moment of inertia, is the acceleration of the torsional angle, k dir is the directional coefficient, K min is the minimum torsional stiffness requirement.

5. The method of claim 1, wherein, The static three-dimensional body shape data comprises height, weight, leg length, and ischial tuberosity spacing, the dynamic riding posture data comprises pedaling frequency, center of gravity distribution, and joint activity angle, and the personalized customized bicycle parameters comprise sizes of bicycle component parts.

6. The method of claim 1, wherein, After the step of determining the three-dimensional model of the personalized customized bicycle based on the personalized customized bicycle parameters, the method further comprises: acquiring heart rate data of historical riding of the user; determining a heart rate fluctuation curve of the historical riding of the user based on the heart rate data; determining a seat cushion comfort adjustment parameter based on the heart rate fluctuation curve; adjusting seat cushion comfort in the three-dimensional model of the personalized customized bicycle based on the seat cushion comfort adjustment parameter.

7. The method of claim 1, wherein, After the step of determining the three-dimensional model of the personalized customized bicycle based on the personalized customized bicycle parameters, the method further comprises: acquiring route data of historical riding of the user; determining a friction adjustment coefficient of a wheel based on the route data of the historical riding of the user; adjusting a pattern of a tire in the three-dimensional model of the personalized customized bicycle based on the friction adjustment coefficient of the wheel.

8. A deep learning-based design system, characterized by, The method comprises: an acquisition unit configured to acquire static three-dimensional body shape data and dynamic riding posture data of a current user; a first data processing unit configured to sequentially extract features from the static three-dimensional body shape data and the dynamic riding posture data to obtain static three-dimensional body shape features and dynamic riding posture features; a second data processing unit configured to perform feature fusion on the static three-dimensional body shape features and the dynamic riding posture features according to a preset ratio to obtain fused features; a third data processing unit configured to input the fused features into a preset deep learning model to obtain personalized customized bicycle parameters of the current user; a fourth data processing unit configured to determine a three-dimensional model of a personalized customized bicycle based on the personalized customized bicycle parameters; and a fifth data processing unit configured to, when the three-dimensional model of the personalized customized bicycle does not satisfy a personalized constraint condition of the bicycle, re-execute the step of acquiring the static three-dimensional body shape data and the dynamic riding posture data of the current user. a fifth data processing unit configured to re-perform the step "obtaining static three-dimensional body shape data and dynamic riding posture data of the current user" when the three-dimensional model of the personalized custom bicycle does not satisfy the individualization constraints of the bicycle.

9. An electronic device, comprising: A computer program product comprising a computer readable medium having stored thereon computer program means, the computer program product comprising computer program instructions which, when executed by a computer, cause the computer to carry out the method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, A computer program product comprising a computer readable medium having stored thereon computer program means, the computer program product comprising computer program instructions which, when executed by a computer, cause the computer to carry out the method according to any one of claims 1-7.

Images