Ebook: Innovative Design and Intelligent Manufacturing

Innovative design and intelligent manufacturing represent an increasingly prominent area in the ever evolving and innovative global manufacturing industry. It combines the principles of innovative design with the application of intelligent manufacturing technologies, bringing significant changes and opportunities to production and product development. Innovative design plays a crucial role in intelligent manufacturing, emphasizing creative thinking in the stages of product design and development and highlighting user experience, functionality, and sustainability. Through the integration of innovative design with intelligent manufacturing technologies, a new vitality and inspiration are injected into the manufacturing industry. The intersection and fusion of innovative design and intelligent manufacturing offer immense potential for future industrial development. By combining creative thinking with advanced technology, we can achieve a more intelligent, efficient, and sustainable manufacturing industry.

The 2024 International Conference on Innovative Design and Intelligent Manufacturing (IDIM 2024) took place in Bali, Indonesia, on 20 and 21 July 2024. The aim of the conference is to gather scholars, experts, and industry professionals from around the world to discuss the latest advancements and challenges in the fields of innovative design and intelligent manufacturing. The focus is on interdisciplinary intersections and exploring how design innovation and intelligent manufacturing technologies can drive industrial progress and societal development. We look forward to sharing and discussing research findings and insights in the fields of innovative design and intelligent manufacturing, including digitization and automation in manufacturing processes, smart manufacturing and Industry 4.0, additive manufacturing and 3D printing, smart sensors and the Internet of Things (IoT), innovative design and user experience, human-machine collaboration and interfaces, sustainable manufacturing, and environmental technologies, among others.

IDIM 2024 was sponsored by Zhejiang Sci-Tech University of China, and technically co-organized by Shanghai Jiao Tong University. It was jointly organized by Nanjing University of Posts and Telecommunications of China, the American International Institute of Design Engineers, KES International, the Indian Institute of Technology, Tanta University of Egypt, and Xiamen University of Technology, China.

In total, 369 online submissions were received from researchers. All papers were reviewed by between 2 and 4 independent experts. Based on the reviewers comments, 127 papers were accepted for publication in this book. All papers were presented either orally, as posters, or online at the event. We believe that this conference will foster cooperation among the organizations and researchers involved in these merging fields.

We would like to thank all contributors for their efforts in submitting their manuscripts on time and we take this opportunity to express our gratitude to the reviewers for their contribution. Thanks are also due to Mr. Maarten Fröhlich of IOS Press for accepting this volume for publication and to his colleagues during the publication process of the volume.

Lakhmi C. Jain, Valentina Emilia Balas, Qun Wu and Fuqian Shi

The ability to fulfil one’s responsibilities both at the workplace and at home while still finding time for interest and hobbies on a personal level is known as work-life balance. There is a need to implement intelligent technologies for shaping work-life balance, as physical, mental, and emotional well-being can all be enhanced by striking a better work-life balance. The need for more transparent and understandable models is growing along with our reliance on intelligent technologies. Furthermore, the current gold standard for establishing credibility and using AI in vital fields is the capacity to explain the model in general. This research seeks to employ XAI to comprehend, illustrate, and elucidate the findings derived from utilizing several machine learning models to determine the primary factors impacting the work-life balance of individuals in the private sector. The support vector machines (SVMs), random forests, and tree-based classifiers from XGBoost were trained and tested on a work-life balance dataset collected from private sector employees. Among all the models, the SVM provides the highest accuracy values for all the features of the feature set compared to the other two algorithms. The attribute importance for the best feature (working style) of the SVM model is explained using the feature importance plot, summary plot, and heatmap from the explainable artificial intelligence models SHAP and LIME. This comprehension makes it easier for decision makers and human resources (HR) practitioners to comprehend the model’s predictions and to implement initiatives that support each employee’s demand for a healthy balance between work and personal life.

In order to solve the problem that text feature mining is insufficient due to the single feature word vector input form at the coding end, which further affects the extraction of text semantic information by the neural network model, the analysis of long text summary generation technology based on deep learning is proposed. This paper introduces the word frequency feature and semantic feature of words, and integrates the TF-IDF value of feature words and the semantic value obtained by LSA through naive Bayesian formula to form a new word vector. It effectively integrates the multi-dimensional features of the text, improves the ability to understand the meaning of words, and helps the encoder to learn the text information. In order to solve the problem that the traditional recurrent neural network has insufficient ability to encode long text and inaccurate acquisition of intermediate semantics, a deep proxy communication mechanism is introduced. The encoder with this structure can obtain global information more accurately and improve the ability to understand text. The experimental results show that each evaluation index has achieved good results. Compared with the benchmark generative summary method point generator+coverage, the automatic summary generation model LTSGDL in this paper has improved about 2.2% on the three summary evaluation indexes.

The global aging population is constantly deepening, especially in first tier cities in China. According to research, the aging population in the central urban areas of first tier cities is higher. Therefore, in order to achieve coordinated socio-economic development, the elderly care service facilities in the central urban area should be reasonably laid out and planned. In this paper, the coverage, resource balance and facility layout rationality of elderly care facilities resources are evaluated through the “Spatial Accessibility” methods. The research results show that it is very necessary to reasonably increase the supply and layout of elderly care service facilities in areas with low coverage of community living circles. We need to further study and evaluate the accessibility of elderly care service facilities through potential models.

In order to solve the problem that high structural redundancy can meet the low accuracy requirements, resulting in a large increase in computing time, the method of intelligent distribution network state estimation based on deep learning is proposed. The structure and input-output properties of single-layer linear neural networks and the IEEE 30 node reliability test system were simulated using the Matlab neural network toolbox. The simulation results show that the single layer linear neural network and the time required is shorter than the bp neural network at the same accuracy, and with the improvement of accuracy, the advantage of computing speed increases significantly. At MSE 0. 4, the single layer linear neural network is 3,814.408s faster than the bp neural network. Single-layer linear neural network is simpler and faster than bp neural network, which can greatly improve the computing speed, and is more suitable for VSSE calculation.

In order to explore a time-saving, labor-saving and efficient and accurate construction monitoring method, the method of project construction progress monitoring based on UAV 3 d modeling is proposed. The coordinates of three-dimensional real scene model and BIM model are converted into the unified coordinate system, and import MicroStation and platform to realize the precise integration of three-dimensional real scene model and BIM model; then based on MicroStation platform, and a construction progress analysis plug-in is developed to realize the matching, segmentation and calculation of three-dimensional real scene model and BIM model, reflecting the actual construction progress with the ratio of completed component volume to the total volume of BIM component, so as to realize the automatic monitoring of project construction progress. The results show that the construction progress of the first to eighth span is higher than 90%; the concrete for the transverse connection of the ninth to fourteenth span girder is not poured, and the construction progress is less than 90%. The construction progress of the first seventh and ninth and tenth main girder on the right side of the bridge is above 97%; the construction progress of the first to seventh main girder on the right side is higher than 95%.

In order to solve the problem of high leakage rate in the existing automatic control system of electric power inspection robot, the automatic system of electric power inspection robot was proposed. The hardware design of the automatic control system of the power inspection robot is embedded Linux design, motor driver design and infrared thermal image thermometer design; The software design includes motor driver initialization, motion control programming, interrupt positioning and calibration programming. Through the design of system hardware and software, the operation of the automatic control system of power inspection robot based on embedded Linux is realized. The experimental results show that the leak detection rate of the designed system is far lower than that of the existing system, and its minimum value can reach 10.25%.

In order to analyze the faults of equipment in the process of operation in time and feed them back to the information system synchronously, and provide decision-making information for the setting of maintenance scheme, the method of intelligent monitoring and fault diagnosis in textile automatic production is put forward. The condition monitoring system of textile machinery can monitor the signal parameters of textile equipment in operation, and then analyze the failure problems faced by textile equipment in the current control mode by comparing the system reference parameters. Based on this, taking vibration signal analysis as the breakthrough point, this paper expounds the condition monitoring signal analysis and fault diagnosis of textile machinery. The results show that the monitoring times are 8 times, and then the actual efficiency of fault diagnosis is obtained by analyzing the data, and the false alarm rate is less than 0.5%.

In order to realize the innovation of the teaching mode of information management and information system profession, and better cultivate high-quality, application-oriented talents, the integration of student management system and education teaching in the background of multidisciplinary integration: strategy and practice. The article researches the teaching reform and practice of information management and information system specialty under the background of multidisciplinary integration. It analyzes the objectives and characteristics of information management and information system professional talents training, and puts forward the relevant measures of information management and information system professional teaching reform on this basis. The results show that after a period of teaching, most of the students in the class have good feedback on this teaching method.

In order to alleviate the rural water environment pollution problem, which restricts the rural ecological environment and hinders the sustainable development of agriculture, the rural sewage treatment system based on smart water services was proposed. The system is developed in B/S mode and adopts distributed component structure, which can be applied to a variety of hardware platforms and operating systems. Each application end obtains on-site data and operation and maintenance conditions through the data cloud platform to realize remote centralized management of sewage treatment stations. The results show that the sewage treatment system has a good removal effect on COD Cr, BOD 5, NH 3 - N, TN, TP, SS and other pollutants in sewage, and the removal rate is more than 80%.

In order to solve the problems of long time, high cost and low efficiency in the lightweight design and optimization process of new automobile frame, the lightweight optimization design method of new automobile frame based on deep learning algorithm was proposed. According to the load and length of the target frame, a mathematical model is established with the minimum weight of the longitudinal beam as the objective function and the boundary conditions, strength, stiffness and stability of the longitudinal beam as the constraints. In MATLAB, the radial basis function neural network algorithm is used to optimize the cross section of the frame longitudinal beam, and the optimal cross section size of the target longitudinal beam is obtained. The finite element models of the new frame 1 obtained from the optimization design and the new frame 2 obtained from the experience design are established respectively, and the working conditions of the two are compared and analyzed by using ANSYS. The experimental results show that frame 1 is 19.5% lighter than frame 2 on the premise of meeting the requirements of frame design and use.

In order to broaden the application of factor analysis, the economic benefit evaluation analysis of tobacco industry based on factor analysis was proposed. Using factor analysis and SPSS statistical software correctly, the economic benefits of 7 cigarette industrial enterprises in a certain place were evaluated quantitatively and qualitatively. The results showed that: using Euclidean distance and class average method, the threshold value was 1.6, and seven cigarette enterprises were divided into four categories: the first category: the second cigarette factory in A city. Category II: Cigarette Factory in City B, Cigarette Factory in City C and Cigarette Factory in City D. Category III: No. 1 Cigarette Factory in City A, and No. 1 Cigarette Factory in City E. Category IV: F City Cigarette Factory. The first cigarette factory in city A has a big difference in comprehensive rankingF_(Summarize) Only ranked 5th; Cigarette Factory in City B, Cigarette Factory in City C, and Cigarette Factory in City D, where comprehensive F is ranked higher in turn, due to asset operation factorsF_“1 ”, capital preservation and enhancement factorsF_“3 ” It is better than No. 1 Cigarette Factory in City A, so it ranks before No. 1 Cigarette Factory in City A.

In order to alleviate the increasing number of vehicles every year and make traffic congestion become more and more serious; an intelligent networked vehicle safety detection mechanism based on deep learning is proposed. Intelligent traffic management can effectively alleviate traffic congestion, and vehicle detection is an important part of intelligent transportation implementation. Traditional vehicle detection methods are inefficient and also rely on manual operation, and have poor robustness. Vehicle detection based on deep learning can solve the above problems well. The results show that with the training, box_loss, obj_loss and cls_loss all approach zero. The model can accurately detect the vehicles on the way, and mark the vehicles with boxes, with a high confidence level between 0.69 and 0.88.

In order to obtain effective information from accumulated data and make awards and punishments for teachers such as appointment, promotion, or bonus increase, and provide convincing basis for this decision, an evaluation model of higher mathematics teaching quality based on improved ID3 algorithm is proposed. This paper introduces the definition and classification of data mining, and introduces the ID3 algorithm of decision tree in detail. According to the ID3 algorithm, a large number of teaching evaluation data samples collected in colleges and universities are analyzed to obtain information gain on different attributes and generate the final decision tree, which can be converted into a set of if then rules. Generate rules and decision trees, and then analyze and predict the new data. The results show that A1 teaching content attribute is the most important in this evaluation system. From here, we can get a fair and objective evaluation. Secondly, teaching methods are also important indicators.

In order to free intangible cultural heritage skills from the traditional mode and actively try new development direction, the application research and experience assessment of intelligent virtual reality technology in museum exhibition is proposed. Firstly, the dilemma faced by intangible cultural heritage is outlined, and the prospect and significance of the application of virtual reality technology in the field of intangible cultural heritage are discussed; secondly, virtual reality technology is utilized to enable the audience to understand intangible cultural heritage in an immersive way through digital display, immersive experience and interactive presentation; lastly, the design idea of digital museum of intangible cultural heritage based on virtual reality technology is proposed, and the Finally, the design idea of digital museum of intangible cultural heritage based on virtual reality technology is proposed, and the design practice and feasibility verification are carried out by taking Fragrant Cloud Veil for example. The results show that: however, the Fragrant Cloud Yarn Museum has certain deficiencies in the publicity of Fragrant Cloud Yarn culture, and the inheritance and innovation of Fragrant Cloud Yarn is facing new challenges, and it is crucial to let more young groups understand Fragrant Cloud Yarn culture.

In order to use computer vision technology to identify the current shooting objects and obtain relevant information, an intelligent identification and navigation system for museum exhibits based on computer vision and deep learning is proposed. System 1 recognizes objects in the museum on the server by uploading image frames via smart phones; System 2 recognizes objects in the area in real time at the smartphone end. The experimental results show that the server-side recognition method takes about 3s, while the mobile side real-time object recognition method only takes about 1s, which can basically achieve real-time recognition without network communication.

This research proposes a scheme of power information system based on quantum communication, aiming at improving the reliability and security of power service transmission. The new algorithm LD-WFQ (improved weighted fair queue) prioritizes the data packets that are about to timeout by estimating the expected time consumption of the data packets to be encrypted, so as to reduce the quantum encryption timeout rate of low priority services. The experimental results show that the LD-WFQ algorithm does not affect the performance of high priority services while reducing the ratio of overtime data packets, and shows good optimization effects under different key request rates. Compared with the traditional WFQ algorithm, the LD-WFQ algorithm has obvious advantages in meeting the requirements of power business encryption and improving the utilization rate of quantum keys, and has practical application significance.

The research and test of the security protection technology of distribution network quantum encryption information based on 5G multi access edge computing applies the communication technology and quantum encryption technology of 5G power private network to the terminal environment of distribution network. Through the integration of quantum key and power service master station, terminal and quantum key, 5G communication and quantum encryption equipment, It enables safe communication between the distribution terminal equipment and the distribution master station, and the system response time is controllable. The results show that in the 5G deployment environment, the maximum delay in 1000 tests is 121.23 ms, while in the 5G+quantum environment, the maximum delay is 200.23 ms.

In order to identify customers with default risk and avoid credit risk, the application of real-time data analysis and predictive maintenance in credit risk management of commercial banks has been proposed. This paper will use CatBoost algorithm to study credit risk of credit card. This paper first preprocesses the data of 24 real-time variables, such as credit line, gender, age, education, marital status, repayment amount, repayment status, and bills payable, and selects 19 of them as the input variables of the model to establish a credit card user credit risk prediction model based on CatBoost algorithm. The results show that the accuracy of CatBoost is 91.73%, which is the highest among the five models, and the accuracy of Logistic is 74.39%, which is the lowest among the five models. Compared with other algorithms, CatBoost algorithm has higher classification accuracy for credit default prediction of credit card users.

In order to solve the constraints on flexibility and scalability of traditional rule-based assessment methods in the comprehensive processing of massive, multi-dimensional and heterogeneous tax-related data for tax risk analysis of large enterprises, an intelligent prediction model for enterprise tax risk is proposed. Aiming at the typical scenarios of tax risk of related transactions of large enterprises, the characterization extraction method of tax data features is defined, and the tax risk analysis and prediction model based on artificial neural network multilayer perceptron is constructed. In order to evaluate the performance of the proposed model, a test dataset is constructed using real tax data and expert labeled data, and experiments are conducted under different positive and negative sample ratios and sample capacity sizes, and compared with several widely used machine learning models. The experimental results show that the model in this paper achieves better performance than the comparative models in both positive and negative sample balanced and unbalanced cases. Among them, when the ratio of positive and negative samples is 5:5, all the models achieve the optimal results relative to the data imbalance, while the model in this paper outperforms all the comparative models, and achieves the best precision rate, recall rate, F1 value and AUC value.

In order to realize the desire of all parties involved in instant retail, it can better understand the behavior of users, discover the hidden interests of users and the behavior rules of group users, and an intelligent recommendation system for instant retail services based on machine learning is proposed. Analyze and process the relevant data of e-commerce websites, use the improved rough set attribute reduction algorithm and Apriori improved algorithm for data mining, use the two improved algorithms together as a scheme to apply to the recommendation system, and then compare the recommendation efficiency of the system through empirical analysis of relevant data. The experimental results show that the time to generate frequent item sets based on the improved Apriori algorithm is significantly less than that based on the Apriori algorithm when the minimum support is fixed. In terms of the efficiency of generating frequent item sets, the Apriori improved algorithm is more efficient. With the minimum support getting smaller and smaller, the comparison becomes more obvious. When the number of transactions is fixed, the time used by the recommendation system to apply Scheme II is obviously less than that used by Scheme I. Moreover, with the continuous increase of the number of transactions in the transaction database, the difference in the time used to apply the two schemes increases significantly. In terms of recommendation efficiency of instant retail intelligent recommendation system, Scheme II is better than Scheme I.

In order to solve the problem that the particle swarm algorithm does not have high search ability in the late iteration and the particles tend to fall into the local optimum when it is introduced into a nonlinear financial risk model (objective), a financial investment risk control model based on an improved particle swarm algorithm is proposed. Based on the optimization of inertia weights and the variation of individual position of each particle, an improved particle swarm algorithm is proposed. The particle swarm algorithm is used to select the optimal control parameters to minimize the total risk value of the financial system. The simulation results show that: in the traditional algorithm, the beginning stage drops faster, and the local optimal phenomenon appears soon, but the improved algorithm makes the particles jump out of the local optimal trap soon by improving the inertia weights and other values, so as to reach the optimal faster, and the convergence of the improved algorithm is advanced more than 5 generations.

In order to effectively manage financial risks, the application of data analysis and data mining technology in financial investment risk management has been proposed. Taking the daily closing price data of Shanghai Stock Exchange Index and Shenzhen Stock Exchange Index as the research object, the marginal distribution is obtained by using the nonparametric kernel distribution estimation function method, the parameters of the commonly used Copula function are estimated by Matlab software, and the Euclidean distance is used as the evaluation index of the Copula model. Then, based on binary normal Copula and t-Copula, we use the new method proposed in this paper to construct a new Copula function. Through comparative analysis, we find that the constructed Copula function can better fit financial data than the commonly used Copula function. Finally, Monte Carlo method is used to calculate the value at risk of the portfolio under different weight values. The results show that when the investment weights are different, the VaR values obtained differ greatly. It can be further found that when the investment weight of the Shanghai Stock Exchange Index is 0.6 and the investment weight of the Shenzhen Stock Exchange Index is 0.4, the VaR value obtained by using these four Copula functions is the smallest, so it can be concluded that the risk of investors selecting this portfolio is the smallest.

In order to solve the problem of typical target recognition, the method of aerial target detection and recognition based on hyperspectral characteristics is proposed. First, the minimum noise separation transform is performed on the hyperspectral image to calculate the intrinsic dimension, and at the same time, the image is denoised. Then, the information divergence is used as the projection index, and the projection index value is adaptively segmented to obtain the spectral curve to be extracted. Finally, the target and its position are identified by spectral angle matching. The results show that the intrinsic dimension of the image is 35 by MNF transform. The projection image is obtained by sequential projection pursuit. Through the adaptive threshold judgment, the value 0.2475 corresponding to the slope of - 1 in the projection index value curve is selected as the threshold to eliminate the background, and then the projection pursuit method is used to extract the end element spectral curve from the projection image. The algorithm in this paper can effectively suppress the influence of background and noise, and can recognize typical targets in the airport more effectively.

This research proposes a video content understanding method based on deep learning to effectively understand video content. First, use ResNet to extract global features and Places365-CNNs migration learning to extract deep scene features. Then, scene vectors are generated by multi-layer perceptron and used as the input of LSTM network to encode and decode video images and description statements. Finally, by pre training on the MSCOCO dataset, accurate description statements are generated for video key frames. The experimental results show that with the increase of data volume, the performance of the model is significantly improved, especially in CIDER-D evaluation, which is superior to other models. The conclusion shows that the proposed model shows high accuracy and performance improvement in describing video content.