ArticlePublisher preview available

Data-informed inverse design by product usage information: a review, framework and outlook

To read the full-text of this research, you can request a copy directly from the authors.

Abstract and Figures

A significant body of knowledge exists on inverse problems and extensive research has been conducted on data-driven design in the past decade. This paper provides a comprehensive review of the state-of-the-art methods and practice reported in the literature dealing with many different aspects of data-informed inverse design. By reviewing the origins and common practice of inverse problems in engineering design, the paper presents a closed-loop decision framework of product usage data-informed inverse design. Specifically reviewed areas of focus include data-informed inverse requirement analysis by user generated content, data-informed inverse conceptual design for product innovation, data-informed inverse embodiment design for product families and product platforming, data-informed inverse analysis and optimization in detailed design, along with prevailing techniques for product usage data collection and analytics. The paper also discusses the challenges of data-informed inverse design and the prospects for future research.
This content is subject to copyright. Terms and conditions apply.
Journal of Intelligent Manufacturing (2020) 31:529–552
Data-informed inverse design by product usage information: a review,
framework and outlook
Liang Hou1,2 ·Roger J. Jiao2
Received: 6 September 2018 / Accepted: 16 January 2019 / Published online: 23 January 2019
© Springer Science+Business Media, LLC, part of Springer Nature 2019
A significant body of knowledge exists on inverse problems and extensive research has been conducted on data-driven design
in the past decade. This paper provides a comprehensive review of the state-of-the-art methods and practice reported in
the literature dealing with many different aspects of data-informed inverse design. By reviewing the origins and common
practice of inverse problems in engineering design, the paper presents a closed-loop decision framework of product usage
data-informed inverse design. Specifically reviewed areas of focus include data-informed inverse requirement analysis by
user generated content, data-informed inverse conceptual design for product innovation, data-informed inverse embodiment
design for product families and product platforming, data-informed inverse analysis and optimization in detailed design,
along with prevailing techniques for product usage data collection and analytics. The paper also discusses the challenges of
data-informed inverse design and the prospects for future research.
Keywords Inverse design ·Product usage information ·Data-informed design ·Data analytics ·Cyber-physical systems
With the emerging technologies of product-embedded sen-
sors and information devices, product operating data can be
captured remotely and continuously (Kiritsis et al. 2003;
Huang et al. 2017). The growing digitization of products
and smart sensing technologies covered under the umbrella
of Internet of Things and cyber-physical systems support
the possibility to collect an increasing amount of product
usage information to be accessible for product design teams
(Lützenberger et al. 2016; Kong et al. 2019; Thürer et al.
2019). E-commerce platforms and social media facilitate
companies’ access to the massive user generated data, which
empowers a data-driven approach to continuous design
improvement and next-generation product prediction (Porter
and Heppelmann 2015; Issa et al. 2017). Opresnik et al.
(2013) systematized this data-driven decision process as an
information feedback loop of collecting, storing and ana-
BRoger J. Jiao
1Department of Mechanical and Electrical Engineering,
Xiamen University, Xiamen 361005, China
2School of Mechanical Engineering, Georgia Institute of
Technology, Atlanta, GA 30332-0405, USA
lyzing data from customers and end-users of the products,
with the goal to discover new needs or identify changes in
usage patterns, and in turn to provide information about new
product offerings back to the customers. By exploiting huge,
versatile and highly contextualized product through-life data,
design engineers can harness their organization’s competitive
edge by uncovering patterns, novel insights, and knowl-
edge through data-driven design (Zhao et al. 2007;Lietal.
2019b). While data-driven design makes better informed
decisions possible for developing better products, enormous
and multiplex user- and product-generated data brings about
unprecedented challenges, alongside unmatched opportuni-
ties, for advancing the theory, methods, tools, and practice of
engineering design for products, systems, and services (Kim
et al. 2017; Hyun et al. 2017).
Product design typically entails a forward decision pro-
cess that is technically driven and focused on intrinsic
performances of the products (Yannou et al. 2013). Con-
ventionally a variety of design specifications and system
conditions are either acquired through hypothetical market
studies or assumed a priori based on anecdotal experience,
which inevitably involves subjective judgments and approx-
imation, leading to deviation from true customer satisfaction
(Zhang et al. 2017). Elicitation of customer requirements is
commonly achieved through interviews, focus groups, user
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
... In such a context, (Zheng et al., 2019c) stressed the need for a systematic development process of a Smart PSS dwelling on the two aspects: the "data-driven" paradigm and the "value co-creation" process. The datadriven paradigm promises the shift towards the "informed" (Hou and Jiao, 2020) development process and the value co-creation process is presented "win-win" for co-existing stakeholders (Gaiardelli et al., 2021). From this viewpoint, embracing the possibility enabled by IoT, it becomes of uttermost significance to understand the various opportunities to collect data within the lifecycle of the Smart PSS and to be able to collect data and derive knowledge from its usage phase. ...
... Within engineering design, literature is rich with applications in different stages of the process. Readers are directed to the state-of-the-art review by (Hou and Jiao, 2020) for a broader view of the landscape, where they have reported contributions in requirement analysis, conceptual design, embodiment design and product architecting, and detailed design. To this date, not just supporting design decisions, but data-driven design has found its role in radical innovation of products (Bertoni, 2018), services (Engel and Ebel, 2019), and PSS (Zheng et al., 2018). ...
... There seems to be an agreement among authors concerning the existing categories of operational data based on the source of data: user-generated data and sensed data (Hou and Jiao, 2020;Li et al., 2021;Zheng et al., 2019b). User-generated data is usually the data acquired from online platforms, surveys, etc. ...
Full-text available
The emergence of IoT has propelled the traditionally known Product-Service System (PSS) to be characterized by smarter technologies, enabling them to collect and process data from the operational stage and facilitate communication between the customer and the provider. Commonly referred to as Smart Product-Service Systems (Smart PSS), these systems promise to create value at a personal level by collecting and effectively utilizing the operational data. However, one of the fundamental challenges is the lack of awareness as to what kind of data can be collected from the operational stage and what can be achieved from this data. This paper systematically reviews scientific literature to underline the kind of data being collected from the operational stage, the purposes being achieved from that data, and how they lead to value creation. The systematic review of 60 representative studies enabled the definition of the operational scenario that comprises 4 dimensions of data and 10 classes of data within these dimensions to generically identify what kind of data is being collected. The intend presented by various authors led to the generalization of 5 themes that target different purposes of collecting data. Further, the papers were classified with regards to functional or non-functional requirements to see how data in different approaches are leveraged for value creation. Finally, the discussion highlights the current gaps in the literature and raises several opportunities for future contributions.
... Balasubramanian et al. divide the use phase data of cars into five types: External conditions, technical status of equipment, product usage, personal data and preferences, and direct communications (Balasubramanian et al., 2016). Hou and Jiao distinguish user-generated data, product operating data, and environmental data (Hou and Jiao, 2020). A comparison of the classifications shows that there are overlaps and similarities of classes, e.g., the division into user-oriented, product-oriented, and context-oriented data. ...
... Traditionally, the acquisition of high-quality use phase information has been difficult (Deng et al., 2021); however, due to the transformation of mechatronic products to cyber-physical systems, extensive use phase data from the product generations in the field are now available (Porter and Heppelmann, 2014). These data can be investigated with statistical analysis, data mining, and machine learning methods (Hou and Jiao, 2020;Igba et al., 2015). The results of the data analysis lead to new insights about the product and its users (Meyer and Wiederkehr et al., 2021). ...
Full-text available
The ongoing digitalization of products offers product managers new potentials to plan future product generations based on data from the use phase instead of assumptions. However, product managers often face difficulties in identifying promising opportunities for analyzing use phase data. In this paper, we propose a method for planning the analysis of use phase data in product planning. It leads product managers from the identification of promising investigation needs to the derivation of specific use cases. The application of the method is shown using the example of a manufacturing company.
... (3) The output of the system is represented in a graphical manner. In the end, operator or engineer knows the result, they can utilize the supervised machine learning process to get the probability [22][23][24] The entire implementation system is developed in a flowchart represented in Figure 3. (1) User interface will transfer the data between the user to the system (2) Target was achieved using a reasoning machine and user input ...
... Finally, the user may use a system of certainty to conclude based on the empirical relations. It may add the past information and also possibly update the essential parts [23][24][25]. ...
Full-text available
Integrating machine learning into one of the manufacturing processes, i.e., forging, is mainly concerned with making the system more intelligent by incorporating them to exhibit global understanding. Sometimes the engineer/operator can find the defects during or after the forging operation. So, the system will need some input to identify the different types of categorized defects. And also, according to that, we will develop the intelligent fault diagnosis process. We should calculate the statistical probability theory. Now, we implement the system which is the structure of the fault analysis system for the forging process. In the structure, we demonstrate the defect of the forged part, use the given imported probability to find the possible causes, and provide some remainders to reduce the fault. For enhancement of feature needs, this work includes more integration of AI with forging.
... Data of this type are easier to capture and enable real-time working-stage recognition. The information is mainly composed of two parts: one part is that of, for example, the CAN Bus and mainly involves the characterization value of important attribute parameters of the complex equipment; the other part comes from the unstructured data that can reflect the objective operating state of the product and the personalized environment in the actual working condition [11,12]. Data such as the inertial measurement unit (IMU) sensor of a smartphone and the engine torque and fuel usage in the CAN Bus can be used [8,13]. ...
Full-text available
With the maturity of sensor and data acquisition technology, the intelligent development of multi-sensor integrated loader becomes inevitable. This paper focuses on the realization of intelligent recognition of loader's working stage via low-cost and efficient sensor configuration of bulk operation data. A feature selection method, Redundancy-Complementariness Dispersion-and-Relevance-based (RCDR) is introduced to select the optimal configuration with fewer sensors. By comparing different combinations of window size and various classifiers, it is found that the sensor set configured based on RCDR feature selection can achieve an accuracy of 94.17% in working-stages recognition. Arguably, the method is potent in configuring a subset of sensors with fewer sensors and accurately recognizing working stages in various types of low-cost operation data without introducing an intelligent calibration system (IF-Then strategy). Future research is expected to tackle the limited applicability of the model caused by data discontinuity, window size combination difference and the change of loader type.
... In this study, the key issue is the efficient optimization of the pressure hull, because the FE simulation, especially the nonlinear buckling analysis, can be computationally expensive in practical engineering. To reduce the computational cost, the data-driven approach has been increasingly adopted in structural optimization in recent years, which suggests itself to be a new way of inverse design for product innovation (Hou and Jiao, 2020). For example, F. presented a data-driven approach for designing hierarchical hybrid structures with multiple lattice configurations. ...
Full-text available
The design of the pressure hull, one of the most important components, has great significance for improving the performance of autonomous underwater vehicles (AUVs). In this paper, a novel ring-stiffened arched cylindrical (RAC) hull is presented based on the traditional ring-stiffened cylindrical hull, which has the advantages of smooth outline, large compressibility, and small buoyancy factor. Then, a data-driven approach for efficiently designing the pressure hull with an irregular structure is proposed based on the finite element simulation, and its framework is introduced in detail. With the data-driven approach, the optimization design of the RAC structure for AUV is performed, which considers the effect of temperature variation on the deformation. The optimization results indicate that the mass of the pressure hull can be reduced by 10.3% by replacing the traditional ring-stiffened cylindrical hull with the RAC hull proposed in this paper. Finally, some pressure tests are carried out to verify the proposed approach, which is also appropriate for designing irregular structures in other engineering fields.
... e query terms have been accordingly processed in the second module to form the set of RDF triples of the query term vocabulary. Since semantic similarity matching is between RDF triples, this study uses a trans-based algorithm to achieve semantic matching between RDF triples with similarity values above a threshold value, and the retrieved library document with the correlation is used as the retrieval result and returned to the user [18]. Since the retrieval results are not unique in most cases, it is necessary to rank the retrieval results according to the relevance, and the system uses the transalgorithm to calculate the semantic similarity value between RDF triples as the basis for the relevance, and the larger the semantic similarity value is, the greater the relevance is, and the higher the corresponding retrieval results are ranked. ...
Full-text available
This study presents an in-depth study and analysis of IoT semantic association and decision-making using a partial differential fuzzy unsupervised approach. It focuses on a semantic annotation framework for device metadata and a knowledge base construction method to further improve the interoperability of IoT domain knowledge by building a unified IoT domain knowledge base and designing and implementing a semantic IoT knowledge management and application generation system. The main proposal is an IoT generic domain ontology, which reuses the existing excellent ontologies of IoT as much as possible, extracts the commonly used concepts of the domain and combines them, and provides a unified semantic template for IoT applications. On the other hand, by applying the entity linking technique to the extension of the knowledge base and linking the structured metadata of devices to the corresponding entities of the background knowledge base, the domain knowledge base can be made to share the rich background knowledge. At the same time, the interoperability of heterogeneous IoT metadata between applications is enhanced by unifying data and concepts from different device applications to the same background knowledge base through entity alignment techniques. The semantic representation of events applicable to IoT application scenarios is investigated, and an IoT event ontology for representing abstract events and event relationships in IoT is designed; next, a domain ontology with IoT sensing and control event representation capability is constructed based on the IoT event ontology, in which the typical domain ontology (SSN) that can be used for IoT applications is followed by the ontology reuse principle is improved and extended to support the description of event types and interevent relationships, and the IoT event model is associated with the improved IoT base ontology through an ontology alignment approach. Finally, the IoT sensing and control ontology are validated by semantic modeling of device composition, component relationships, and operational processes based on the IoT sensing and control ontology.
Industrial Product-Service Systems (IPS²) have been increasingly researched and practiced in a variety of fields and sectors. Nevertheless, such solutions are not as significantly implemented as expected by the CIRP keynote in 2010: it is not a dominant business of manufacturing companies today despite their economic and environmental advantages. One reason for this delay is assumed to be that the initial efforts were mainly on business and service when design was an afterthought. The promising digital technologies that have become easily implementable in practice will in the next years facilitate design and implementation of IPS² smartly to satisfy users and contribute to sustainability. This keynote analyses literature and practice in the last decade, aiming to propose the main characteristics of IPS² of the future and design processes adequate for IPS². The design process of both production systems for sustainability and high-value systems for sustainable development goals may need to be supported differently than today. The list of recommendations for future research on IPS² design is provided at the end of the paper.
Product design needs to take full account of requirements from the lifecycle of product. However, researcher considered more requirements from product manufacturing/assemble stage, and less requirements from product operation stage. Identifying subsystems to be improved and risk analysis is two methods applied to product design considering the effect of failure at product operation stage. But these methods are labor-intensive and time-consuming tasks and need a lot of expert knowledge. There are huge amounts of data including inspection reports and maintenance records in product operation stage. These data can offer meaningful feedback on next-generation product design. The past few years, digital twin has gained considerable attention because it is characterized by two-way interactions between the digital and physical worlds. This paper proposes a digital twin (DT) framework for product to-be-designed analysis based on operation data. DT data is collected and store from the operation stage of current product. DT data is processed and analyzed to acquire the operation knowledge by Clustering algorithm, Frequent Pattern-growth algorithm, Multi-attribute decision-making, etc. The failure knowledge is used for the design of next-generation product. It can help designer to identify subsystems to be improved and offer the failure information for risk analysis. The failure knowledge also predicts the easily failure subsystems at operation stage of current product. A case of Tunnel Boring Machine is given to illustrate the implementation process of the framework and realize the part function of the framework.
Full-text available
The successful planning of future product generations requires reliable insights into the actual products’ problems and potentials for improvement. A valuable source for these insights is the product use phase. In practice, product planners are often forced to work with assumptions and speculations as insights from the use phase are insufficiently identified and documented. A new opportunity to address this problem arises from the ongoing digitalization that enables products to generate and collect data during their utilization. Analyzing these data could enable their manufacturers to generate and exploit insights concerning product performance and user behavior, revealing problems and potentials for improvement. However, research on analyzing use phase data in product planning of manufacturing companies is scarce. Therefore, we conducted an exploratory interview study with decision-makers of eight manufacturing companies. The result of this paper is a detailed description of the potentials and challenges that the interviewees associated with analyzing use phase data in product planning. The potentials explain the intended purpose and generic application examples. The challenges concern the products, the data, the customers, the implementation, and the employees. By gathering the potentials and challenges through expert interviews, our study structures the topic from the perspective of the potential users and shows the needs for future research.
Full-text available
Reducing costs and increasing equipment availability (uptime) are among the main goals of industrial ventures. Well defined interval durations between maintenance inspections provide major support in achieving these targets. However, in order to establish the best interval length, process behavior, cycle times and related costs must be clearly known, and future estimates for these parameters must be established. This paper applies process mining techniques in developing a probabilistic model in Bayesian Networks integrated to predictive models. The probability of a given activity occurring in the probabilistic model output establishes the forecast boundaries for predictive models, responsible for estimating process cycle times. Availability (uptime) and cost functions are mathematically defined and an iterative process is performed in the length of intervals between maintenance inspections until the time and costs wasted are minimized and the best interval duration is found. The probabilistic model enables simulating changes in the event occurrence probability, allowing a number of different scenarios to be visualized and providing better support to managers in scheduling maintenance activities. The results show that production losses can be further reduced through optimally defined intervals between maintenance inspections. © 2018 Springer Science+Business Media, LLC, part of Springer Nature
Full-text available
The rapid development of information and communication technology (ICT) enables a promising market of information densely product, i.e. smart, connected product (SCP), and also changes the way of user-designer interaction in the product development process. For SCP, massive data generated by users drives its design innovation and somehow determines its final success. Nevertheless, most existing works only look at the new functionalities or values that are derived in the one-way communication by introducing novel data analytics methods. Few work discusses about an effective and systematic approach to enable individual user innovation in such context, i.e. co-development process, which sets the fundamental basis of the prevailing concept of data-driven design. Aiming to fill this gap, this paper proposes a generic data-driven cyber-physical approach for personalised SCP co-development in a cloud-based environment. A novel concept of smart, connected, open architecture product (SCOAP) is hence introduced with a generic cyber-physical model established in a cloud-based environment, of which the interaction processes are enabled by co-development toolkits with smartness and connectedness. Both the personalized SCP modelling method and the establishment of its cyber-physical product model are described in details. To further demonstrate the proposed approach, a case study of a smart wearable device (i.e. i-BRE respiratory mask) development process is given with general discussions.
Full-text available
The Industry 4.0 program and corresponding international initiatives continue to transform the industrial workforce and their work. The service-oriented, customer-centric and demand-driven production is pushing forward the progress of industrial automation. Even though, it does not mean that human can be fully replaced by machines/robots. There is an increasing awareness that human presence is not only one type of manufacturing capability, but also contributes to the overall system’s fault tolerant. How to achieve the seamless integration between human and machines/robots and harness human’s full potential is a critical issue for the success of Industry 4.0. In this research, a human-centric empowering technology: industrial wearable system is proposed. The aim of this system is to establish a human–cyber–physical symbiosis to support real time, trusting, and dynamic interaction among operators, machines and production systems. In order to design a substantial framework, three world-leading R&D groups in this field are investigated. Five design considerations have been identified from real-life pilot projects. The future trends and research opportunities also show great promise of industrial wearable system in the next generation of manufacturing.
Full-text available
Product design is greatly influenced by product configuration processes and can be suspended or result in failure if the configuration process consumes too much time, cost, or resources; such results can also occur if the end products manufactured based on configurations failed to satisfy customers. Therefore, a configuration approach that saves time, cost, and resources, as well as highly satisfies customers, is necessary and significant. Against the background, this study proposes a configuration approach that uses online data to map customer requirements into product configurations, including the product transaction data and customer review data. The approach generates feasible configurations initially by using transaction data. Next, the approach produces training samples based on positive customer review data. Lastly, the intelligent classifier is trained by the training samples and is utilized to select final configurations from feasible configurations to satisfy customer requirements. A real-world design case of smartphones is used to illustrate the proposed approach, and the results indicate that this approach saves time, cost, and resources and is competitive compared with other product configuration methods. This novel configuration approach provides designers and companies with a superior and efficient method to complete configuration tasks with competiveness and low risk and adds value to the usability and analysis of online data.
Full-text available
With the advent of new generation information technologies in industry and product design, the big data-driven product design era has arrived. However, the big data-driven product design mainly places emphasis on the analysis of physical data rather than the virtual models, in other words, the convergence between product physical and virtual space is usually absent. Digital twin, a new emerging and fast growing technology which connects the physical and virtual world, has attracted much attention worldwide recently. This paper presents a new method for product design based on the digital twin approach. The development of product design is briefly introduced first. The framework of digital twin-driven product design (DTPD) is then proposed and analysed. A case is presented to illustrate the application of the proposed DTPD method.
Full-text available
As a cost saving and profit-making strategy, a modular design is being employed in developing complex products and systems (CoPS) in recent decades. At the early stage of design, the reliability of a product can be improved by identifying the influential function modules based on the modular function architecture. In this study, the weighted LeaderRank algorithm and susceptible-infected-recovered (SIR) model of weighted and directed complex networks (WDCNs) are employed to identify the influential function modules of modular CoPS at the conceptual design stage. First, the structure of the function module is obtained and is mapped into a WDCN. Second, based on the similarity between the behaviors of nodes in the WDCN and function modules in the CoPS, a node-identification approach based on the weighted LeaderRank algorithm is employed to identify the influential function modules, whose influences are then verified through the SIR model. The influential function modules of a modular large tonnage crawler crane are determined as a case study to demonstrate the effectiveness and validity of the developed method.
This work investigates the application of the inverse analysis to an illumination design of a three-dimensional rectangular enclosure. The problem consists of finding the luminous fluxes on the light source elements, located on the top of the enclosure, that satisfies a prescribed uniform luminous flux on the design surface, located on the bottom surface. The solution assumes that all the surfaces emit and reflect diffusely, and that the hemispheric spectral emissivities are wavelength independent in the visible region of the spectrum. The inverse analysis is described by a system of linear equations that is expected to be ill-conditioned since it involves the solution of a Fredholm integral equation of the first kind. To tackle the ill-conditioned system of equations, the TSVD (Truncated Singular Value Decomposition) regularization method is applied. In addition to presenting a methodology to solve for the luminous exchanges in a enclosure, starting from a thermal radiation energy balance, this work considers two design cases: one in which the light source elements cover the entire top surface; and one in which a reduced number of light sources are considered, a more practical solution. In both cases, the proposed inverse design is capable of providing a solution that satisfies the prescribed luminous flux on the design surface within average and maximum errors less than 1.0 % and 5.0 %, respectively.
In this paper, based on more than ten years' studies on this dedicated research thrust, a comprehensive review concerning information mining from big consumer opinion data in order to assist product design is presented. First, the research background and the essential terminologies regarding online consumer opinion data are introduced. Next, studies concerning information extraction and information utilization of big consumer opinion data for product design are reviewed. Studies on information extraction of big consumer opinion data are explained from various perspectives, including data acquisition, opinion target recognition, feature identification and sentiment analysis, opinion summarization and sampling, etc. Reviews on information utilization of big consumer opinion data for product design are explored in terms of how to extract critical customer needs from big consumer opinion data, how to connect the voice of the customers with product design, how to make effective comparisons and reasonable ranking on similar products, how to identify ever-evolving customer concerns efficiently, and so on. Furthermore, significant and practical aspects of research trends are highlighted for future studies. This survey will facilitate researchers and practitioners to understand the latest development of relevant studies and applications centered on how big consumer opinion data can be processed, analyzed, and exploited in aiding product design.
In this article, we study the balancing principle for Tikhonov regularization in Hilbert scales for deterministic and statistical nonlinear inverse problems. While the rates of convergence in deterministic setting is order optimal, they prove to be order optimal up to a logarithmic term in the stochastic framework. The two-step approach allows us to consider a data-driven algorithm in a general error model for which an exponential behaviour of the tail of the estimator chosen in the first step is valid. Finally, we compute the overall rate of convergence for a Hammerstein operator equation and for a parameter identification problem. Moreover, we illustrate these rates for the last application after we study some large sample properties of the local polynomial estimator in a general stochastic framework.