Xavier Fischer

Interactive Design: Then and Now

An obvious fact of the present industrial epoch is the advancement of computing techniques that has strongly contributed to the emergence of engineering processes becoming more and more efficient, rapid and high-powerful. Computing Science has allowed efficient virtual tools to emerge. Nevertheless, today all industrial experts share the idea that computational tools do not represent a sufficient solutions for leading to innovation. Some strong transfers between Industry and Research have been developped in order to make around the virtual tools new techniques that lead to whole and complete solutions really ensuring a support to decision making in product engineering.

Identification of design variables and criterion variables towards the integration of user requirements into preliminary design

Industrial designers have to design products that meet User Requirements. User Requirements are usually used as validation elements at the end of the Product Design process. Yet, at this stage in the Product Design process, many decisions have already been made. Decisions made by Designers in the Preliminary Design phase account for 70% of total product development costs. Therefore, iterating a Design Process is very expensive. The object of the present approach is to integrate User Requirements at the beginning of Preliminary Design, using functional specifications. The general object of the present approach is to find links between a Sensory Characterisation and a Functional Characterisation of a product. Using such links, Designers will be able to evaluate the influence of design decisions on User sensations. User requirements can play a major role as physical or financial aspects in Preliminary Design.

Model reduction technique for mechanical behaviour modelling: Efficiency criteria and validity domain assessment

The current paper presents the study of a neural network-based technique used to create fast, reduced, non-linear behavioural models. The studied approach is the use of artificial neural networks (ANNs) as a model reduction technique to create more efficient models, mostly in terms of computational speed. The test case is the deformation of a cantilever beam under large deflections (geometrical non-linearity). A reduced model is created by means of a multi-layer feed-forward neural network, a type of ANN reported as ‘universal approximator’ in the literature. Then it is compared with two finite-element models: linear (inaccurate for large deflections but fast) and non-linear (accurate but slow). Under large displacements, the reduced model approximates well the non-linear model while having similar speed to the linear model. Unfortunately, the resulting model presents a shortening of its validity domain, as being incapable of approximating the deformed configuration of the cantilever beam under small displacements. In other words, the ANN-based model provides a very good compromise between accuracy and speed within its validity domain, despite the low fidelity presented: accurate for large displacements but inaccurate for small displacements.

Multidisciplinary design process based on virtual prototyping for microsystem design

Microsystem design involves the mixing of two main disciplines: mechanical engineering and electronics. Multidisciplinary design approach is not easy to implement and becomes more difficult when the challenge consists in simultaneously handling new knowledge that characterizes the field of microsystem industry. We present herewith a new multidisciplinary engineering approach used to lead the detailed design of micro-accelerometers based on virtual tools. Virtual prototyping that is usually implemented in one sole domain of competence is developed to be manipulated by either experts from electronics or mechanics. This new tool improves collaboration and the rapid identification of highly satisfying solutions. The virtual tool allows exploration of micromechanical devices and system design issues and objectives. It uses interactions with the designer in order to achieve solutions integrating the physics rules of the design and the experience of the designer.

Qualitative Constraints in Integrated Design

Designers use mechanical calculus and their technical knowledge to give, propose and validate technical choices. Product perception and design reasoning are modeled using graphs. The knowledge used is defined through fuzzy-logic rules, qualitative analysis, analytical forms and reasoning is represented as a system of constraints. We give an example which aims at defining the technological choices of a lattice work of beams.

Conception Inversée Intégrée : prédimensionnement de produits par satisfaction de contraintes

Mechanical design is based on aims of schedule of conditions. Design is an inverted integrated process: designers use aims to do design choices. This process, which requires a translation of all objectives, allows the synthesis of ideal dimensions (choices of materials, dimensions and shapes) from explicit and implicit rules. Rules translate all process laws and mechanical laws: mechanical calculus is integrated in design. Inverted Integrated Design (IID) process is modelized from a constraint based representation and a qualitative valuation. Constraints are analytical or are build from neural networks. Consistency techniques are used in CSPs to seek all consistent dimensions. We apply our work to the design of some pressure apparatus.

Component based modeling for structuring measurement Preliminary modeling applied to automotive braking behavior

This paper exposes a methodological approach that intends to lead later to an interactive braking system simulation model. Our Component Based Modeling (CBM) method begins by the product definition and the analysis of significant moments to determine the actions of the driver and the interactions produced between the components of the braking system, the driver and the environment. All our approach is based on the concept of product lifecycle. Development, qualification and updating of the automotive braking system model are going to be done thanks to instrumentation with sensors and micro-sensors. These elements have the advantage of being easily integrable and non intrusive.

Granular modelling for virtual prototyping in interactive design

This paper introduces the concept of granular modelling as a strategy to construct human-centred virtual prototype and environment models. This approach aims at providing a granular strategy for virtual prototype models in order to allow the change from a coarse-grain description to a fine-grain one according to the level of specification desired. The granular model is built from the aggregation of several component models and their interaction models, representing the interaction between two components. Component models must include information not only about themselves but also about the environmental part related to them. This is important as environmental conditions may have different effects at different scales/levels. Additionally, it leads to a higher level of communication, only allowed through the interaction models. This implies that new different objects might be integrated only by including their component models and their required interaction. The foreseen advantage is that the human is integrated in a simple manner: a component model corresponding to a haptic interface and an interaction model relating it to the virtual product. The first step, identification and organic description, consists of establishing an interaction graph from the hierarchical structure of the virtual product and all the objects that will be present in its future environment. Subsequently, the component and interactions models are constructed. The model is expected to be implemented into a multi-agent system where every component model derives into a component agent.

A Tutor Agent for supporting distributed knowledge modelling in interactive product design

Currently, the process of problem analysis in product design involves multiple partners in a distributed environment. A tutoring approach to support knowledge modelling will be presented in this article, focusing on variables as a key component of knowledge. A set of Tutor Agents intends to aid distributed partners to coherently elicitate their know-how in order to use to analyse design problems. A granular Knowledge Base is proposed to standardise distributed knowledge, providing Product Life-Cycle (PLC) constraints (based on product knowledge) to construct optimisation models. Furthermore, a case will be presented in order to clarify the approach. The results will enable design teams to integrate constraints from PLC knowledge in order to optimise the evolution of design concepts.

Integrated Design and Manufacturing in Mechanical Engineering

Automated concept generation is non-trivial task. The complexity of this problem is mainly due to lack of formal representation frameworks that lend themselves easily to a computational approach. Generative grammar has emerged as a potential solution to this problem and presents a number of different possibilities for conceptual design automation. A novel search method is presented: it has been developed specifically for search trees defined by a special class of generative grammar in which rules of the grammar have parameters associated with them. A novel feature of the proposed search is Human in the loop approach in which learning about the search space is achieved by querying the user. The user fatigue restricts the maximum number of comparisons of candidate solutions (30–50). Prom the data gathered from the comparisons, a stochastic decision making process proposed in this paper quickly converges to a region of design space which best meet the users preference. The method is implemented and applied to a grammar for shampoo bottle concept generation. It is shown through multiple user-guided and automated experiments that the method has ability to learn and adopt through human computer interaction process.