People-centred Computational Environments for Design Discovery: Establishing a Fully Integrated Framework

随着技术的进步，设计正在成为一项越来越复杂的活动。新的计算技术加上不断增长的计算能力，现在意味着有机会支持几年前还不存在的设计活动。其中一个例子是人工智能和可视化技术的发展。随着这些技术的进步，人们对设计过程和产品使用的期望和看法也发生了变化。在过去，人往往是牺牲品。以大约世纪前的重工业为例。在工具的设计和使用上几乎没有投入时间和精力，因为如果一名工人因设计不当而死亡或受伤，总有另一名工人接替他/她的位置。这种苛刻的态度在大约50-60年前开始改变，人们越来越意识到为人类设计的重要性。这是由计算机的发展和增长，以及需要设计的形式，可以方便和容易地使用，并理解人类的系统界面。一个系统驱动的设计方法的例子是CAD（计算机辅助设计），这是成功地采用了许多项目。然而，CAD是在设计生命周期结束时使用的，此时产品接近最终设计，即它是物理的，有形的和可理解的。早期设计被称为概念设计，在此期间，想法被产生并转化为产品模型。它提供了一种评估设计方案的方法，并允许深入了解设计问题。据估计，80-90%的设计成本是在（概念）设计阶段的前10-20%确定的。因此，概念设计是一个至关重要的阶段，因为在这个阶段可以缩短产品生命周期，降低成本，提高设计质量。概念设计不仅重要，而且非常复杂，因为一个好的设计必须满足多种要求，例如提供合理的结构，美观的外观，灵活的操作，以及创新。因此，利用相对简单的设计模型搜索设计空间花费了相当多的时间。然而，鉴于人类信息处理的局限性，特别是搜索过程和概念设计，更一般地说，将受益于计算系统的发展，这些计算系统在以不确定性和通常定义不好的概念和设计表示为特征的阶段提供更好的支持。限制当前CAD和设计有效性的另一个问题是，它在很大程度上是特定领域的，计算工具通常由最终将使用它的学科的个人设计。这种方法可能会导致“隧道视野”，因为设计人员将使用其领域的知识和工具，并且可能会错过其他领域的相关知识和技术。因此，多学科的方法可以证明在设计过程中是有用的;新的工具和想法不仅可以用来改善最终的设计，而且还可以使一个通用的概念设计系统，可以在不同的学科的发展。它研究了最新的计算智能技术与更成熟的使能计算技术的整合，以在设计的概念阶段提供实用性。这意味着将采取以人为本的办法，确保界面的设计符合用户的期望和能力，同时通过确保用户不断评价系统的产出来减少不确定性和风险。此外，该研究采用了多学科方法，主要学者和工业家共同代表工程设计（民用，航空航天，机械和通信）;药物;生物传感器;软件;虚拟产品;和“可穿戴技术”。该项目是雄心勃勃的，因为所提出的以用户为中心的计算系统的复杂性。然而，通过澳大利亚人权委员会/EPSRC资助的由参与拟议研究的主要学者/工业家组成的小组，已经取得了进展。在过去的12个月里，这个小组举行了会议，并确定了与这一研究领域有关的主要问题。该提案现在寻求12个月的资金，以支持专门的研究工作，以建立在现有的平台上，并详细计划，在发展这种以人为本的概念设计计算方法所需的下一步。

With advances in technology, design is becoming an increasingly complex activity. New computational techniques coupled with increasing computing power now mean that opportunities are available to support design activity which did not exist just a few years ago. One such example is the development of artificial intelligence and visualisation techniques. Coupled with these technologys advancements is a change in people's expectations and perceptions associated with teh design process and the use of products. In the past, the human was often the expendable part. Take, for example, the heavy industries of around a century ago. Little time and effort were applied to the design and use of tools, because if a worker was killed or injured as a result of poor design, there was always another to take his/her place. This harsh attitude began to change some 50-60 years ago with a growing realisation of the importance of designing for the human. This was fuelled by the development and growth in computers, and the need to design interfaces for systems in a form that could readily and easily used, and understood by humans.One example of a system-driven design approach is CAD (computer-aided design) which is successfully employed in numerous projects. However, CAD is employed towards the end of the design lifecycle when a product is close to final design i.e. it is physical, tangible and comprehensible. Early stage design is known as conceptual design, during which, ideas are generated and transformed into a product model. It provides a means for evaluating design alternatives and permit insights into the design problem. It has been estimated that 80-90% of design costs are determined in the first 10-20% of the (conceptual) design phase. Conceptual design is, therefore, a crucial stage as it possible during this stage to shorten the product lifecycle, reduce costs, and improve design quality. Not only is conceptual design important, it is also very complicated as a good design must fulfil multiple requirements relating to, for example, providing a reasonable structure, an appealing appearance, flexible operations, and be innovative. Consequently, considerable time is spent searching through design space utilising relatively simple design models. However, given the limitations of human information processing, the search process in particular and conceptual design more generally would benefit from the developmentof computational systems that provide better support at a stage characterised by uncertainty and often poorly defined concepts and design representations.A further issue that limits the effectiveness of current CAD and design more generally is that it is largely domain specific in the sense that a computational tool would typically be designed by individuals from the discipline that will eventually use it. This approach could result in 'tunnel vision' as the designers will work with knowledge and tools from their field and could potentially miss out on relevant knowledge and techniques from other domains. Consequently, a multidisciplinary approach could prove useful in the design process; new tools and ideas can be used not only to improve the final design, but also to enable the development of a generic conceptual design system that could be utilised within different disciplines.This proposal addresses these shortcomings. It investigates the integration of the latest computational intelligence techniques with more established enabling computational technologies to provide utility during the conceptual stages of design. Implicit in this is that a human-centred approach will be taken to ensure the design of the interface matches the user's expectations and capabilities whilst reducing uncertainty and risk by ensuring constant user evaluation of the system's output. Further, the study takes a multidisciplinary approach with key academics and industrialists working together representing design across engineering (civils, aerospace, mechanical and communications); drugs; biosensors;software;virtual products; and 'wearable technologies'. The project is ambitious because of the complexity of the proposed user-centric computational systems. Progress has already been made, however, via, the AHRC/EPSRC-funded Cluster comprising the key academics/industrialists involved in the proposed study. During the past 12 months this group met and identified primary issues assocaited with this research domain. The proposal now seeks 12 months funding to support dedicated research effort to build on the existing platform, and to plan, in detail, the next steps needed in the development of this people-centered computational approach to conceptual design.