Real Time Sensing & Product Ethonography in Manufacturing
Current limitation in industry
Designing products relies on a deep understanding of how people make use of them. Traditionally, this happens through design research, which is costly and labour intensive. When products go out into the world, it is hard to know what has happened to them – without collecting products at end of life, or carrying out extensive customer surveys, it is difficult to get good feedback, and iterating on the designs can take years.
Emerging smart devices are starting to build up an understanding of their environments – for example, Roomba vacuums map out the houses that they clean. Similarly, connected IoT devices such as Nest thermostats build up a picture of their users, in order to provide a personalised experience and save on energy bills. However, this intelligence is focussed on the way these devices operate, supporting their function of improving their environments. This is distinct from using the data to improve the devices themselves. Currently, a real-time understanding of product performance is limited to complex, single function products such as Rolls Royce jet engines. Low cost sensing and data analysis allows us to gain insights from a wider range of domestic and consumer devices, giving immediate feedback and supporting rapid design iteration.
Current limitation in scope of research
Capturing context is a key challenge in understanding human behaviour and interactions with objects. Laboratory studies do not capture daily practices, so research ‘in the wild’ is needed. However, this means that the important moments we would like to understand are likely happen far away from researchers, so we need better techniques to understand what took place. If someone drops a product, how were they holding it? What were they trying to do? What happened just before and just after the drop? Modern low-cost sensors can help to capture enough context around events to provide answers to these questions, supporting rich design insights.
The Value added to research
In order to make this happen, we need to develop a science of continuous ethnography that looks at sensor information and relates it to produce usage. This means:
• Understanding how to do activity and novelty detection in the wild with diverse sensors and objects.
• Creating methodologies for relating sensor data to social and individual practices at scale, over long periods of time
• Techniques for selective data capture allow research and design questions to be refined post-deployment.
The Value added to industry
This part of Chatty Factories will look at how this information can be rapidly fed back into the design of the objects themselves. There are several challenges, which will individually support different parts of the design and manufacture process, and lead towards the development of “use driven design”, where things are refined based on the ways in which they are used in practice. By looking at the entire lifecycle, from manufacture, through to shipping, salesrooms, delivery, use and end-of-life, we will add value in the following ways:
• Adding sensors to everyday objects, lets us build up a real-time view of the way that they are used, and capture subtle details about what happens and when. This will give manufacturers a greater insight into the way that their products perform. In contrast to traditional studies, we will look at the objects, not the people, which will help to reduce concerns about privacy and surveillance.
• We will add intelligence to the objects we are working with. On-device processing can make sense of data as it comes in, reducing the amount of data that is shared. Processed data can be simpler and less invasive. This will allow data to be collected in more places, and shared more easily. Looking at multiple sources of data together will allow us to build up a more comprehensive picture of the contexts in which things are used, allowing manufacturers to see how they fit into different environments.
• We will then develop ways to make use of this data in the design process, allowing product designers to quickly and efficiently respond to data as it comes in, creating a more responsive product development cycle.
Visualising the future
Overall, our value add is to provide a rich understanding of the ways that things are used, where artificial intelligence is combined with observation, providing timely data that shapes the evolution of products from the start of manufacture through to the end of life.
• Continuous digital ethnography helps us understand product use at scale.
• Intelligent objects reduce data at source, and can make data collection less invasive.
• Relating data to use supports a dynamic, use-driven design process.