Respond by: N/A
The intention of this study is to collect a dataset of how different individual complete Activities of Daily Living (ADLs) in a home environment. By collecting data from a variety of sensing sources (multi-modal), we are able to apply a range of techniques with the intention of enabling: (i) activity recognition, i.e. development of intelligent systems that can autonomously determine activities of a human at home; and (ii) natural activity simulations, i.e. generation of realistic streams of sensor data that mimic a human performing their daily activities.
The role of participants in this study is to execute a prescribed list of activities (ADLs) so that corresponding sensor information can be matched to those activities through annotation. The annotated dataset can then be used as outlined above.
You will attend an assigned session at the Robotic Assisted Living Testbed (RALT), at The Lyell Centre, Heriot-Watt University. The experiment itself should take no more than two hours, unless you choose to participate in additional sessions.
Upon arrival, you will be briefed with information that will help you in exercising your role in the study. Much of this information is also contained in this document, for your reference.
During your session, you will be introduced to the testbed facilities, which seek to mimic a modern one bedroom smart apartment. You will be asked to treat the apartment as if it were your own home, so that we can capture your natural behaviour. You will be provided with a list of activities which you will be expected to complete during your time in the testbed, which you should attempt to follow as exactly as you can. If you make mistakes, please do not worry, just continue on as best as you can.
Data will be collected from around the environment, from the following sources:
You will also be asked to wear an Apple Watch on your dominant hand, which will allow us to collect information about how you move your hands during activities.
Once you have completed all of the prescribed activities, or have opted to end the session, the data from your session will be immediately collated and stored in a local database of participant data for later analysis.
If you wish to take part, please contact the organisers of the study listed on this page.
We are recruiting adults over the age of 18.
Note that the study will involve performing activities in a home environment that has not been modified for those with specific access requirements.
Taking part is entirely voluntary. You may choose not to take part or subsequently cease participation at any time. We will ask you to sign a consent form to show you agreed to take part. You may exercise your right to withdraw at any time by informing a member of the research team.
Provided you are not directly involved in the research (i.e. not a member of CARE Group), you will receive a £20 Amazon voucher for your participation in this study. In electing to participate, your time and effort will contribute to the research effort in improving activity recognition and assistive robotic systems.
Basic demographic information (gender, age, ethnicity and handedness) will be collected prior to the commencement of data collection. Each participant will be given a code known only to the researcher and supervisor. Your demographic information will be seen only by members of the research team.
You can ask for your participant data (i.e. name and demographic information) to be deleted at any time by emailing or contacting one of the people listed below. However, while we can delete our copy of your data from the collected activity dataset and inform known third parties who have republished your data, we cannot guarantee erasure of all copies of the data that may exist elsewhere due to the nature of dissemination of information on the internet.
You should be aware that the activity dataset we collect will include video footage. You are able to chose (in the Consent Form) whether this information is private only to the research team, or can be made publicly available as part of a published ADL dataset. If you choose to be included in the public dataset, the implication is that it is theoretically possible that someone may be able to identify you. Note that we will not publish your name or other demographic information, and in the published dataset you will be known only by your participant number.
Our intention is to create a general purpose and rich dataset for Activities of Daily Living (ADL), to enable other researchers in the field of activity recognition to experiment with and apply different approaches to a variety of sending modalities. Providing unaltered video footage from multiple vantage points allows our dataset to best imitate realistic conditions for a robot in the home. This may allow, for example, the application of computer vision-based strategies for pose detection or gaze detection.
It is entirely optional whether your video recordings are included in the public dataset. You may provide consent and opt wear a suitable face covering during the data recording.
If you withdraw from the study all the information and data collected from you, to date, will be destroyed and your name removed from all the study files.
The results will be reviewed by the researchers and their supervisor and will be used to disseminate the results of the project. Subsequently, written material and datasets associated with the project may proceed to scientific publication in specialist scientific journals. Datasets collected may be made available online, via the university website and/or those of the third-party scientific publications.
The research is organised by persons academically affiliated with Heriot-Watt University, The University of Edinburgh, and the Edinburgh Centre for Robotics (a joint venture between Heriot-Watt University and The University of Edinburgh). Members of the research team are supported by the Engineering and Physical Sciences Research Council (grant EP/L016834/1), EPSRC Centre for Doctoral Training in Robotics and Autonomous Systems.