Sporto: A Robot for Self Awareness
FINAL PROJECT STATEMENT
1. Problem Area
2. Design and Technology
3. Evaluation
4. Schedule
In current American society, obesity is one of the biggest social and health problems. In 1999, a staggering 61% of US adults were considered overweight (Flegal et al.). While many people are aware of the importance of exercising, they find it hard to sustain their motivation and quickly revert to their old habits. We plan to build a robot that, by its actions, encourages self-reflection and motivates long-term behavioral changes in its users.
For this human-robot interaction, trust will need to be established between a robot and a person. Humans relate to robots as independent entities, which is different from the way we respond to a computer screen. Since people do not follow messages or advice from people that they don’t trust, people would also not believe or follow messages/advice from robots that they don’t trust.
We would like to explore the human-robot interaction that enhances a feeling of trust between people and robots. A robot’s autonomy and physical embodiment are some of the factors that influence people to view robots as independent entities. We would like to test different combinations of these two factors by measuring how much people trust and follow the advice given by robots.
There are several ways that the robot could show whether or not the user is meeting his or her goal.
1. Morphological manifestation: The robot could actually become fatter, which could be achieved with a servo mechanical solution or by inflation.
2. Emotional expression: A robot could give audible noises or physical expressions that are happy or sad.
3. Suggest activity: Speech, sound, behaviors, or movement could also be used to recommend that a person do a certain activity. For example, when they are exercising enough the robot might sleep or relax, and when they are eating more or have not exercise enough, a robot will run, or walk.
To represent a robot’s behaviors, we could use a simple jointed structure and maybe eyes or lips, but it will vary depending on its design. The robot will be designed to reside where college students spend most of time without moving such as their office area. If the time allows, we would like to explore how this robot in a shared space can encourage positive group interactions to encourage each other’s healthy lifestyle. A robot might connect two people who have sit around for extended periods of time in the same office. Groups of people like sports teams could share a robot to encourage the whole team to work together to reach an exercise goal.
The form of the robot will have an effect on how the user listens to it. The key is to make the robot motivate the person. In order to not depress the user, and compel them to keep up their activity, it will need to be encouraging rather than angry. A fat robot may need to be a cute representation, not an ugly one. A person also might interact differently depending on its representation; people may be less likely to listen to a robot that takes the form of an animal than a robot that takes the form of a famous weight trainer. These are all things that will need to be considered when choosing a design.
Input
We would like to use BodyMedia’s body-monitoring armband to measure a person’s activity. From this information, we would like to sense how much exercise people do everyday, and when they do it. The built-in wireless capabilities of the unit would provide a seamless interface, allowing users to keep the robot updated with minmal effort. If this technology is not available to us, a device using a simple accelerometer could be used to determine activity levels. With a pedometer-based solution, our prototypes would require the user to enter data manually via a web page. A Qwerk would allow us to create a simple web-based interface which informs the robot in near-real-time.
We plan to conduct two evaluation studies: design evaluation, and prototype evaluation.
Design evaluation
During the design evaluation, we plan to conduct informal interviews with 5-10 participants about our multiple design alternatives. This study will be mainly interview based, and we aim to gather which design alternative (i) conveys its message in a non-annoying manner to people, and (ii) is most credible for people. From this information we will choose which design to use for our prototype.
Prototype evaluation
Our test group will consist of the general college student population. This will not only make it easier to find test subjects, but it will allow us to gain more information than if our target user was a more specific group. It will also allow us to see how obese people react to the robot versus physically fit people. If it were only aimed at the obese would come embedded with a negative connotation and cause embarrassment.
We would like to conduct the evaluation study with five participants in our target group for a twelve day period. The first Monday through Friday will measure a person’s current behavior. We will ask each participant to wear the activity monitor for the entire day, except when they sleep, or take a shower. They will need to note when it has been taken off. The robot will then be provided. The Saturday and Sunday will be used to make people to be accustomed to the new design, and remove the “novelty effect.” The final Monday through Friday will be used to gather the real data of people’s response to our robot. This schedule removes the fact that people will be more inclined to be active on weekends since they have more time. It will also allow us to compare a Monday through Friday period without the robot to a Monday through Friday period with the robot.
We will take a qualitative survey to measure:
Quantitative results could include:
Questions will also be asked of the study participants to find out how the robot could be improved.
Week 5-7: User analysis and design exploration
Week 8: Design evaluation
Week 8-11: Prototype building
Week 11 : Prototype evaluation preparation
Week 12-13 : Prototype evaluation
Week 14-15 : Summary report