A Robot For Health
Human-robot interaction is now well enough understood to allow us to build useful sociable robot systems that can function outside of the laboratory. This is the first project to develop and deploy a sociable robot system to investigate long-term human-robot interaction in people’s homes in the context of helping people with their behavior change goals (see research page). Specifically, the sociable robot system is designed assist people who are trying to lose or maintain weight. We selected this application domain because it supports a long-term study where the creation of such a system might make a practical difference. To develop this application, we collaborated with Dr. Caroline Apovian at The Nutrition and Weight Management Center at Boston Medical Center.
Overweight and obesity are currently significant problems in the United States. They are increasing and estimates of the cost to the US economy for obesity-related health problems range from US 75 to 125 billion dollars per year as of 2004. In the United States, the National Center for Health Statistics at the Centers for Disease Control and Prevention report that 65% of the adult population is overweight or obese (31% obese and 34% overweight, calculated using the body mass index, or BMI). According to the World Health Organization, this is an international problem, with over 1 billion of the world’s adult population overweight, with 300 million of these considered obese, and they state that almost all countries (high-income and low-income alike) are experiencing an “obesity epidemic”. It is also known that of those who do lose weight, 90 to 95% are unable to keep the weight off long-term. The length of time looked at in studies of weight regain varies, but is commonly between six months and one year.
Obesity is a problem that is being addressed in many ways. Most current work relates to new pharmaceuticals, diets, bariatric surgery, or other treatment regimes. Some clinical research in the last decade has looked at new isnterventions that are focused on creating behavior change and leading to long-term weight loss. The design and study of novel and potentially more successful means to effect behavior change is a promising direction in recent bariatrics work. Some of these include internet-based interventions using either text- or character-based interfaces.
Our earlier results in controlled Human-Robot Interaction studies have shown that a robot can be seen as more credible and informative than a character on the screen. Hence, there is reason to believe that a robot may be a more effective mechanism for conveying the behavior change message. Results showing that a robot can be more engaging than an animated character lends itself to the possibility of creating a set of longitudinal interactions, or a relationship, that is longer-lasting than previous techniques and therefore also more likely to have the opportunity to create long-term behavior change.
The website for Cory Kidd’s company, Intuitive Automata Inc., is now live and will soon make a commercial version of the weight loss robot.
One aspect of social robots that we are concerned with in HRI studies pertains to how people perceive them. For instance, when a robot is designed to be depended on by a person for completing a task, it must be seen as trustworthy. This is not a feature that we know how to turn off or on, so it likely has to do with how the robot interacts with a person. In interactions where a person is relying on a robot for information, they must believe that the information is credible. Based on earlier research, this is one feature that could be affected by the presence or proximity of the robot. If information is being conveyed in a situation where a robot is teaching a person, then the robot must be capable of engaging the person accordingly. As with other aspects of social robots, there are many issues that affect engagement, as a number of studies have shown.
Social Presence in Animated Verses Robotic Characters
What we discovered is that a robot is an effective partner in an interaction because of its physical embodiment. Through our pair of experiments, we found that a robot is seen as more engaging than an animated character and is perceived as more credible and informative, as well as being more enjoyable to interact with.
Social Presence in Co-present Verses Remote robots
Overall, we see these studies as establishing some of the basic design parameters for robots that are intended to interact with humans in social situations. There remains much more work to be done before we have determined most of the parameters that will create a successful interaction, but the results presented here allow us to begin creating the kinds of interactions that will move human-robot interaction out of the lab in the near future.
Our sociable robot system addresses a main shortcoming of existing weight loss programs, that of long-term adherence. Combining what we have learned during several years of research in human-robot interaction with current weight loss and weight management techniques, we have constructed a system that allows people to manage weight-related data and interact with a robotic coach.
A relationship model has been developed to appropriately create and manage the relationship between the coach and the user. Based on what is known about human relationships, human-robot interaction, and relationships with agents, the model allows the user’s interactions with the coach to evolve over time, allowing the system to establish a pattern of engagement that attempts to keep the user using the system, and therefore maintaining their weight loss and maintenance behaviors, longer than they otherwise would.
Much like a human coach, the robotic coach attempts to become a part of the person’s social support network. By integrating into this existing network, this system can take advantage of the known benefits of social support in weight loss and weight maintenance. By helping a user keep track of data relevant to their weight loss program, the user has the option of sharing the data with family, friends, or caregivers as a way to gain their support in encouraging progress.
The coach offers feedback on recent behavior and makes recommend
ations for near-term behavior. The feedback is based on comparing recent diet-related behavior, such as calories consumed and exercise performed, with goals set by the user. Recommendations come from general information on diet, nutrition, and exercise and are tailored to the individual based on the current stage of the relationship between the coach and the user.
The system maintains a database that keeps track of interactions with the user, information gathered from the user, and goals set by the user. This is used in the relationship model and for the feedback to determine how each interaction should occur.
The sociable robot system consists of an interactive robot coach and a computer that maintains the necessary information. The robot is able to orient its head and eyes to perform simple non-verbal communication cues such as mutual gaze and joint attention. It also has a synthesized voice in addition to displaying text on a screen. Earlier work we conducted in human-robot interaction has shown that a robot can be more engaging than a character on the screen, which leads to our using a physical robot in this system rather than an agent on a PDA, phone, or computer. Other pieces of the system will allow automation and simplification of the system for users.
In a controlled long-term evaluation study, we compared the robotic coach to a standalone desktop computer running the same software and to a traditional paper log. A current challenge in weight management is not losing weight but rather in getting individuals to keep off weight that is lost. The results of our study show that participants track their calorie consumption and exercise for
nearly twice as long when using the robot than with the other two interventions and develop
a closer relationship with the robot. Both of these are indicators of longer-term success at weight loss and maintenance.
The results from our study support the following hypotheses:
- Participants interacting with the robot will use the system for an overall longer period of time than participants who have a computer or a paper log.
- Participants interacting with the robot will rate the system higher on regular responses to the short version of the working alliance inventory (WAI-SF) than participants who have a computer.
- Participants interacting with the robot will rate the system higher on responses to the full version of the working alliance inventory (WAI) administered at the end of the experiment than participants who have a computer or participants using paper logs.
- Participants interacting with the robot will rate the system higher on responses to a scale of trust administered at the end of the experiment than participants who have a computer or participants using paper logs.
- Participants interacting with the robot will rate the system higher on responses to a scale of engagement administered at the end of the experiment than participants who have a computer or participants using paper logs.
- Participants interacting with the robot will rate the system higher on responses to a scale of reliability administered at the end of the experiment than participants who have a computer or participants using paper logs.
- Participants interacting with the robot will have a closer bond by the end of the experiment than participants who have a computer or participants using paper logs.
- There will be no difference in the amount of weight lost across all three groups at the end of the six week study period.
Cory Kidd, Designing for Long-Term Human-Robot Interaction and Application to Weight Loss. January 2008. Ph.D. Media Arts and Sciences, MIT. [PDF]
- C. Kidd and C. Breazeal (2008). “Robots at Home: Understanding Long-Term Human-Robot Interaction”. Proceedings of the 2008 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2008). Nice, France. [PDF]
- Cory D. Kidd and Cynthia Breazeal (2007). A Robotic Weight Loss Coach. Twenty-Second Conference on Artificial Intelligence. Vancouver, British Columbia, Canada. (AAAI ’07) [PDF]
- Kidd, C. & Breazeal, C. (2006) “Designing a Sociable Robot System for Weight Maintenance.” In Proceedings of IEEE Consumer Communications and Networking Conference (CCNC-06). Las Vegas, NV. Volume 1, 253-257. [PDF]
- Kidd, C. and Breazeal, C. (2005) “Sociable Robot Systems for Real-World Problems,” Proceedings of Fourteenth IEEE Workshop on Robot and Human Interactive Communication (Ro-Man-05), Nashville, TN. 353-358. [PDF]
- Cory D. Kidd and Cynthia Breazeal (2004) “Effect of a Robot on Engagement and User Perceptions.” Proceedings of IROS 2004. Sendai, Japan. [PDF]
- C. Kidd (2003), Sociable Robots: the Role of Presence and Task in Human-Robot Interaction. Master of Science thesis in Media Arts and Sciences, MIT Media Laboratory. [PDF]