Jacob Robinson and Dr. Mark Colton, BYU Mechanical Engineering
Introduction
Autism is a spectrum disorder that is characterized by an impairment of the child’s social relationships, communication skills, and imaginative thought, resulting in a decreased quality of life for the child and his or her family, as well as significant economic costs to society. Although clinical treatment has helped some autistic children learn social skills, many others do not respond well to clinical treatment. Some research groups have investigated the use of robots as a tool to help therapists connect to the patient easier. Autistic children are more inclined to interact with a mechanical device because they exhibit repeatable actions that are easy for the child to understand.
Studies conducted by faculty and students from the mechanical engineering, computer science, and communication disorders departments at Brigham Young University have shown that the use of a humanoid robot in connection with clinical therapy has helped child-therapist relations and has resulted in an increase in joint attention and other desirable behaviors [2]. The current robot, named Troy, (shown in figure 1) is capable of simple interactions, which the therapist uses in low doses to help the child feel more open to other therapeutic activities. The robot uses a rectangular LCD monitor in a case to represent its face and head.
A new version of Troy is currently being developed with the aim of making the robot more interactive and expressive. The research presented here addresses the deficiencies of the current head by designing a new head that allows for more robot-child interaction and physical modularity to enable future customization to meet the individual needs of the children and therapists.
Methodology
To understand what would make a good head and face for an assistive robot for children with autism, a literature review was conducted. Current research shows a wide range of types of robots and clinical usage. However, one point is clear: To interact emotionally with humans, the robot must not only cover the discrete basic emotions, but must be able to switch continuously from one emotion to another [3]. Therefore, the new robot head needed to be capable of more emotional expression. From consulting with the researchers who designed Troy, it was decided that the LCD screen was a good design that allowed a range of emotions to be easily programmed by changing the facial expression. But one drawback was the set shape of the rectangular screen. Ideally Troy would be built in a way that its appearance could be adjusted to be more human-like or more machine-like depending on the needs of the child.
Using these goals of modularity and expressiveness, concepts were generated that could meet these requirements. Figure 2 shows the CAD model of the concept that met these requirements best. The model uses one μLCD screen (mouth) and two μOLED screens (eyes) which a face can be displayed on. The screens were mounted on the existing robot’s neck for simplicity. This design allowed for the most modularity of the head while still allowing the face to be easily programmed with various expressions. The shape of the head could then be adjusted by adding a shell designed to the desired shape.
This concept was prototyped using one μLCD screen (mouth) and two μOLED screens (eyes) from 4D Systems (see figure 3). The screens were mounted to the head using a frame constructed from clear acrylic and were then programmed using the 4D Systems integrated development environment (IDE) called 4D Workshop4.
Discussion of Results
The resulting prototype shown in figure 3 is a good start to the development of a more interactive and modular robot head for Troy. The mouth display is a touch screen, thus providing the option of touch interaction with the child. As the clinicians come to better understand the needs of the children they are working with, head shells can be shaped and installed. The separate displays for the mouth and each of the eyes creates a face that is flexible and free to adjust. Overall this makes the new head for Troy much more modular and capable of a wider range of emotions that before.
Conclusions
This research has been a key part of the development of the new Troy. With the introduction of the separate screens for the mouth and eyes, the head can now be finalized with a mounting structure that allows for variations of face configuration and head shells to be attached. Assistive robotics with autistic children is still a new field and requires learning through trial and error, especially since every child has different needs. This new head design will provide the modularity that is crucial to this field. Also, as I continue into my master’s research I will be working to make even more parts of Troy modular and flexible. That way the clinicians will have the most freedom with one robot.
References
- M.A. Goodrich, M.B. Colton, B. Brinton, M. Fujiki, J.A. Atherton, D. Ricks, M.H. Maxfield, A. Acerson, “Incorporating a Robot into an Autism Therapy Team,” IEEE Intelligent Systems, Vol. 27, No. 2, 2012, pp. 52-59.
- D.J. Ricks, M.B. Colton, and M.A. Goodrich, “Design and Evaluation of a Clinical Upper-Body Humanoid Robot for Autism Therapy,” Proc. International Conference on Applied Bionics and Biomechanics (ICABB10), Venice, Italy, 14-16 October, 2010.
- Greet Van de Perre, et al. “About the design of the social robot Probo, facilitator for ASD therapies,” 9th National Congress on Theoretical and Applied Mechanics, 2012
Figure 3: Final prototype of robot head