INTERACTIVE ROBOTS ON AUTISM SPECTRUM DISORDER
Abstract
The purpose of this study is to examine the therapeutic relationship between Interactive Robots and children with Autism Spectrum Disorder. The disabilities caused by Autism Spectrum Disorder follow a trend of impairment in social interactions and communication as well as repetitive interests and behaviors. Children with autism, find it extremely difficult to integrate in to social spheres, and to learn social skills and behaviors. With the accelerated pace of Artificial Intelligence development, social robots are being used as interactive tools to help children with Autism Spectrum Disorder play with them and to evoke desired behaviors. These interactive robots are programmed to develop children’s social interactions, spontaneous awareness and their connection to the surrounding environment and people. The use of intelligent interfaces and automatons for children with Autism Spectrum Disorder, beginning at an early age, could permit these children to attain normalized developmental breakthroughs and produce community connections which they might not be able to do on their own. While this new and emerging technology is very promising, some of today’s challenges like cost and accessibility of this technology still need to be overcome.
Introduction
Autism is a peculiar mental disorder induced mainly by neuropsychological hereditary brain dysfunction. Termed “Autism Spectrum Disorder” ( Tartaro & Cassell), the developmental disabilities caused by this disorder follow a pattern of repetitive interests and behaviors and degeneration in social interaction and communication (NAS, 2005). Autism is a lifetime developmental impairment that unfavorably affects the way a person communicates and connects with the surrounding environment and people.
​
Furthermore, children with autism demonstrate some degree of different behaviors from children without this disorder and find it difficult to learn social skills due to their interaction impairment ( Won, Mah, & Kim, 2013). Supporting Autistic children by providing adequate assistance is a challenging task; nevertheless, advanced technology and improvements in the field of robotics and Artificial Intelligence (AI) can assist children with autism improve their ability to interact socially. These tools provide an exclusive opportunity to bolster and advance the treatment of autism ( Scassellati, Admoni, & Mataric, 2012).
​
These interactive robots provide personalized therapy for children with autism, which can be tailored to their individual problems and preferences. The robots act as actors, performing suitable behaviors in specific situations to educate the child (Robins B, Otero N, Ferrari E, & Dautenhahn K, 2007). Hence, these robots are designed to improve on Autistic childrens’ social interactive behaviors and as well reduce anxiety and stress suffered by children with autism.
​
While the benefits of this robotic technology to treat Autistic children is unquestionable, cost and accessibility challenges still exist with this technology. Further, the technology is today not mature enough to fully automate therapy and still requires the support from parents and therapists. These challenges will be discussed in a later section.
Autistic Spectrum Disorder (ASD)
ASD has become the most prevailing neurological and developmental disorder that is mainly diagnosed in children today. The United States Centers for Disease Control and Prevention estimates that 1 out of every 88 Americans has been accurately diagnosed as Autistic (Baio, 2010). Children with ASD typically begin showing symptoms as early as twelve months of age, though two to three years is the usual age of diagnosis (Rowan, 2012). The symptom of autism falls in a spectrum, no child is affected by the disorder in the same way. Severity varies from curtailing a regular life to an overwhelming impediment that may require governmental care (Melillo, 2012). “The core impediments in Autism Spectrum Disorder (ASD) are often characterized on three key extensions: Community relationships/interaction, Social communication and Imagination” (John-John & al, 2013).
Interactive Robotics and their use in treating Autism
Socially interactive robots are helpful to communicate, express and discern emotions, cultivate social relationships, construe natural cues, and advance social competencies ( Fong T, Nourbakhsh I, & Dautenhahn K, 2003). Social robots are being used as tools to provide therapeutic treatment to children with autism, to play with them, and to evoke desired behaviors.
​
They generate interesting, attractive and purposeful interplay situations that force children to interact with them. Although there are several other interactive software agents and computer aided therapy models developed for autism therapy to activate different skills, the interactive robots can do a better job in helping the autistic children (Kozima H, Michalowski MP, & Nakagawa C , 2009). There have been many interactive robots built in the recent past by several agencies and universities in different countries to interact with the autistic children. These robots differ in their appearance, behavior and the type of activities that they perform.
​
Interactions with robots play a major role in engaging autistic children because robots can gain the attention of the children better than humans. These child-robot interactions will help autistic children to overcome the impediments and improve their communication skills, emotional awareness and understand the world better. The interactive robots exhibit positive behaviors during the therapy sessions and educate the children with autism (H & Zlatev J, 2000). The important problem of the autistic children is that they find it very hard to read and understand facial expressions of the humans. Interactions with humans can evoke unwanted sensory stimulations causing stress to the Autistic children. Child-robot interactions are notably different because these are designed to exhibit only a limited set of basic emotions excluding any sensory overload ( Michaud F, et al., 2007).
Research behind developing Therapeutically Interactive Robots
The efficiency of the therapy depends significantly on the design and functionalities of the robot.
Not all the features of the robot will attract autistic children. The receptiveness of Autistic children may vary widely. However, it must be noted that there are some similarities in children’s reactions to certain characteristics of the robots.
​
Several research activities have been performed to determine the best design features that would suit Autistic children. This research included feedback collected from a panel of experts, therapists, parents and teachers (Ferrari E, Robins B, & Dautenhahn K, 2009). However, children with autism were not included in this study, due to their difficulties in communicating properly, which one could view as partially compromising the results of this study. (Duquette A, Nadeau I , & F Michaud, 2003).
Autistic children are noted for having a short concentration span, so it’s important for the robots to visually engage the children (Dubowski J, Dautenhahn K, & Robins B, 2004).
​
The color combination of the robots must be chosen in a way to not over stimulate the children but, at the same time, to grab their attention. Different shapes, lights and other graphical parts can attract the autistic children. Sharp edges and bright colors are not advisable ( Cabibihan JJ Cute & Hoa TD,, 2012). “Teaching children with autism to interact with the robots may increase a fluidity of thought and reduce conceptual fragmentation.” ( Peter M. David Fulton & Sherratt D, 2002)
Challenges with usage of Robotics for ASD treatment
Cost is a key impediment to this robot technology being accessible to every autistic child. Further, robot technology today cannot fully replace the need for the care and education provided by parents and therapists, which further increase costs. Research suggests that children with cognitive disorders, especially those with autistic spectrum disorders, are accompanied by their parents and therapists while interacting with the robots, which may necessitate that one parent stay home to care for the autistic child, thereby further stressing the family finances.
The use of robotic technology for autistic spectrum disorder is an emerging area. The robotics study demonstrated that the robotic care using this new technology can be very expensive. The complexity of a robot should be suitable for a particular child with autism, so that the robot’s behavior should attract the child and grab its attention and draw out various actions from the child. It’s vital for the children with autism to interact dyadicaly with a sense of security and curiosity, in which they gradually get used to the robot’s actions and response. The way this dyadic interaction works with minimum intervention from therapists varies for different kids (Cocoro Nakagawa & Hideki Kozima , 2004). This requires significant sophistication and customization in robot design. The interactive robots would need to be able to build a profile of the user and understand the requirements to cater to the needs of the users with minimum human intervention (Todd H. Wagner & et al, 2011). All of these requirements increase the cost of developing such robots, which is a significant impediment in providing access to this technology for all autistic kids.
​
Autistic children have the right to free public education. With the increase of children with the autism spectrum, the public schools are finding more ways to educate the autistic children. Parents of children with autism often go through stress and anxiety disorders because of the cost of the robotic treatment. It’s important to reduce the cost of the treatment by automating the functions of the robots and ensuring maturity of the technology through public and government funding for developing such robots.
While early intervention services are crucial in the development of children with ASD, not many parents accept the fact that their kids are autistic due to the fear of stigma.
Additionally, a lack of awareness of the probability of ASD, adverse effects without early intervention, resources and the latest technology available for treatment has a negative impact on the inception of the treatment (Johnny L.Matson & Alison M.Kozlowski, 2010). Educating parents on the available options to treat ASD can significantly reduce cost by ensuring more early intervention.
Current and Emerging Advances in AI
Emerging advances in artificial intelligence coupled with the quenchless global demand for digital technology drives the need to enhance the quality, effectiveness and approachability of therapeutic treatments (Dimitrov, 2016). Understanding machine learning, pattern recognition and natural language processing have caused a great impact in detecting and assessing the spectrum of the disorder. Virtual reality has made way to create intelligent, compassionate and interactive human robots/agents, skillful in providing care and surroundings favorable to therapeutic treatments. The advantage of this technology is to assist children with ASD and also to improve their mental health. A very good example is the Aurora project (Dautenhahn, 1999), which studies the future implementation of robots as therapeutic toys, especially for children with autism. Kaspar, a child-size humanoid automation was designed by the Adaptive Systems Research Group. Its serviceability varies with projects on advancement and cognitive automations ( Dautenhahn & Robins, 2004). An innovation such as this exhibits a promising future for assisting a number of autistic kids. These Interactive robots are designed to develop children’s social responsiveness, spontaneous awareness, and communication with the surrounding community and atmosphere. They also help in teaching the kids to elicit socially acceptable behaviors. Further benefits of AI- mechanized treatments include: triadic interactions, special education adviser can team up with the designers and the robotic software engineers to progress and extended ecosystems for autism therapies. This technology can be deployed in a classroom, where students can interact with the socially interactive devices in a unified and modifiable approach (Dautenhahn, 2003). The usual methods of monitoring the cognitive health can be very expensive and time consuming, but the emergence of new digital treatments have a very high tendency of reaching people bypassing the stigma, disgrace and disbelief (Arkin, 2010).
User Needs
With the accelerated pace of development it’s absolutely important for the designers to understand the user’s goals, mental/cognitive restraints, evaluation of the stage and treatment options when designing a system/robot aimed at improving the cognitive, social and emotional conditions of the kids with autism. The eventual goal of the user is to improve his/her cognitive capabilities, improve social skills, eye contact and joint physical expressions. All the child-robot interactions are not only designed for the child to learn the important skills during the therapy sessions, but also to be able to apply those lessons when interacting day to day with the people around. The ultimate goal must be to improve the child’s social interaction with the people around, and not just with the robots (Cabibihan, Javed, Jr, & Mariam, 2013). Generally, autistic children get very agitated when other children try to play/interact with them. The limbic area which controls emotions is affected for the autistic kids; especially the amygdale is the area of the brain responsible for aggression and emotions. So, it’s important for the designers to understand the exact problem areas and the requirements of the users to build an intelligent interface (Sousa, 2007). Some of the important areas of requirements includes imitation, eye contact, self-initiation, improve social skills and interactions (S, 2006). The robots feature should be customizable, if the autistic child gets intimidated by the appearance of the robots eyes, it should be removable. ( Nadeau I , Duquette A, & Michaud F, 2003)
Designing for future social therapies
Future AI and robotic technologies have a very high potential of being an extra set of eyes and ears for teachers, therapists and the parents. While the children with autism/ASD interact with the robots in school or at home, information about the displayed social behaviors and other quick assessments to measure if a child understands the activity should be captured regularly. This data can be used to analyze strengths and weaknesses of the children with autism. The child’s emotional development and learning capacity can be evaluated by the teachers, by slowly exposing them to diverse social environments (How Social Robotics is Revolutionising Therapy for Autistic Children, 2013).
Combined AI and robotic technologies should be incorporated in designing the therapy for children with autism, whether it be virtual software like Siri or hardware like the interactive robots. These help Autistic children successfully maneuver the different settings and social interactions, while implanting a sense of calmness in them. The normal developmental milestones and social connections can be easily achieved by the Autistic children by using combined AI and robotics technology from a young age. These technologies play a vital role in bolstering the kids’ confidence and social skills that they need to live independently. Records show that already a small number of Autistic kids have transitioned to the mainstream classes from special programs designed for autism (Goldstein).
So, it’s important to design an intelligent robotic system that understands and adapts quickly to the user’s needs and also manages all the tasks by itself without any human intervention. Most importantly, the technology must store data and provide feedback for further analysis. These robots should have facial recognition features to store the user’s personalized needs and provide more autonomous care.
Intelligent User Interfaces
The four areas of impediments in communication for children with ASD are (1) delay in spoken language, (2) difficulty to interact or start a conversation, (3) repetitive use of language and (4) lack of social imitative play (Center for Autism and Related Disorders [CARD], 2013). In addition, a serious sensitivity to noise and smell is also prevalent with the Autistic kids. Endorsing the social nature of human beings and its significance for artificial intelligence (AI) is an interesting challenge that requires interdisciplinary references. Human –robot interaction (HRI) plays a vital role in incorporating these viewpoints (Maulsby D, Greenberg S, & Mander R). Designing a multi modal interface such as text, audio, visual and haptic would be very useful for these kids to function efficiently. The cognitive load can be reduced with the help of the multisensory approach. (Vega M.D, 2013).
​
Further, the intelligent user interface should also be capable of diagnosing and fixing bugs within its constantly developing and adaptive framework. To accomplish this, it must know all the technical details of the parts (both hardware and software modules) that it is made up of and how they are assembled together. This type of constant diagnosis can save time and money for the users. Ideally, the machine should be able to learn, update and adapt its skills by observing the patterns and behaviors of the users.
Current and Future considerations for Robotic Therapy The insatiable global demand of digital technology pushes the need to improve the quality, efficiency and approachability of therapeutic treatments.
The spectrum of the disorder can be assessed by implementing many of the AI techniques such as machine learning, pattern recognition and natural language processing. Virtual reality has paved way to generate intelligent, empathetic and interactive humanoids that are adept in caring and providing surroundings suitable for therapeutic treatments.
Varied considerations need to be incorporated when designing the interactive robots. Firstly the robots must be visually appealing and engaging to a child. A mindful balance must be found so that the robots do not stimulate the autistic child unnecessarily.
Secondly, designing immensely realistic (human like) robots is not advisable, because if the robots are too human like then the autistic children might not be interested in interacting with the robot. Providing a partially mechanical look/appearance is essential to trigger child’s interest, but at the same time the robots should not be too mechanical because it can distract the child’s attention in analyzing its components rather than interacting with the robot (Nicole Giullian, et al., 2016).
Thirdly, the functioning of interactive robots should be understandable to the therapists. Navigating the robots in therapeutic environments and interacting with it effectively. Triadic choreographies must be designed effectually for the robots to follow the instructions given by the operators/therapists. A therapy session can be compared to a dance number: the whole session is based on the choreographed dance moves. The robots might not be adequately trained to perform different actions. The involvement of child and therapists in this performance also plays an important role.
A current study supports preliminary proof showing the beneficial action of long term active music therapy in a group of children/ young adults with severe autism. Since music seemed to be of clinical advantage for specific core domains of autism, a randomized controlled trial of long-term music therapy in autism is guaranteed (Boso Marianna, Politi Pierluigi, Minazzi Vera, Abbamonte Marte, & Emanuel Enzo, 2007).
One of the biggest mysteries right now is that there are a lot of inconsistencies among different people on the Autistic spectrum. It’s absolutely essential to pin down and classify that spectrum into people who reason in a specific way and provide more support.
More real time data of user interaction information specific to the user’s preference should be stored in the system in order to quickly retrieve the data and provide support to the users without wasting time and money.
Conclusion
Considering the number of children genetically affected by Autism Spectrum Disorder, which causes impairment of social interactions and communication, the need for developing low cost and accessible therapy is essential today. Thanks to the accelerated pace in technological advancement in Artificial Intelligence and robotics, interactive robots have been designed to partially alleviate this problem and provide therapeutic treatment in order to stabilize emotional awareness and communication with the surrounding atmosphere. While several challenges like cost and accessibility of this emerging technology remain, the promise and usefulness of this technology to help treat ASD is unquestionable. On-going research in AI, intelligent user interfaces and robotics is expected to continue to drive down cost, increase the ability to treat autonomously with minimal human intervention and provide diagnostic feedback to therapists and parents.