Posted in Uncategorized

How to read an academic paper (for class)

These answers are all based on the Supporting Clinicians in Robot-Assisted Therapy
for Autism Spectrum Disorder: Creating and Editing
Robot Animations with Full-Body Motion Tracking by J. Alan Atherton and Michael A. Goodrich (2011)

Did the abstract tell you the three things I said it should? If not, what did it tell you?

What is the research topic?

Creating a way for clinicians to no longer have to rely on programmers every time they need to modify the actions of a robot. Instead software and hardware that can track the movement of the user and then apply this to the robot is proposed as a useful solution with the clinicians, computer scientists and mechanical engineers needing to be collaboratively involved in the design process of this tool.

What did the authors do?

It doesn’t state in the abstract what the authors did, it just talks about their experience of clinician’s desire to change the robot’s behaviour to suit the needs of different children and that because these clinicians can’t program they need to find a solution that will allow them to change the robots behaviour.

What they discovered

That robots may be more useful to clinicians if they are able to create and modify actions for the robot. And that for any idea on how to fix this problem; clinicians, computer scientists, and mechanical engineers need to work collaboratively in the design process.

What seems to be the research question(s) they were trying to answer?

  • How to empower clinicians to customize the robots behaviour to suit the needs of each child?
  • Ways to change the robots behaviour without the need for traditional computer programming.
  • How to make the solution to the problem of changing the robot’s behaviour without traditional computer programming simple for the clinicians to use.
  • How to design an interface that a clinician will find easy to use?
  • Why will a full body gesture-based method for motion capture help the clinicians solve this problem?
  • Why do clinicians need to be able to program the robot?

What Method(s) did they use to answer the question(s)

  • Review research papers and reports on the subject.
  • Talking with Clinicians who currently use SARs.
  • Programmed their own motion sensor program.
  • Discussing with their design group which includes professors and clinicians who are involved with therapy for children with autism, mechanical engineering professors and students, computer science professors and students.

How Credible do you think the paper is?

To decide whether a paper is credible or not, first you must look at the authors themselves, what credibility do they have themselves:


J. Alan Atherton

Work: Department of Computer Science Brigham Young University Provo, UT.

Are they an expert: According to google scholar he is a PhD candidate and has been cited in 119 articles since 2012. His focus is Human-Robot Interaction, Human-Computer Interaction, Computer Graphics, Animation and Artificial Intelligence.

Michael A. Goodrich

Work: Department of Computer Science Brigham Young University Provo, UT. (Professor, Department Chair)

Are they an expert: According to the Brigham Young University website he received his PhD in 1996 in Electrical and Computer Engineering under the direction of Dr. Wynn Stirling. After graduation, he completed two years of postdoctoral research in an associate position at Nissan CBR. His doctoral work was in intelligent control, and his post-doctoral research was in computational models of intelligent human behaviour. He spent one year as a research assistant professor in the Computer Science Department at Brigham Young University where he is currently a Professor of Computer Science. His research interests are; human-robot interaction, human-vehicle interaction, multi-agent learning, artificial intelligence and decision theory.


According to Google Scholar it has only been cited 12 times but those that have sited it have themselves been sited 30 to 140 times which shows us that those who wrote those reports believe that it is credible.

It seems to have been published by the university and so is hard to pinpoint whether it is peer reviewed or not. As it has been used by multiple



  • Robots can be beneficial to therapy for children with ASD.
  • Clinicians need to be able to modify robot behaviour.
  • Motion capture may enable intuitive generation of complex animations and algorithmic analysis inside a gesture-based user interface can help to modify the captured animations intuitively and efficiently for non-experts.

What they seem to mean

  • Studies have shown that robots help children with ASD learn social behaviours that seem to easily move over to being used not just with robots, but with parents and children their age too.
  • Clinicians can’t all take courses in programming, nor can they turn to a programmer every time they need to make the robot do something different for each child. This is why a method that is easy for the clinicians to learn and understand needs to be created before robots become more readily available as tools used within therapy environments. The robot needs to behave differently for each child because no two people with ASD are the same.
  • Motion capture seems a viable solution because with the technology getting cheaper and more readily available and with motion being a skill that people are born with it becomes a logical way forward into programming these robots to be able to be used easily by the end user. Since the hardware is already existing within things like game consoles the next step is to figure out how to create software that anyone can easily pick up in either a mini course or using an instruction manual.

Do I agree with what was written in the conclusion?

I do agree with the conclusion, because as I’ve found through my research of Autism is that no two children with Autism behave the same. A one solution fits all robot wouldn’t work well as a therapy tool. Programmers can’t be on hand 24/7 and this is why Motion Tracking would be the perfect solution for Therapists. To be able to change the behavior of the robots will save them time and be a good assist tool in their therapy.

What I learned from the paper

The SARs need to be programmed to suit the child

Socially Assistive Robots can be reprogrammed to suit any child’s needs in a therapy situation. But currently this is a hassle and it leaves many children with a one size suits all situation. The robot’s need to be programmed to each be as unique as the child itself that is using said robot otherwise the therapy sessions will have no real need for the SAR.

Microsoft’s Kinect sensor is cheap and powerful enough to help solve the issue.

The reason why sensors were not seen as a solution in the past was because they were too expensive. Since Microsoft developed the Kinect, this has become a much more affordable approach, even cheaper than requesting a programmer to do the reprogramming every time a new child starts a therapy session. The Kinect combined with PrimeSense’s NITE full body tracking has made it so programmers can design a software for the robots that takes these motions and relates the human movement to the robots movement. A Nintendo Wii remote is used for the discrete events.

Description of the paper

Robots are currently being used as tools by therapists for children with Autism. They can teach Autistic Children social skills that they find difficult to learn from human to human interaction. These social skills can be then carried over into interacting with other children over time. The robot will play imitation games with the child and when the child shows that it is uninterested in the activity the robot can sense this and change its actions. When they show that they understand this task well the therapist introduces the child to another child with autism and the robot will play a turn taking game with them, allowing the children to interact with each other through the robot.

When a therapist needs the robot’s behaviour to change to suit each child, they must rely on programmers to come in and reprogram the robot for them. This can take up a lot of therapist’s time and money. Because the field of robots being used as therapy assistants is still in an exploratory stage, not many solutions have been tested to solve this solution.

This paper discusses the idea of using a motion sensor to allow a therapist to easily reprogram the robots themselves. The only step programmers would need to be involved in is the software that picks up on the therapists’ movements. Motion sensors weren’t considered in the past because they used to be a very expensive piece of hardware. But with the release of Microsoft Kinect it has become a more viable solution, allowing therapist to easily create unique intimation games for their patients.


One thought on “How to read an academic paper (for class)

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s