DEVELOPMENT OF ASSISTIVE ROBOT FOR CHILDREN WITH AUTISM SPECTRUM DISORDER

Recent years have seen an ever-expanding number of implementations of robotic technology in health and assisted living. Robots of all sizes and forms are currently utilized both for high-precision diagnostic, physical and cognitive treatment. The value of maximizing patient productivity and minimizi

Project Title

DEVELOPMENT OF ASSISTIVE ROBOT FOR CHILDREN WITH AUTISM SPECTRUM DISORDER

Project Area of Specialization

Robotics

Project Summary

Recent years have seen an ever-expanding number of implementations of robotic technology in health and assisted living. Robots of all sizes and forms are currently utilized both for high-precision diagnostic, physical and cognitive treatment. The value of maximizing patient productivity and minimizing the scope of mistakes makes it vital to adopt assistive robot technology and healthcare assistance. Robots can do some jobs better, affordable, and faster than humans. According to the Autism Society of Pakistan statistics from South Asia estimates that there are 350,000 children with autism spectrum disorder (ASD) in Pakistan. The developmental disability due to ASD is often characterized by social, emotional, and communication challenges. Teaching and training of these autistic children requires thorough professionals and specialized equipment. Autistic people, both infants and adults may have moderate to serious conditions. Early intervention and care have been seen to alleviate these difficulties, but such facilities are nearly absent in Pakistan. Psychiatric system of Pakistan is weak, with meager manpower and resources. To fill the acute shortage of child psychiatrists, we need to provide training and courses, but it will be costly and time-consuming process. This is the scenario where assistive robots play a significant role in diagnosis and treatment of autism in a more affordable and personalized way. Several studies indicate that autistic children can learn way better from robots than from human instructors/psychiatrists. In this project, the proposed assistive robot shall resemble a cat that is a child friendly character and draws more children than a human face with low-cost as compared to pre-build costly robots such as Nao and Zeno. As ASD is incurable, many children on spectrum needs 20 hours of therapy a week which interests more toward a cost-effective assistive robot than to pay a professional psychiatrist every day. The proposed robot has various modules such as puzzle games which offers as a mental exercise, storytelling, verbal exercises, and normal movements such as nodding head and hand shaking to develop social skills in autistic children under the supervision of a caretaker/parent. The 7 degree of freedom cat robot with servos at each joint and two arms which perform dance movements and handshaking with Raspberry Pi as the main controller. The robot will recognize and track the child using camera mounted on its head along processing the image with OpenCV library and passing processed image through a convolutional neural network to classify the facial expression of the child and automatically activate the appropriate module to make the child remain engaged with robot during training activity. The robot has also a screen with graphical user interface to display words and feelings through pictures or text medium. The algorithm for the robot will be developed in Python.

Project Objectives

We aim to design, develop and control a 7-DOF tabletop cat robot . The key objectives of this project are stated below:

1. To develop the robot structure that is liked by children.
2. To add recognition capabilities into the robot.
3. To develop graphical user interface (GUI) to enable the user to control the robot.
4. To establish human-robot interaction capabilities using natural language processing (NLP).

Project Implementation Method

The project starts with the design of the 7-DOF tabletop robot with a defined workspace. The minimum and maximum reach of the robot arms; after modeling the CAD design will be used to develop the robot. 5 MG91 and 2 SG90D Servo motors will be used to control the robot. Raspberry Pi is used as the brain of the robot. The robot has two modes that are automatic and manual. In automatic mode, a camera and a microphone are used as input signals to robot. When a trigger word is detected by microphone the robot will be activated and will look for the child before it, after detecting and recognizing child the robot will seek a command from the child to perform a function such as storytelling, playing a game, verbal exercise, showing gestures and playing music. Amid training session, the robot takes care of child’s engagement in activity by the camera mounted in the head of the robot by capturing the image of child and sends it to Raspberry Pi. The software on Raspberry Pi process the image with OpenCV library to normalize and convert from RGB to gray and then feed that image into a facial emotion recognizer using a convolutional neural network. The output of emotion recognizer is about child’s mood if he is upset then robot will either play music or perform some dance moves to get his attention back else will continue the training. In manual mode, the robot is operated by a graphical user interface through an LCD display installed on its body where a user has authority to opt for any desired function to be accomplished by the robot. After every session of robot with the child, the input responses are gathered and stored for further analysis using suitable machine learning algorithms which shall be helpful in making a tailor-made strategy for child. The effectiveness of robot as compared to human therapist will be measured by engagement rate and improvement in social skills through a survey form. This survey form will be filled by the parent/caretaker of the autistic child.

A flow chart is shown in the Figure 2:

(images/DEVELOPMENT OF ASSISTIVE ROBOT FOR CHILDREN WITH AUTISM SPECTRUM DISORDER _1639950148.png)

Benefits of the Project

The utilization of a robotic platform permits us to make orderly social prompts to animate children and produce response constitutes of explicit social behaviors. In this way, robot-based interactions offer us assistance to discover unique chances in which quantitative quality measurements concerned with pattern behavior can be analyzed. These measurements give a chance to compare children in a standardized way as well as the possibility of penetrating the advancement of a single person over time and environment. The most interesting benefits of assistive robots in training of children with autism spectrum disorder (ASD) for social rules and skills, as robots may be useful not only in replacing of professional therapists and reducing skilled staff required in the treatment of ASD, but also encourages the children in learning process for social skills and rules like distrust. This provides positivity for autistic children in the future, where many social robots will exist in daily life. In Pakistan majority of autism cases don’t even get diagnosed due to the lack of doctor’s expertise and a dearth of professional medical expertise. The cost of early developed assistive robots such as Nao and Zeno are much higher, and our proposed assistive robot will be cost-effective so that everyone can afford it. After completion of project, we will keep the robot in university to provide free service to autistic children of Jamshoro/Hyderabad free of cost.

Educational Tool: The assistive robot shall be used to teach children interactively using a GUI.

Healthcare Assistant: In human therapeutic sessions it is difficult to engage the child for longer periods, whereas an assistive robot provides us an opportunity to create a therapeutic setting where children is engaging in an interaction with more interest and having fun amid training, something that is typically hard for the autistic children in a traditional way.

Technical Details of Final Deliverable

The final product has an approximately 12” by 12” base with two arms and a head that uses 5 MG91 and 2 SG90D Servo motors, and a servo driver such as 16-Channel Adafruit 12-bit PWM Servo Shield shall serve as Servo controller. A Pi Camera is mounted in the head of the robot for tracking, recognition, and visual recordings. The Raspberry Pi is mounted on the same base as robot. When robot delivers lessons verbally, graphical visuals will be shown on the screen fixed on the chest to help make learning experience better and understandable. The touchscreen LCD display is fixed on the chest also used for manually operating robot’s functions. The outer layer of the robot will be a fur of cat to make it look like a cat character. Raspberry Pi, LCD and servo controller are all powered separately. The algorithms will be implemented using Python programming language within the Raspberry Pi.

Final Deliverable of the Project

HW/SW integrated system

Core Industry

Medical

Other Industries

Education , Health

Core Technology

Robotics

Other Technologies

Artificial Intelligence(AI)

Sustainable Development Goals

Good Health and Well-Being for People, Quality Education

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com