Development of educational robot for slow learners

Project Title

Development of educational robot for slow learners

Project Area of Specialization Artificial IntelligenceProject Summary

Many slow learner children (such as those diagnosed with autism) face speech problems. They require speech therapy to improve their learning and communication skills. Speech therapy is not only expensive but qualified speech therapists are available in large cities only. Hyderabad, which happens to be second largest city in the province of Sindh, has nearly no qualified practicing speech therapists. Even, where speech therapists are available, they lose interest due to repetitive and monotonous nature of this job. They can get irritated due to this monotony, resulting in lower quality of therapy. It has been found that robotic speech therapy gives better results as compared to human speech therapist. Facing and replying the same answers to the same questions sometimes makes them answer either harshly or not at all. Studies also show that children take harsh behaviors and lack of attention seriously and deep down they start feeling excessive fear, worry or anxiety, dramatic changes in eating or sleeping, long lasting sadness or irritability. This project will be developed to be provided in schools and hospitals to aid students/children with deep learning disabilities. Since robots are machines and machines are invented to do repetitive work more efficiently, thus they do not get tired of doing the same task again and again. This project will encourage the children to learn in a way that feels more like creativity and less like homework.

Project Objectives

1. To develop robot structure
2. To add speech recognition capabilities to the robot
3. To incorporate interactivity programmatically

Project Implementation Method

1. The selection of robot structure.
2. The selection of hardware for robot.
3. Selection of software.
4. Development of Python code for speech recognition.
5. Integration of microphone and speaker modules for speech recognition and speaking respectively.
6. Integration of suitable camera module for facial expression recognition. If the expressions show that the question is understood, then it expects for an answer else it will repeat the question with some additional information.
7. Machine learning will be used to measure the accuracy of given pronunciation/answer. If the pronunciation/answer is correct then the robot will praise else it will try to lead child to correct pronunciation/answer.
8. A screen will be used to display related information.
9. Project results and report.

Benefits of the Project

1. Use as a speech therapist:

Meets the demand for the speech therapist.

1. Need:

As speech therapy is rare in Pakistan. We came to following conclusions:

1. There are almost no speech therapists.
2. If there are, they aren’t as good to be helpful.
3. And the good therapists are far more expensive that even a middle-class family can afford.
4. Language barriers: People usually are native speakers of one language and therapist other.

Thus, subject to the funding, we will develop the robot that is free of cost. This robot will be offered in schools and hospitals. This is helpful for special, deaf and dumb students/children.

1. Repetitive task:

Humans are good at teaching, better than robots for now (maybe). But robots are really the best option for repetitive tasks. They don’t get tired teaching the same thing again and again.

1. Anti-biasing:

Human nature gets attracted to things that tend to response in their way. This creates a biasing and preference. Robots neither get biased nor prefer some student over another.

1. No language barrier:

This project could be used for multiple languages. User can speak the suitable language and the answer will be given accordingly.

1. Cost effective:

This robot will be much cheaper than existing robots. Following are some links that prove it:

1. Pepper: 1700 USD (approximately 3 lac PKR)

https://www.mediapost.com/publications/article/276387/pepper-the-robot-will-be-your-companion-for-a-pri.html

1. Kasper: 2118 USD (approximately 3.5 lac PKR)

http://kannanking.blogspot.com/2011/03/kaspar-kids-robot-robot-can-interact.html

1. Nao: 18490 USD (approximately 29 lacs PKR)

https://www.robotlab.com/store/robotlab-nao-autism-pack

Technical Details of Final Deliverable

The project in the end would have a robot with dimensions of approximately (2’, 2’, 4’) length, width and height respectively. An on-head camera will capture facial expressions and send it to raspberry pi where a python algorithm will be used to estimate the understanding level of the children about the question that was asked. Speech will be used as input and output for both questions and answers of the child and robot respectively. Along with speech, the robot will also show the related information of the question on the peripheral device (i.e. LCD mounted on body).

Final Deliverable of the Project HW/SW integrated systemCore Industry EducationOther Industries Medical Core Technology Artificial Intelligence(AI)Other Technologies RoboticsSustainable Development Goals Good Health and Well-Being for People, Quality Education, Industry, Innovation and InfrastructureRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	75820
Raspberry pi 4 model B	Equipment	1	9750	9750
Microphone module	Equipment	1	2700	2700
Raspberry pi case	Equipment	1	500	500
Speaker	Equipment	1	3000	3000
LCD on face	Equipment	1	7800	7800
LCD on body	Equipment	1	10700	10700
Battery	Equipment	1	7000	7000
Camera module	Equipment	1	900	900
Battery charger	Equipment	1	1500	1500
Body and parts	Equipment	1	20000	20000
Raspberry pi adapter	Equipment	1	1060	1060
Wheels for mobility	Equipment	4	220	880
HDMI converter	Miscellaneous	1	530	530
Memory card	Equipment	1	1000	1000
Jumper wires	Miscellaneous	50	10	500
Documents/Report printing	Miscellaneous	300	5	1500
Shipping cost (all times combined)	Miscellaneous	1	1500	1500
Screws and bolts	Miscellaneous	1	5000	5000