An educational robot for children that uses a camera to recognize numbers and English letters (EasyOCR), fruits (YOLO) and colors (HSV), then speaks the item's name.
A project to design and build an interactive educational robot for children based on artificial intelligence and computer vision, carried out by a team of students with technical assistance from Techno Enjaz. Using a front-facing digital camera, the robot captures the card or object a child shows it and recognizes numbers and English letters with EasyOCR, fruits with the YOLO algorithm, and colors with the HSV color space, then immediately plays an educational audio clip that names the item.
The robot's body was made with 3D printing and houses an Arduino Nano board, an HC-05 Bluetooth module, and three MG995 servo motors (one for the arm, two for the legs), powered by two Li-ion 18650 batteries. The voice clips were pre-generated on the ElevenLabs platform with a custom robot voice. In practical testing, the robot successfully recognized all four categories, displaying the item name and a confidence score and playing the matching audio; the report does not provide quantitative recognition-accuracy measurements.
| Item | Details |
|---|---|
| Project type | Interactive educational robot (embedded hardware + computer-vision software) |
| Field | Educational robotics, computer vision, embedded systems, early learning |
| Project status | Assembled final prototype, tested in practice with cards and objects in front of the camera |
| Techno Enjaz role | Technical support and assistance to the students during project development |
| Core technologies | Python, EasyOCR, YOLO, HSV with OpenCV, ElevenLabs, Arduino Nano, HC-05, MG995 servos |
| Outputs | Recognition of numbers, English letters, fruits, and colors; instant voice response; welcome gesture; GUI with live feed and controls |
Many tools for teaching children still rely on rote learning and memorization with limited interaction, while early childhood calls for methods that combine visual, auditory, and physical engagement. Children struggle to learn basic concepts such as numbers, letters, and colors when they are presented only theoretically, without a way to link what they see to what they learn.
Many educational apps and electronic systems also present static content and do not use computer vision to recognize real objects in the child's environment. The project therefore set out to build a robot that recognizes the item a child holds up, answers with a clear voice, and does so in an engaging physical body that moves.
The project was carried out academically by the students, with technical assistance and support from Techno Enjaz during development. It is presented among the office's work as an example of its support for projects that combine computer vision with embedded systems and Arduino microcontrollers in educational applications.
The system follows a repeating path from image capture to voice response:
The system block diagram divides the work into five layers: input (camera), preprocessing, artificial intelligence (the three models), decision-making (selecting the audio clip), and output (sound and motion).
| Category | Technology | How it is used in the project |
|---|---|---|
| Numbers and English letters | EasyOCR | CRNN architecture combining a CNN for feature extraction, an LSTM for character sequences, and CTC for decoding the text |
| Fruits | YOLO | Detects the object in a single pass over the image, returning a bounding box, class, and confidence score |
| Colors | HSV color space | Converts the image from RGB to HSV, then applies color thresholding to extract a mask and determine the dominant color |
| Voice response | ElevenLabs | Pre-generated audio files for all words and phrases, played immediately after recognition |
HSV was chosen because it separates color information from brightness, making color recognition more stable under changing lighting than RGB. EasyOCR was chosen for its ease of use with Python and OpenCV and its built-in multilingual support. The project preferred pre-generated voice clips over on-the-fly speech generation to keep responses fast and pronunciation quality consistent; the voice was designed with a text prompt asking for a clear educational robot voice in Modern Standard Arabic, with a gentle, encouraging tone suited to children.
| Component | Role in the robot |
|---|---|
| Arduino Nano (ATmega328P, 16 MHz) | Main controller for the servo motors |
| HC-05 Bluetooth module | Wireless serial (UART) link between the robot and the software when needed |
| 3 × MG995 servo motors | One for the arm, two for the legs; metal gears, 11 kg·cm torque at 6V |
| LM2596 step-down converter | Lowers the battery voltage to the level required by the board and motors |
| 2 × Li-ion 18650 batteries | Power source (3.7V nominal, 3300 mAh per cell) |
| Digital camera | Captures images of items at the front of the body |
| Speaker | Outputs the educational audio |
The control signals of the three motors are wired to pins D9, D10, and D11 on the Arduino, the HC-05 is connected to the TX/RX pins, and all modules share a common ground to reduce noise and communication errors. The components were chosen after comparison with alternatives such as Arduino Uno and ESP32, HC-06, SG90 and MG996R, and the 7805 regulator, based on size, power consumption, torque, and ease of integration.
The computer-vision program was developed in Python with a graphical interface that shows the live camera feed, including:
The report shows four application screens: recognition of numbers, English letters, fruits, and colors.
According to the practical testing documented in the project:
These are functional results for a prototype. The report contains no project-specific quantitative tables, such as per-category recognition accuracy, the number of images tested, or response time, so no specific accuracy figures are attributed to the robot. General figures cited in the report about EasyOCR or YOLO are known properties of those tools, not measurements of this robot.
The project compares itself with commercial educational robots such as Anki Cozmo, Miko 3, Roybi, and Wonder Workshop Dash. According to that analysis, those systems focus on conversation, language learning, coding, or ready-made content, with limited support for visually recognizing real objects in the child's environment. The project defines the gap it targets as combining computer vision with bilingual voice interaction in a single, extensible platform.
According to the outlook in the project, the robot could be developed by:
If you are working on an interactive educational robot for children and need technical help developing the idea or building a prototype, you can contact the Techno Enjaz office to discuss your project's requirements and the appropriate scope of support.
An educational robot with a 3D-printed body, built by a team of students with technical assistance from Techno Enjaz. A front camera captures the card or object a child shows it, the robot recognizes numbers, English letters, fruits, or colors, and then plays an audio clip naming the item.
It uses EasyOCR (a CRNN architecture) for numbers and English letters, the YOLO algorithm to detect fruits, and conversion to the HSV color space with color thresholding to find the dominant color. The software is written in Python with a GUI showing the live feed and confidence score.
The robot includes an Arduino Nano controller, an HC-05 Bluetooth module for wireless communication, three MG995 servo motors (one for the arm, two for the legs), an LM2596 step-down converter, and two Li-ion 18650 batteries, plus a digital camera and a speaker.
Practical testing showed the robot recognizing all four categories directly from camera images, displaying the item name and confidence score, playing the matching audio, and running stably during continuous operation. However, the report includes no quantitative measurements such as per-category accuracy or the number of images tested.
Recognition is limited to numbers, English letters, fruits, and colors, and the only movement actually implemented is a single welcome gesture at startup. Voice clips are pre-recorded, the robot holds no open conversation, and it does not track the child's progress; these are listed as future developments.
Reach out to Techno Enjaz engineering team to discuss technical feasibility, prototype hardware, and development scope.
Discover more engineering prototypes and academic implementations
A remotely controlled ground robot prototype combining Arduino UNO, three wireless links (Bluetooth, Wi-Fi, 433 MHz RF) and a camera whose images are analyzed with YOLO.
A robotic hand that mimics a human hand opening and closing in real time via camera hand tracking, sending finger states over Bluetooth to an Arduino UNO driving servos.
A prototype that measures a tank's water-level percentage with an HC-SR04 sensor and Arduino Nano, showing it on an LCD and sending it via HC-05 Bluetooth to a phone app.
