Voice-
Controlled
Robot With
the ESP32
21BEC1154-HARIESH P
21BEC1142-HARISH M
21BEC1294-ENIYAN KM
Objective
This project aims to
develop a voice-
controlled robot
leveraging the power of
the ESP32
microcontroller. The robot
is designed to respond to
voice commands,
providing a hands-free
user experience.
2
• The core of this project is the implementation of a
Introduction voice recognition model trained with TensorFlow. This
model, which runs on the ESP32 using TensorFlow
Lite, is capable of recognizing specific “wake words”
or commands. Once a recognized command is
detected, the robot performs the corresponding action.
3
Concept
The concept revolves around
integrating a robust speech
recognition system into a robot,
allowing it to interpret and respond
to spoken commands or queries
from users
4
Methodology
Slide Title: Methodology
1. Microphone Setup
•The system is compatible with both I2S and analogue microphones, with I2S microphones recommended for
their superior noise characteristics.
2. Voice Recognition
•Implement voice recognition using a model trained with TensorFlow.
•The model runs on the ESP32 microcontroller using TensorFlow Lite.
3. Pre-Trained Model
• A pre-trained model is included for immediate use.
• The model is trained for “wake word” recognition, allowing the robot to respond to specific voice
commands.
4. Implementation
• This project will be implemented with the help of the a robo car,microphone system ,esp32.
• Tensorflow lite is runned with the help of ESP32 microcontroller .
5
Expected Outcome
1. Voice-Activated Control
• The robot will be able to respond to specific voice commands, providing a hands-free user experience.
2. High Recognition Accuracy
• With the use of TensorFlow and a well-trained model, the system is expected to have high accuracy in
recognizing and responding to the designated “wake words”.
3. Flexibility
• The system will allow for the use of different types of microphones and even the ability to train your own
voice recognition model, providing flexibility in its application.
4. Real-Time Operation
• The robot is expected to operate in real-time, with minimal delay between the voice command and the
corresponding action.
5. Scalability
• The methodologies used in this project can be scaled and applied to more complex systems, paving the
way for more advanced voice-controlled applications in the future.
6
Hardware Model Implementation
1. Microphone Breakout Board
•The core of the voice recognition system is a microphone breakout board based on the ICS-43434. This board captures the voice
commands.
2. ESP32 Microcontroller
•The ESP32 microcontroller is the brain of the robot. It processes the voice commands and controls the robot’s movements.
3. I2S or Analogue Microphone
•An I2S or analogue microphone is used to capture the voice commands. The choice of microphone can be adapted based on the
specific requirements and constraints of the project.
4. TensorFlow Lite
•The voice recognition model, trained using TensorFlow, is deployed on the ESP32 using TensorFlow Lite. This allows the model to
run efficiently on the resource-constrained microcontroller.
5. Robot Hardware
•The robot’s hardware, including motors, sensors, and power supply, are controlled by the ESP32 based on the processed voice
commands.
7
PRODUCT DETAILS:
M IC R OP H ON E ESP32 S E RVO M OT OR S
8
PROGRESS
1. Project Understanding and Planning :
•Completed understanding the project requirements and objectives.
•Successfully planned out the project timeline and resources needed.
2. Preliminary Research :
•Conducted research on the necessary hardware including the ESP32 microcontroller and the
microphone breakout board.
•Researched on the functionality and setup of the hardware components.
3. Laying basements for Tensorflow:
• Started working on Tensorflow lite.
• Acquired public audio samples for data pre processing and training.
9
Thank You
21BEC1142-HARISH M
21BEC1294-ENIYAN KM
21BEC1154-HARIESH p
10
