EEG based smart devices control through IoT for paralyzed patients
The human brain is made up of billions of neurons that interact with one another via electricity. Electrical activity, often known as brainwaves or Electroencephalogram (EEG) signals, is a technique for measuring brain electrical activity. An EEG headset is used to record electrical activity. The EE
2025-06-28 16:26:56 - Adil Khan
EEG based smart devices control through IoT for paralyzed patients
Project Area of Specialization NeuroTechProject SummaryThe human brain is made up of billions of neurons that interact with one another via electricity. Electrical activity, often known as brainwaves or Electroencephalogram (EEG) signals, is a technique for measuring brain electrical activity. An EEG headset is used to record electrical activity. The EEG headgear converts data into packets and delivers them via a wireless media (Bluetooth / Wi-Fi). The signal are acquired with the help of an EEG headset which has a specific number of rows and columns of electrodes which are placed on the subjects scalp and there are 16 or 32 individual electrodes. These electrodes acquire the scalp level movement of the electrons and based on these activities thousands of signals are generated. These signals are extracted into a computer or likewise device through BCI (Brain Computer Interface).The next step is the preprocessing of signals, Preprocessing is the process of converting raw data into a format that is more suited for future analysis and user understanding. Preprocessing refers to the removal of noise from EEG data in order to come closer to the real brain signals. The noise produced is removed with the help preprocessing filters which include (Low-Pass filter, High-Pass filter, Band-Pass filter and Notch filter). Then comes the phase of feature extraction, EEG signals are complex. We use computers to apply intricate automated processing algorithms to EEG data, allowing us to retrieve 'hidden' information. There are several techniques, including time domain features (mean, standard deviation, entropy,), frequency domain features (Fourier transform, wavelets,), and finally synchronicity features, which examine the relationship between two or more EEG channels (coherence, correlation, mutual information, etc.). We have a standard set of frequency bands to differentiate them in groups and then use it on the basis of acquired data, The frequency bands are – delta: below 3.5 Hz (0.1–3.5 Hz), theta: 4–7.5 Hz, alpha: 8–13 Hz, beta: 14–40 Hz, and gamma: above 40 Hz. To make our model accurate over number of subjects we use machine learning to train our model. To train the model two methods are used which are Cross-validation using K folds, support vector. Now these classified signals can be used to control a number of applications based on the need of the user.
Project Objectives1. The main objective is to select a system that supports paralyzed patients by processing EEG signals and send the signals to distanced smart devices.
2. The secondary objective is to process the EEG signals to control external devices. As it is beneficial and is the most efficient form of smart devices control for a paralyzed person as it flawlessly work together towards helping a disabled person.
Project Implementation Method- We will use headset (Neurosky Mindwave or Muse 2) among these two headsets which gives us the most accurate signals. The NeuroSky MindWave Mobile 2 is an EEG headset that safely monitors and transmits the power spectrum (alpha waves, beta waves, and so on) data over Bluetooth Low Energy (BLE) or Bluetooth Classic to interact wirelessly with your PC, iOS, or Android smartphone. The device is used in our project to acquire brainwaves then we can use them in any of our applications, and delivering them to hardware.
- Put this headset on and see your brainwaves alter in real time! You can track your attention and relaxation levels with the MindWave Mobile 2.
- For the purpose of noise removal the acquired signals are preprocessed and the noise produced is removed with the help preprocessing filters which include (Low-Pass filter, High-Pass filter, Band-Pass filter and Notch filter).
- The signal acquiring process is set to 5 to 7 seconds long so that the accuracy of the signal would be high and amount of time taken would be considerable.
- To classify the signals we have a standard set of frequency bands to differentiate them in groups and then use it on the basis of acquired data, The frequency bands are – delta: below 3.5 Hz (0.1–3.5 Hz), theta: 4–7.5 Hz, alpha: 8–13 Hz, beta: 14–40 Hz, and gamma: above 40 Hz.
- The models are trained using machine learning and the methods used are Cross-validation using K folds.
- Data must be partitioned into subsets in order to verify the complete input data. The statistical process of dividing the sample data into k subgroups is known as k-fold cross-validation and the second method is support vector machine (SVM) builds a hyper plane. SVM classifier can handle high-dimensional data. Kernel is a decision-making method that may generate nonlinear decision boundaries] SVM is a linear two-class classifier in its most basic form.
- The number of models trained will be 3 counting each on an “ON and OFF” state which will be equivalent to training 6 models.
- To interface the processed and classified signal with the IoT devices we use open BCI (Brain Computer Interface) to send the processed signal to Adafruit server from there we will be using ESP-32 to control the appliances. The on and off signals are sent to control the devices through IoT.
- Paralyzed people are almost completely neglected in practically every aspect of life, and this is especially true in modern culture. A crippled person is unable to perform anything on his own and must always rely on the assistance of others. They see themselves as a burden to their families and society.
- The advantage of this technology is that it will allow them to govern their own life and accomplish things on their own, as well as interface with electronics and appliances in their houses via brain signals via an EEG headset.
- This project requires gathering data from numerous patients and assessing it in terms of age, gender, location, and patient history characteristics, among other things.
- The other benefit is to use EEG signals to operate home appliances, wheelchairs, TV remotes, and other devices with minor changes to the project architecture.
- Long-distance control of smart gadgets would be the most important advantage, allowing the crippled person to be self-sufficient.
Software:
Our final deliverable will be a trained machine learning model which will have six classes model. Adafruit is a platform for displaying, responding to, and interacting with data from your project. Adafruit provides us with libraries for a variety of programming languages, as well as user interface support which are built-in functions of the Adafruit. Dashboard will show the data using charts and graphs to allow us to make better decisions. We will be using Adafruit to send real time codes from the code script written in python language to the module ESP-32.
Hardware:
The IoT architecture will consist of the following components. ESP-32 to interface the devices with the adafruit server, relays to control the ON and OFF operation of the devices. Smart models which will be trained includes a fan, a TV and a bulb. Our final machine learning model and code will directly run on ESP-32 without the need of a computer. We can also save the data on cloud and send the control signal to devices directly. The ESP32 may work as a stand-alone system or as a slave device to a host MCU, lowering the communication stack overhead on the primary application CPU. Through its SPI / SDIO or I2C / UART interfaces, the ESP32 can communicate with other systems to offer Wi-Fi and Bluetooth capability.
Final Deliverable of the Project HW/SW integrated systemCore Industry ITOther Industries Education , Medical Core Technology NeuroTechOther Technologies Artificial Intelligence(AI), Internet of Things (IoT), Wearables and ImplantablesSustainable Development Goals Good Health and Well-Being for People, Industry, Innovation and InfrastructureRequired Resources| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 70000 | |||
| BT module (HC-05) | Equipment | 1 | 1000 | 1000 |
| Arduino | Equipment | 1 | 1000 | 1000 |
| Communication Modules | Equipment | 1 | 5000 | 5000 |
| Electrical Components | Equipment | 1 | 5000 | 5000 |
| Casing/Setup Cost | Equipment | 1 | 8000 | 8000 |
| LCD | Equipment | 1 | 1000 | 1000 |
| ESP 32 | Equipment | 3 | 1000 | 3000 |
| Neurosky headband | Equipment | 1 | 14000 | 14000 |
| EEG Headset | Equipment | 1 | 20000 | 20000 |
| Ten20 Paste Jars 3-Pack | Equipment | 1 | 1000 | 1000 |
| Biosensing Board | Equipment | 1 | 10000 | 10000 |
| Miscellaneous | Miscellaneous | 1 | 1000 | 1000 |