Smart Chair
More and more common activities are leading to a sedentary lifestyle forcing us to sit several hours every day. The postures of any individual can reveal their sitting habit, mood, and even predict health risks such as pressure ulcers or lower back pain. Mining the hidden information of the po
2025-06-28 16:35:08 - Adil Khan
Smart Chair
Project Area of Specialization Artificial IntelligenceProject SummaryMore and more common activities are leading to a sedentary lifestyle forcing us to sit several hours every day. The postures of any individual can reveal their sitting habit, mood, and even predict health risks such as pressure ulcers or lower back pain. Mining the hidden information of the postures can reveal their wellness and general health conditions. In-seat actions contain significant hidden information, which not only reflects the current physical health status but also can report mental states. Considering this, we design a chair, based on a pressure detection module (deployed on the seat), to recognize and monitor in-seat activities through sensor. The individual is asked to perform and interact with different objects, his/her data is recorded and then analyzed for activity recognition and object recognition being used to perform the activity. Our results show that the proposed method, by fusion of time- and frequency-domain feature sets from all the different deployed sensors, can achieve high accuracy in recognizing the considered element of HOI.
Project ObjectivesAim
The primary aim of the project is to design a smart chair that can detect and correct posture of a person.
Objectives:
- Design hardware of the chair.
- Data acquisition and feature extraction.
- Object recognition.
- Activity recognition.
- User identification.
- Posture detection.
- Emotional state detection.
Hardware design
In the first phase, a chair is designed that consists of pressure sensors (FSR-406). Two cushions i.e., sitting part and backseat of the chair are equipped with pressure sensors. The sensing system includes a total of 12 pressure sensors in which 7 sensors are deployed in the bottom cushion, while 5 sensors are embedded in the backseat of the chair. The sensing module data is sent to Arduino based circuit design to store the data as a comma separated values (csv) file. The sampling rate of each pressure sensor is 10 samples/second. These sensors collect the data generated by the pressure of body while performing different activities.
Data Acquisition
A total of 8 participants (5 male and 3 female) were involved in this study with age range of 19-23 years with an average age of 21 years. The data were acquired for two cases: (i) while interacting with multiple objects (for object identification) and (ii) while performing activities with the object (for activity recognition). For object identification, the participant was asked to interact with two different objects i.e., book and a laptop for 30 seconds while sitting on the smart chair. Then we recorded the pressure sensor data during this span of interaction. A total of 10 trials are made for each participant. A gap of 10 seconds is given between each trial. In the second phase, the participant is asked to perform activities with the objects. A total of 10 trials are made for each activity.
Feature Extraction
Following features are extracted :
Fast Fourier transform, Principal Components, Mean, Root Mean Square, Variance and Standard Deviation.
Object Identification
In the first phase it is tested whether an object can be automatically classified while examining individual’s posture. A highest classification accuracy of 98.75% is achieved using MLP classifier for recognizing two objects using pressure sensor based smart chair.
Activity Recognition
Based on object identification, we further investigated the performed activities with the objects using same set of features and classifiers. The average classifier accuracy for activities performed with the laptop i.e., open, type or close . MLP again achieves the highest accuracy rate of 78% as compared to NB, SVM, and J48 for three activities.
User Identification
Furthermore, we also investigated whether users can be identified while interacting with an object. Since, user identification is useful for biometrics and customized settings. Here we only investigated individual differences while interacting with a laptop. Same set of classifiers are applied to recognize users while interacting with the object.In this case, it is observed that NB classifies the users of laptop with a highest accuracy of 98.75% as compared to MLP, J48, and SVM.
Benefits of the ProjectFollwoing are the benefits of the projects
- The chair identifies the posture and gives the correct posture so that the user can avoid chances of physical pain or any other disorder.
- It can be used to identify User's, recognise their activities and also the object being used to perform the activity.
- The data collected from postures of people can be used for research purpose.
The final deliverable will be a smart chair having pressure sensors which will collect data whose features will be extracted and present posture of user will be identified. Then we will train a machine learning model accroding to the correct posture so that the chair also tells the user correct posture. Also, we are planning to use MUSE brain device that records EEG signals so that we can predict emotional state of the person sitting on the chair. we will make GUI desktop application to display the output of detected posture and correct posture. If possible we will make android application for the same purpose.
Final Deliverable of the Project Hardware SystemCore Industry ITOther IndustriesCore Technology Artificial Intelligence(AI)Other TechnologiesSustainable Development Goals Good Health and Well-Being for PeopleRequired Resources| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 75000 | |||
| MUSE 2: Brain sensing headband | Equipment | 1 | 48000 | 48000 |
| Pressure sensor FSR 406 | Equipment | 14 | 1000 | 14000 |
| Arduino mega | Equipment | 1 | 1500 | 1500 |
| Chair | Equipment | 1 | 6000 | 6000 |
| Cushions | Equipment | 2 | 250 | 500 |
| Stationary | Miscellaneous | 2 | 1000 | 2000 |
| Printing | Miscellaneous | 30 | 100 | 3000 |