i-CANe
I-CANe is an intelligent cane for the visually impaired that allows easier navigation on streets by identifying obstacles that are on ground or overhead. I-CANe will consist of a cane equipped with sensors and batteries to create a compact navigation device that can intelligently identify obstacles
2025-06-28 16:33:00 - Adil Khan
i-CANe
Project Area of Specialization Artificial IntelligenceProject SummaryI-CANe is an intelligent cane for the visually impaired that allows easier navigation on streets by identifying obstacles that are on ground or overhead. I-CANe will consist of a cane equipped with sensors and batteries to create a compact navigation device that can intelligently identify obstacles and notify the users with sounds. A GPS functionality will allow user’s location to be detected. The cane will allow users to identify obstacles with a single device, reducing the use of many different devices for on ground and overhead obstacles. We will be using computer vision techniques such as Artificial intelligence, image classification, object detection as well as sensors, to allow their simultaneous detection in real time.
Navigating the streets is a difficult task for the visually impaired. While the standard navigating cane allows users to detect obstacles on the ground through ultrasonic sensors, it does not detect shoulder level obstacles, neither does it give a detailed picture of the obstacles. Likewise, with the white cane. The existing canes do not offer complete obstacle detection or safety features like GPS tracking. Therefore, we propose a cane that offers a complete obstacle detection, identification through vision to allow user to better understand his surroundings for navigation. Moreover, the cane will have GPS tracking, which will allow family members to track location and ensure safety of the user.
The navigation process is controlled by Raspberry Pi. To improve the navigation Raspberry Pi is being used to help the visually impaired person. The obstacle is detected using ultrasonic sensor and the image is captured using camera. This is intimated to user using microphone.
Project ObjectivesThe aim of this project is to use a technology for a visually impaired person to help him in navigation, and the detection of obstacles while walking without any guide. The project will be expected to carry out the following tasks:
- To create an enhanced version of the navigating cane for the visually impaired to make him/her independent.
- Real time on-ground and side obstacles detection with sensors.
- Front obstacle detection through vision.
- Real time obstacles identification.
- Provide GPS tracking of user and GSM alerts.
The very basic goal of this system is to make our visually disabled user independent. The workflow is illustrated in Figure 3.1 as attached below. The workflow is divided into two parts; object detection and identification. Live feed is obtained through the raspberry pi camera in the form of images. Obstacle presence is detected through the ultrasonic sensors placed at various angles and places of the cane and their distance from the user is also communicated. Realtime images are sent to the cloud for computation and through proposed datasets and AI operations, objects are identified through the earpiece connected for the user. For constant connectivity a GSM module is added. The GSM module is also contributing to the SOS features in case of emergency to alert the users loved ones. The GPRS module is being used to provide location tracking services. Headphones are connected through the audio jack present in raspberry pi3B. Similarly, the Bluetooth and Wi-Fi being used are also built in the raspberry. To ensure maximum battery life and efficiency and external power unit is attached to the cane. The power unit comprises of rechargeable Lithium Batteries -18650. A 3 Cell Battery Management System will act as the heart of our charging circuit. It will regulate voltage across the module to ensure all parallel batteries stay at 3.7V no more and no less. Another component added to the power unit is the buck convertor to ensure that voltage conversion between high and low voltages are done efficiently, thus extending battery life. The goals of the power unit are to make it compact and modular. The I-CANe is charged through a 12V charger connected to a wall power outlet.
| Benefits | Limitations |
| Detection of obstacles of left, right, and front side. | Limited range. It will not be able to detect large buildings. |
| Identification of obstacle with AI | Not water-proof. |
| Independency and confidence | Only for outdoor detections and identifications. |
| Detection of ground level to shoulder level. | Light intensity limitations |
| Low cost. | |
| Voice feedback for user guidance | |
| GPS tracking | |
| Complete and compact system | |
| Optimised and long battery life. |
Benefits
Detection of obstacles of left, right, and front side.
Identification of obstacle with AI
Independency and confidence
Detection of ground level to shoulder level.
Low cost.
Voice feedback for user guidance
GPS tracking
Complete and compact system
Optimised and long battery life.Technical Details of Final Deliverable- 18650 Li-Ion Rechargeable Cells
- 3S 25A BMS Board The 3S 25A Battery Management System charge protection board ensures the security of the battery pack. It will protect the I-CANe during charging and discharging. It will monitor the voltage across all cells and balance the voltage between them accordingly to ensure equal levels of charging throughout.
- INA 219 DC Current Monitor For I-CANe we plan on utilizing the INA219’s ability to measure battery health and status. The voltage indicates the charge levels where as current indicates short circuits and potential sources of damage.
- LM2596 Buck Convertor For excellent load regulation and high efficiency, we chose the Dual USB output LM2596 Buck Convertor for I-CANe, it will help us achieve a stable fixed output of 5V with a continuous 3A.
- HC-SR04 Ultrasonic Sensor We propose to use four HC-SR04 ultrasonic sensors to provide maximum coverage for object detection for the user. It uses sonar to determine the distance to an object and can notify to the user accordingly for angles where the rpi camera does not cover. We aim to offer excellent non-contact range detection with high accuracy and stable readings.
- Raspberry Pi Camera Board v1. We propose to perform object identification through images but videos could be a method of incorporation in the future.
- TTGO T-Call ESP32 SIM800L is a development board that combines the GSM/GPRS chip SIM800L for SMS and Phone Calls with Wi-Fi and Bluetooth connectivity. The board ensures ADC conversions on low power, with a dual core processor.
- NEO 6M v2 u-blox The NEO 6M v2 u-blox is a GPS module that can track up to 22 satellites and identifies locations anywhere in the world. We chose this particular module as it consumes low power and inexpensive.
-
Raspberry Pi 3B+ The brains and central part of I-CANe is the raspberry Pi3B+. It has a quad core 64 bit CPU, Wi-Fi and Bluetooth connectivity. It has a 1GB RAM and easily connectable to headphones, RPi camera and touch screen – all components that are crucial to I-CANe. The Raspberry Pi 3B+ enables us to incorporate many features whilst still being cheaper than its new variants. We managed its computational heat by attaching two heat sinks to the board and adding an external RPi fan which can be removed later, if need be.
| Elapsed time in (days or weeks or month or quarter) since start of the project | Milestone | Deliverable |
|---|---|---|
| Month 1 | Literature Review | Implementation ideas and plan |
| Month 2 | Power unit | Schematic |
| Month 3 | Power unit | Hardware Implementation |
| Month 4 | Control Unit | Schematic and dataset practice installation |
| Month 5 | Control unit | hardware implementation |
| Month 6 | Data set | Dataset implementation |
| Month 7 | I-CANe | Prototype |