Wednesday, 31 January 2018

Self navigation with Haptic Feedback System

Visually impaired people face challenges while they travel. A white cane is supposed to help in such circumstances which enables the visually impaired people to navigate on roads. But this cane is inefficient since it cannot detect low hanging tree branches or obstacles above the waist level. To overcome the above problem, the following projects employ a method which intimates the user of obstacles above the waist level.

It consists of a pair of camera sensors which gives the signal to the controller which in turns regulates the vibration motors (which are affixed to a vest) using the motor drivers.
This mechanism is achieved by processing the real-time images captured by the camera sensors and then controlling the vibration motors array accordingly (for eg. the vibration motors on the left side of the vest would turn ON if the object in the front is on the left of the user).
The software algorithm employs following libraries:
I) OpenCV: for image processing
II) TensorFlow: for identifying objects


Image1

The user holding the prototype

Things to be added in the Future Release (Audio Feedback)


Block Diagram



Self navigation with haptic feedback system

Visually impaired people face challenges while they travel. A white cane is supposed to help in such circumstances which enables the visually impaired people to navigate on roads. But this cane is inefficient since it cannot detect low hanging tree branches or obstacles above the waist level. To overcome the above problem, the following projects employs a method which intimates the user of obstacles above the waist level.
It consists of a pair of camera sensors which gives signal to the controller which in turns regulates the vibration motors (which are affixed on a vest) using the motor drivers.
This mechanism is achieved by processing the real time images captured by the camera sensors and then controlling the vibration motors array accordingly (for eg. the vibration motors on the left side of the vest would turn ON if the object in the front is on the left of the user).
The software algorithm employs following libraries:
I) OpenCV: for image processing
II) TensorFlow: for identifying objects