This project “Corridor Navigating Autonomous Wheelchair” aims at providing corridor navigating capabilities to powered wheelchair using imaging and proximity sensors. The system analyses each frame of corridor captured by the image sensor mounted on wheelchair and generates appropriate guiding control signals to the motors of the wheelchair. The proximity sensors mounted on the device help in obstacle detection and collision avoidance and also aid in he locomotion. The system also contains manual control which can bypass the automatic controlling mechanism in case the automatic system fails or the user wishes to do so.
The project consists of two versions, one with PC control and the other with FPGA control. The
PC controlled version of the system has been successfully built and tested. Several simulation
tests have been conducted for the FPGA version of the system.
The objective of this project is to control the motion of a wheelchair as it navigates itself in a given location avoiding any obstacle, if encountered. Joystick and gesture based navigation using accelerometer are also included for navigation in manual mode, In automatic mode,Dijkstra's algorithm calculates the shotest path for navigation, Fuzzy logic control algorithm is implemented for the velocity control of wheelchair. Stereoscopic vision mechanism is used for obstacle detection.
The test results varified that the wheelchair navigated effectively in the preloaded map. Obstacle detection using stereo camera was achieved- the obstacle avoiding algorithm took around 3 seconds to resolve direction of movement upon obstacle detection. The velocity control was also achieved successfully. This project shows promise for future automatic navigating vehicle in large industries, hospitals and on the roads as well.