This is a group project done in the 3rd year of our Computer Science Degree Program in University of Colombo School of Computing (UCSC). The concept is to automate a vehicle which transport goods inside which uses same path each time (Eg. Human Driven forklifts used in big companies/workshops like automobile companies). The system is given a destination to move the vehicle into. It first finds the optimal path from the source to the destination. If any obstacles are detected the vehicle stops there and corresponding path is removed from the system so in the next time the vehicle will be moved in an alternative path. An image stream is sent from the vehicle to the computer using a wifi connection. The system recognizes the symbols (numbers here) placed in special places like Bends and Junctions and using a Neural based OCR system it recognizes the numbers. Then it commands the vehicle to move according to the place identified (Turn left if 3 is identified) and guides the vehicle to move from the source to the destination. Those commands are sent through a serial port from the pc, which are sent to the vehicle via a wireless connection using FM modules. The PIC in the vehicle understands the command and drives the motors accordingly. This is not a Path following robot and completely depends on Image Processing and Computer Vision. No sensors are used. All the circuits other than the FM modules are hand made. The project was done by 5 of us namely Sanjaya Amarasinghe, Manupriya Banu, Yuwakanth Rengasami, Pasindu Jayasekara and Manoj Pussawela. (Video edited by Banu, taken at the 3rd year lab, University of Colombo School of Computing (UCSC) in the day of final demo 20th April 2012)

Be the first to comment