Vision based Object Recognition of E-Puck Mobile Robot for Warehouse Application

Abstract

At present, most warehouses still require human services for unloading of goods. Unloading of goods requires a continuous system to ensure the quality of work productivity. Therefore the need of autonomous robot system in warehouse is needed to improve the quality of work. Thus, a localization and recognition algorithm is developed and implemented on the E-puck robot. The task involves the recognition of desired object based on their colour (red and blue) and locating the desired object to the target location (marked by green marker). In addition, the collision avoidance algorithm is also developed and integrated to allow the robot manoeuvre safely in its working environment. The colour histogram technique is used to recognize the desired object and the target location. Based on the experimental results, the developed algorithm is successfully fulfilling the pick and place requirement with success rate of approximately 70% in simulation study and 50% in real implementation.