Development of an algorithm to find loads on the floor relatively to the position of the robot

The Autonomous Guided Vehicles are a product that saw huge increase in the demand in the last decade. The reason behind this is due to the fact that they provide optimized solution to speed up production and reduce the costs. AGVs may di ers accordingly to the speci c task to ful ll or the building characteristics. InSystems Automation GmbH is a German company in Berlin that, among other elds, is investing considerable nancial resources into research and de- velopment of innovative solutions in this eld. These new research plans and solution often come from the customer desire or demand to acquire an auto- mated vehicle that would be able to ful ll a certain task. The research conducted in this work and thesis aims to develop an algorithm for one of those customized AGV called The Stacklifter. The project started two years ago from the idea to have a customized bot that it is able not only to recognize standardized objects from the ground, but to lift and re-locate them within a facility. The major di erence with what is currently available is the development of a customized software that should be able to ful ll the task, performing the recognition in a 2D environment, trying to avoid as much computation as possible as the robot is driven by PLC. It is actually this point that makes a huge di erence as the majority of detection algorithm run on PC and work in a 3D environment exploiting image recognition. However, compared to a PC, a PLC has a huge limit in term of computation, especially since this kind of system is designed to run 24/7.