ARTIFICIAL VISION FOR AUTONOMOUS ROBOT HANDLE DISPLACEMENT CONTROL

摘要

Autonomous systems are increasingly becoming part of our daily routine - from the air-conditioning system that automatically controls temperature to robots that perform tasks independently. However is a big challenge to have a system, as in the last example, that are able to self-orientate and perform certain tasks in an autonomous way since the complexity of decisions and unknowns that comprises this environment and must be addressed. This article proposes a model of integration of artificial vision in controlling of the movement of a robotic manipulator in an autonomous way, where it will be able to create its own trajectory as the task to be performed requires and it can execute it with the ability to alter the trajectory in case of any obstacle or changes in the path already created. In this context, this set becomes a useful tool as in the production system, where there is a quest for flexibility and agility in the development of certain task as well as for ordinary people, in the future, performing tasks in their home. Based on a theoretical reference it is proposed a conceptual model where it is presented the integration between an artificial vision system and a robotic manipulator. In a second phase this model is implemented and submitted to a case study where the system should goes through a labyrinth with any trajectory finding a solution to it. In the development of the case study it was verified that the system had a satisfactory and robust behavior, performing all the moves that were requested even though interventions had to be initially done in order to correct the model. The paper has as main contributions: i) the development of a tool that provides flexibility and autonomy to the productive system, being able to adapt to unscheduled changes ii) the possibility of aggregation of other tools to the set with a shorter time of preparation; iii) the viability of appliance to the existing system.