Object Mapping Estimation on Multi-agent Robot Using Omnidirectional Camera

Abstract

Object detection in a robot vision system is usually used to find out the objects around the robot. In addition, this technique can also be used to recognize the type of object. Beside this, it will generate a position of the object toward the robot and make the robot easier to know the object location and to do its job. However, to position the object in the robot's world model, a mapping process is needed. It requires good precision of object location to generate a good world model map. To do it close to the reality, an estimation can be used. The estimation is the process to predict where the location of an object. One of estimation method in mapping is intersection of two straight line. In this paper, we present the mapping of an object to the robot's world model based on intersection of two straight lines using omnidirectional camera in multi agent robot. It perform well with 6742.87 of MSE and 82.11 of RMSE.