Well-functioning autonomous robot solutions heavily rely on the availability of fast and correct navigation solutions. The presence of dynamic/moving objects in the environment poses a challenge because the risk of collision increases. In order to derive the best and most foreseeing re-routing solutions for cases where the planned route suddenly involves the risk of colliding with a moving object, the robot's navigation system must be provided with information about such objects' positions and velocities. Based on sensor readings providing either 2-dimensional polar range scan or 3-dimensional point cloud data streams, we present an efficient and effective method which detects objects in the environment and derives their positions and velocities. The method has been implemented, based on the Robot Operating System (ROS), and we also present an evaluation of it. It was found that the method results in good accuracy in the position and velocity calculations, a small memory footprint and low CPU usage requirements.