Real vs. Simulated: Analyzing Implementation Algorithms for ROS
GPS navigation is utilized in a wide scope of applications from providing vital information for rescue efforts to the navigation of robots built upon a robot operating system (ROS). The critical aspect of this research is to perform efficient point to point navigation with real-time GPS data. The capabilities of a robot using ROS is limited in that an extended kalman filter is used to provide state-estimation information for the internal GPS navigation to perform efficient path planning. GPS signal can be subject to interference from external sources regardless of where the data is being sent to and thereforeraises possible expectations of error (i.e. noise) in the form of accuracy and precision. It is then imperative to implement an algorithm that would produce newly filtered GPS data that is processed to perform point to point navigation in a robot. The filtered GPS data will then be mapped for comparison between a simulated environment and the University Quad at Cal Poly Pomona. Since the use of a simulated environment can cause little variation in expected results, this data set is expected to be close to the desired outcome of the implementation and filtering algorithm. By analyzing the simulated vs real data sets, the research intends to produce a more efficient method of implementing GPS data for ROS systems.