Robotic systems nowadays become larger and more complex everyday. Due to this does the planning phase of these systems become more important as well, as it is no longer an option to just start working and see where the system will end up. New technologies in the form of simulations have been around for a while, but usually have been used for training purposes rather than for prototyping and testing, while a wide variety of robot simulators already exist such as VREP, Gazebo, 20-sim and ARGoS. Testing robots is a challenging task compared to testing software, whereas software testing usually results into a true or false answer are tests of robots more in a grey area where multiple ranges of answers can be correct or incorrect depending on environments, mechanical properties or user constraints.
My research consists of finding an elegant solution to these testing issues by making use of a simulation of the robot. Using an intuitive way of interfacing with the user for testing, the need for complex software understanding is removed. The testing framework should be as standardized as possible to allow for widespread use and less dependent on the robot at hand. The contribution of this research will be in the form of software that can interpret what the user wants to test within what constraints, automatically runs the simulator while testing for the given properties and eventually returning useful test data to the user. This allows for more in-depth, faster and more flexible ways of testing in comparison to testing on a physical setup. It opens up opportunities to test parts of systems that are otherwise hard to track or hard to reach, reducing the time spent in the testing phase and increasing the overall reliability of the system.