Drone simulation software is recently more and more popular. So far we distinguish two main purposes for drone simulation:

• Research and development;
• Operator's training;

Many inexperienced operators found it useful to practice first, using virtual equipment, where any crash is virtually costless in opposite to the real devices.

On the other hand, most of the drone simulators of this class use a simplified physics model and graphics, that results in not so realistic UI in terms of simulation and graphics quality, yet good enough to practice operator's reflexes.
Use of this kind of simulators is obligatory in many countries when practising and preparing for professional UAVO certification, as the first step towards practical, outdoor part of the training.
Some simulators are composed just like a playground area for model flights while advanced present scenarios the virtual operator must implement (also validated over time and space, including time-limited and quality evaluated virtual exams). Simulators of this kind used to provide various airframes to choose between, different camera positions (one can observe from the ground, or use the onboard camera, simulating FPV flight, and so on. It is a common approach, that simulators are somehow open to connect any controller, as they used to be able to use regular Joystick/Game controller that is visible in the operating system. There used to be a hardware, mostly USB-based receiver that is capable to act like FC, where you can connect various Receivers thus use your own Controller. This approach introduces the natural latency of the RC link.

At the moment, there are multiple solutions, but two of the simulators are freely available thus they are popular among a variety of users:

• AirSim by Microsoft (still Open=Source): https://github.com/microsoft/AirSim; This simulator is also suitable for UGV, not only for UAV, and its purpose is mostly to implement and test high level, AI-related solutions for autonomous vehicles. AirSim can integrate with ROS-based hardware (Robot Operating System) solutions. As it is implemented using popular gaming Unreal Engine, the graphics is astounding yet it requires a powerful GPU to run. The simulator provides rich user experience with a natural-looking environment like cities, parks and so on.
• GAZEBO was developed as pure Open-Source simulator for a variety of physical devices, also for UAVs and UGVs, but not limited to: http://gazebosim.org/; Its origin is to simulate robots. It provides simple graphics, but the benefit is it can run on constrained hardware. The GAZEBO seamlessly integrates with the ROS environment. The simulator also provides ORCE rendering engine for a realistic environment rendering but the real power is simulator's flexibility on configuration and simulation running, starting from constrained Linux-based machines, finishing on Amazon AWS, cloud-based simulations.