Intelligent Drone Dynamics Simulation

During a mission, if a drone carries out machine learning procedures to all the data collected by its numerous sensors, it might make moment-to-moment assessments around how to respond to its position. Proposed paper will use the Newtonian method to build the drone dynamics model. During problem study we require a treating idea that estimates atmospheric density amounts grounded on weather conditions and other physical numbers. So, the model can execute stable flight under the impact of environmental reasons, and can clearly prove the variation in the transfer of the drone under the impact of environmental influences.

As a basic was studied a proportional integral derivative (PID) controller depend on of a proportional unit, integral unit, and derivative unit. Through the simulation procedure, when a drone flight command is read, it is treated by the order interactive module first and is sent to the physical model and control model, which will update the drone’s motion condition based on the command information and environment condition.

In proposed research, the newly developed neural network must be trained first, as how an individual child being skilled to execute activities.

The MATLAB/Simulink environment was employed to perform the simulations. In this part, solutions are performed through altitude and attitude balance using step input and multi-level tracking, and position tracking.

This research has positively submitted the design and implementation of the auto-tuned PID controller using neural network in evaluation to manual tuned PID controller.