Hardware in the loop system development for modeling and control of multirotor vehicles

Unmanned air vehicles were designed to accomplish specific missions thats are dangerous to be fulfilled by man power. Nowadays the big companies want to reduce worker payments in companies budgets, so they tended to robotic applications. Not only big cargo but also restaurants or fast food companies are planning to carry out courier services via quadrotors for delivering cargo to costumers. Main reason for this changing is development of electronic devices, especially sensors, reduction of the prizes, and easily arrival UAVs parts. So, UAVs are going to be modern flying robots in the near future in all of the areas in mind. The topic of the thesis is development of hardware in the loop system to model and control of multirotor vehicles. Other words, we want to model and control multirotor vehicles on a testbed. Why testbed? The testbed gives some advantages. Reducing time consuming such as, eliminating the necessity of going to flight area, cheap workable conditions, and reduction the material and part breakage are most of them. This works is a sample of Model Based Design Process. Model Based Design (MDB) is a process that accelerating prototyping with cost-effective in a short time. MBD process phases are definition of requirement, mathematical modeling of the system(simulation), rapid prototyping, code generation,software in the loop (SIL), processor in the loop (PIL), hardware in the loop (HIL) and validation. First phase is definition of the system or product properties and requirements. The generating mathematical model of the system under boundary conditions effecting the system directly or indirectly in the simulation environment. Regarding to the mathematical model, rapid prototyping and code generation phase is coming to test the system on the simulations carry out on the computer environment that is the most important point to reduce cost and time consume. Software in the lop simulations are carried out to analyzing generated codes on the host computer. Processor in the loop is analyzing the system on target CPU on the embedded controller to reduction of the processor errors. Hardware-in-the-loop (HIL) simulation is a type of real-time simulation to test designed controller with real plant or system input out put via embedded hardware. HIL simulation shows how controller responds, in real time, to realistic virtual simulations. HIL can be use to determine if real physical system (plant) model is valid or not. The Gyroscopic testbed was designed for HIL tests of the quadrotor (IRIS+) for validating and implementing the designed different controller coefficients in real time environments. The HIL testbed is designed in order to how a quadrotor attitude control test easily. Before preliminary design, literature search is done and different test beds are found for both quadrotor testbed and motor propeller testbed. Then working principle of gyroscope is gave inspiration to design testbed. All of the designing processes are carried out on the computer aided drawing environment. Then production and assembly phase are done. Next problem is fixing the quadrotor on the testbench. A support part is designed and produced, then quadrotor vehicle is assembled on the testbench. The motor-propeller testbed is designed with same procedure of HIL testbench. Mathematical modeling of a quadrotor is second chapter of the thesis. Before the implementation of the controller, the simulation of the system give foreknowledge about behavior of the system. The mathematical model of the quadrotor is needed to simulate. Newton-Euler and Euler-Lagrange methods are used to derivate nonlinear model of the system. In the mathematical model, thrust and drag coefficients are used for refer propellers reaction to angular velocities. Some assumptions are taken into account during the modeling. The third chapter of the thesis is related with quadrotor and testbeds parameters estimation. IRIS+ quadrotor vehicle has four motor, one pair rotate clockwise, other rotate counterclockwise with propellers which are self locked type. IRIS+ quadrotor has cross type frame architecture that angle between front arms is one hundred twenty degree but back angle between back arms is one hundred forty degree. Cross type quadrotor has basically two advantages over plus configuration. first, motor need less power for pitch and roll motion and remaining power can be used for increasing translational motion. Second advantage is providing clear vision for camera with obtuse angle in between front propellers. The reference coordinate frames are defined related to cross type frame quadrotors. The quadrotor moments of inertias are founded by experimental studies.Blade element theory is used for calculating the propellers thrust and drag coefficients. Mathematical model of the HIL testbed is derived via Euler Lagrange energy method. The inner and outer circles affect the motion of the quadrotor seriously so differences of angels of the circles and quadrotor are taken into account to effect on moment of inertias. Designing flight controller for simulation and implementation is the last section of the thesis. Main aim of the section is comparison PID and feedback linearization methods outputs both in simulation and real time environments. In simulation, PID and Feedback Linearization Controller is applied on the both nonlinear model of quadrotor and HIL testbench nonlinear model in SIMULINK. PID responses are figured as phi, theta, psi, and Z parameters of the systems outputs. Feedback Linearization controller designed to eliminate nonlinear terms to stabilize the quadrotor vehicle attitude motion with different lambda coefficients. The last part of the simulations is designing feedback linearization controller for HIL testbed with optimum lambda coefficients of phi, theta, and psi. Lambda of Z is't needed because there isn't any motion in Z direction. Implementation of the PID and Feedback Linearization controllers are verified on the Pixhawk open-source autopilot hardware. Pixhawk code is redesigned to implement controllers in real time. The flight data is graphed in MATLAB to visualize motion with changes of angles for both PID and Feedback Linearization controller. Main difference between the results of the PID and Feedback Linearization controller is seen on the response of outputs. PID controller is tracking the signal with higher overshoots and bigger oscillations then Feedback Linearization controller response. But PID is faster than Feedback Linearization Controller. Both of the controller are stayed in effect of moments of inertias of HIL testbed circles. It is easily seen in long time delay on the time response especially in psi angle. Future works are planned to continue in this works line. Especially different nonlinear controller methods are to implement on airvehicles. Moreover, code generation on the autopilot hardware are planned on the UAVs.