Just wondering why and what is the main motive that you chose "Sliding mode controller" for the control of quad-copter.
Chosen of sliding surface is not an easy task (Just my opinion only). Further problems like chattering effect, stability, higher order sliding surface/model and many.
A robust tuned PID controller/modified or upgraded PID may provide a similar or comparable response as an SMC.
On the other hand, PID has many advantages, can't explain in a single paragraph. But in short sentence simple in design, easy to tune and implement. PID has problems also which can be improved in many ways without replacing the complete controller with other advanced or intelligent controllers.
The easiest way to alleviate the chattering is by using the boundary layer technique where in place of directly using signum function, we use a different user defined discontinuous function. You can look for this technique in the given paper.
Zhu, Zheng, Yuanqing Xia, and Mengyin Fu. "Adaptive sliding mode control for attitude stabilization with actuator saturation." IEEE Transactions on Industrial Electronics 58.10 (2011): 4898-4907.
The uncertainty effect can be due to the parameter variations in the system dynamics, surrounding disturbances or/& may be some fault in the measurement sensors or actuators which are not considered in the mathematical model dynamics. I mean to say that there are many sources from where these uncertainties might occur. If you are precisely modeling these uncertainties in the mathematical model then, you will be in a better position to handle these uncertainties using robust control technique.
It is for the user to decide what is their desired performance criteria.
In my opinion the better form of SMC could be adaptive integral higher order SMC but again it would be very complicated to design. It will have a sliding phase from the starting point t=0 (due to the integral part). Adaptive part will deals with the constraint about the apriori knowledge of upper bound of uncertainty. Higher order will take care of the chatting problem.
Moreover, one can not easily decide what is the best form of SMC because there are other criterias like finite time convergence, fixed time convergence of states, alleviating the overestimation of Adaptive gain.
Therfore, it is all boils down to the problem and the performances requirements of the controller.
The sliding mode controller works very nice with nonlinear system. One of the difficulties that you will face is to prove the stability of the designed controller using Lyaponove stability criteria. You can mix between linear and nonlinear controller. For me, I have used PD-SMC for the quadcopter stability and trajectory tracking. You can read my article in the following link.
Article A proportional derivative sliding mode control for an overac...
Thank you Ahmed Alkamachi ,but two questions for you.firstly what about your control input do you still have chattering?.secondly is it good enough to be implemented on a hardware without damaging the actuators ?
If you read my article, you can see that I have used a function that resolve the chattering problem. Also, I have tuned the SMC parameters using genetic algorithm and I have included the actuator limits in the cost function.
The main disadvantages of SMC are the chattering problem. The chattering is the natural price that the SMC pays in order to 1) eliminate completely the matched external disturbances 2) replace the original system by a new certain one, which created according to the desired features, and maybe with a lower order. This elimination and creation are the main advantages of SMC.