Design, optimization and testing of Sliding Mode Controllers for an autonomous lunar nano-drone

The LuNaDrone by Evolunar is an autonomous lunar nano-drone with 6 degrees of freedom capable of carrying out specific missions on the lunar soil. Currently, the drone is at an advanced stage of development, with a well-defined physical structure and a working navigation algorithm. The aim of the thesis is to design and test a new controller for the LuNaDrone embedded in a pre-existing simulation environment. In the field of robust control, the Sliding Mode Controller is an excellent strategy in the presence of parametric uncertainties and disturbances. For this reason, the core of the work focuses on the design of SMCs for the LuNaDrone. In the introductory part of the thesis, the context of the work is outlined, starting with the purpose of the drone's missions: the exploration of the lunar lava tubes, still not in-depth known to the scientific community, or the transport of small payloads. We then specify the movements and the targets that the LuNaDrone must accomplish during each stage of its mission. Next, the drone’s dynamics are described. These are accurately modelled in the simulator by incorporating structural and dynamic constraints through kinematic relations and Euler’s equations. At this stage, a transformation matrix M is introduced, mapping the thrust inputs into the forces and moments used in the model plant. Then, the model is enhanced by integrating actuators and mass variation over time. The control inputs to actuators are properly handled by modulator blocks resulting in an ON/OFF actuation. Subsequently, the control strategy is designed. Sliding surfaces are chosen as combinations of state errors, then boundary layers and tunable parameters are designed. The controller’s stability is proven through Lyapunov theory, which guarantees the convergence of the system to the sliding surface in a finite time. Furthermore, different SMC versions are designed and analyzed. In the following section, the optimization of the model is described, focusing on parameters tuning through Genetic Algorithm in order to improve its performance under different operating conditions. The tuning results are presented and discussed. Finally, the thesis concludes with application hints of the newly developed controllers, including the possible integration of code generated in C on a specific embedded board. This offers a concrete use case scenario with the aim of giving further solidity to the overall project by creating an autonomous and reliable lunar drone.