The recent advancements in aerial robotics and control systems have significantly enhanced the capabilities of drones, particularly in tasks requiring precise maneuvering and stability. Innovations in the integration of robotic arms onto quadcopters have led to improved package delivery systems, with a focus on optimizing both mechanical design and control algorithms. Notably, the development of Proportional-Integral-Derivative (PID) controllers has shown superior performance in trajectory tracking and payload handling, making them a preferred choice for stable flight operations. Additionally, the introduction of learning-based approaches, such as Neural Predictors, has revolutionized the modeling of payload dynamics, enabling more accurate and adaptive control strategies. These predictors have demonstrated substantial reductions in estimation errors, outperforming traditional methods with fewer samples. Furthermore, the field has seen the emergence of fast Physics-Informed Model Predictive Control (PI-MPCS) surrogates, which offer a computationally efficient alternative to traditional MPCs, ensuring stability and robustness in control tasks. These surrogates have shown to be particularly effective in resource-constrained environments, providing rapid and precise control approximations. Lastly, risk-averse control frameworks have been developed to enhance the reliability of MPC algorithms in adverse conditions, particularly in vehicle control where model mismatches can lead to instability. These frameworks leverage uncertainty modeling and sample-based approximations to ensure reliable performance even under challenging conditions. Overall, the integration of advanced control techniques and machine learning models is driving the field towards more efficient, adaptive, and reliable drone operations.