This thesis aims to examine critical components in human-vehicle interaction for unmanned aircraft system (UAS) and develop strategies for improved interaction during uncertain events in the operations. Human-vehicle interaction paradigm for the UAS poses unique challenges than other vehicle platforms. To maintain an “equivalent level” of safety and awareness in the presence of sensory and vestibular isolation, a human pilot is required to exhaust their cognitive ability in many cases, especially in the event of uncertainty. In aviation operations, uncertainty is certain. In this work, we consider both visual line of sight (VLOS) and beyond visual line of sight (BVLOS) operations. We identify two unique uncertain events that affect these two unique unmanned aerial operations and aim to investigate and improve the interaction strategies. In particular, for BVLOS large UAS operations, we take the Detect and Avoid (DAA) application and recognize that the command and communication link plays the most critical role in the remote pilot’s situational awareness as well as decision-making. We implement a decision-theoretic approach to effectively increase the remote pilot’s contribution within safe operating conditions and enhance the human-machine user experience. Our results demonstrate noteworthy improvement in the pilot command reception without inducing any near-mid-air collision (NMAC) instances and enhancement of pilot involvement in decision-making in the encounter resolution. For small UAS VLOS operations, we address the rising operational and navigational challenges due to turbulent wind. We introduce a wind-aware small unmanned aircraft system human-in-the-loop simulation pipeline for operations in low-altitude urban airspace. We design the wind-aware user interface and extend a simulator to integrate spatial-temporal atmospheric boundary layer wind data. Human-in-the-loop experiments are conducted to assess the usability of wind information displays. Results from the ANOVA analysis indicate that wind-aware display significantly reduces (p − value < 0.05) pilot cognitive workload and significantly improves situational awareness in all wind conditions. We also observe significant improvement in mission and performance with wind-aware user interface. Motivated by the responses from subject matter experts we also design disturbance rejection controllers to mitigate wind effects on sUAS navigation. We integrate high-fidelity rotor dynamics into the simulator and learn error dynamics due to the turbulent effect using a Koopman operator. The Koopman-based model captures the highly nonlinear dynamics in quadcopter trajectory tracking in a linear form. We then employ MPC controllers with the learned dynamics and evaluate their performance using Monte Carlo simulations. The stochastic MPC, which treats the wind as a combination of a mean component and a random component, performs the best and reduces the mean tracking error significantly.