The proliferation of autonomous technology has opened new opportunities for safety-critical operations involving autonomous vehicles, particularly in scenarios where these vehicles coexist with humans. A knowledge gap exists in constructing safety constraints using sensor data to facilitate user-assignable control design. Our efforts have tackled issues related to geofencing, leader-follower formation, and obstacle avoidance for autonomous aerial vehicles. Notably, existing safe control methodologies heavily depend on computational methods to incorporate safety constraints into control design. Additionally, qualitative performance measures lack parametric representation in the generated control signal. Consequently, any alterations in system dynamics or obstacle properties necessitate control recalculation. In response to these challenges, we leverage recent advances in control design based on barrier functions to plan and track safe trajectories by isolating obstacle data from control design, establishing a clear connection between performance criteria and control formulation. We interpret safety requirements as physics-based control barrier functions with closed-form expressions. Extensive experiments and simulations have demonstrated the effectiveness of our proposed safe control for unmanned aerial vehicles.

 

Unoccupied Aerial Vehicle (UAV) swarms are increasingly deployed in dynamic and unpredictable environments, making their operational resilience a critical research focus. However, current simulation models for evaluating swarm robustness often oversimplify real-world disruptions, limiting their effectiveness in predicting and mitigating failures. This research introduces several disruption modeling approaches that enhance UAV swarm simulations by incorporating advanced aerodynamic and environmental disturbances, such as wind turbulence, obstacle interactions, and network intrusions. Using a combination of 2D and 3D grid-based simulations, we analyze swarm behavior under varying disruption conditions, including induced airflow variations and complex obstacle geometries. Results indicate that existing UAV control algorithms exhibit significant performance degradation when subjected to accurately modeled disruptions, highlighting a gap in current resilience strategies. By refining disruption simulations, we enable a more realistic assessment of swarm stability, energy consumption, and mission success rates. This talk will present key findings from our research, discuss the implications for swarm adaptability, and explore future directions, including heterogeneous swarm configurations to enhance resilience. Our approach provides a critical step toward developing UAV swarms capable of maintaining operational efficiency in real-world scenarios, advancing their reliability in disaster response, surveillance, and infrastructure monitoring applications