New CLBF framework ensures safe drone flight under temporal logic constraints

A team of researchers from an undisclosed institution (Ruikun Zhou, Yating Yuan, Haocheng Chang, Yinan Li, Yiming Meng) has introduced a new controller synthesis framework for continuous-time dynamical systems subject to Linear Temporal Logic (LTL) specifications and bounded control inputs. Their key innovation is the combination of sequential LTL task decomposition with formally certified Control Lyapunov-Barrier Functions (CLBFs). Rather than relying on explicit state-space abstractions — which are computationally expensive and often impractical for complex systems — the method breaks down global LTL tasks into a sequence of safe-stabilization subproblems. For each subproblem, offline-computed level sets of CLBFs guarantee satisfaction of local safety and stability constraints. The resulting switching feedback controllers then enable efficient online planning and dynamic re-planning, ensuring robust continuous satisfaction of the original temporal logic requirements even under state perturbations.

The framework was validated through both numerical simulations and a hardware demonstration on a Crazyflie quadrotor, a popular open-source micro aerial vehicle. The results show that the approach can handle bounded control inputs and provide formal guarantees without the scalability bottlenecks of abstraction-based methods. This work is particularly relevant for safety-critical autonomous systems — such as drones, robotic arms, and self-driving cars — that must reliably follow complex temporal logic missions while respecting physical limits. By avoiding state-space abstractions and leveraging CLBFs, the method offers a practical path toward deploying formally verified controllers in real-world environments where computational resources are limited and dynamic re-planning is essential.