PRO-SPECT: Probabilistically Safe Scalable Planning for Energy-Aware Coordinated UAV-UGV Teams in Stochastic Environments

A team of researchers has published a new paper on arXiv introducing PRO-SPECT (Probabilistically Safe Scalable Planning for Energy-Aware Coordinated UAV-UGV Teams in Stochastic Environments). The work addresses a critical limitation in current autonomous vehicle team coordination: most planning algorithms either assume deterministic travel times or use fixed safety margins that can be overly conservative or unsafe. PRO-SPECT instead models travel times as random variables and provides mathematical guarantees that the probability of mission failure (specifically, the drone depleting its energy before reaching the mobile ground charging station) stays below a user-defined threshold, such as 1% or 5%.

The algorithm formulates the coordination problem as a Mixed-Integer Program and solves it in polynomial time, making it scalable for practical applications. Unlike previous approaches, PRO-SPECT supports both initial mission planning and dynamic re-planning when disturbances occur, all while maintaining the same probabilistic safety bound. The researchers provide theoretical proofs about solution feasibility and computational complexity, and demonstrate the method's effectiveness through numerical comparisons and simulations, showing it can generate safer, more efficient plans than deterministic or robust-but-conservative approaches.

This represents a significant advance for deploying autonomous vehicle teams in real-world scenarios like infrastructure inspection, precision agriculture, or emergency response, where environmental uncertainty is the norm. By moving from deterministic or fixed-margin planning to provably risk-bounded planning, PRO-SPECT enables more reliable and efficient operations of coordinated drone and ground robot systems in unpredictable conditions.