Multi-UAV Cooperative Trajectory Planning With Dynamic Programming

This paper proposed a cooperative trajectory planning method with Dynamic Programming (DP) to address the real-time generation of safe trajectories for Multi-UAV in complex environments. The DP comprises three key components: (1) adaptive batch sampling of the state space to efficiently reduce computational complexity; (2) application of DP principles to compute feasible trajectories from the sampled states; (3) formulation of the trajectory optimization problem as a quadratic programming (QP) problem through spline curve parameterization, yielding optimal smooth trajectories. Extending the single-UAV planning paradigm, we introduce a distance-constrained information exchange for multi-UAV coordination. Through collaborative environmental sampling, each UAV gains access to extended spatial awareness beyond its individual sensor range, thereby expanding the solution space and enhancing the reliability and robustness of the planning outcomes. Experimental results demonstrate that our framework achieves real-time performance with computation times in the order of milliseconds. Comparative evaluations against state-of-the-art methods reveal superior trajectory quality, making it suitable for multi-UAV applications in complex environments.