Robust moving horizon planning for multi-vehicles area coverage in uncertain environment using mixed-integer-programming

The growing reliance on multi-vehicle systems in various applications necessitates safety, efficiency, and adaptability. Traditional approaches often struggle to balance coverage performance with real-time adaptability and computational efficiency, particularly in uncertain or dynamic environments. To address these challenges, this work proposes a robust moving horizon planning algorithm for multi-vehicle area coverage in uncertain environments using mixed-integer programming. The proposed algorithm enables vehicles to dynamically adapt to wind disturbances and vehicle uncertainties. It systematically incorporates vehicle dynamics and constraints, such as obstacle avoidance, to optimize a performance index that considers uncovered area and energy consumption. This work also proposed a novel tuple encoding to enhance the computation efficiency of the proposed planning algorithm. Moreover, it incorporates a constraint-tightening approach to provide a theoretical guarantee on constraints satisfaction despite the vehicle uncertainties. The algorithm is evaluated under three distinct planning architectures: centralized, decentralized, and distributed, to assess their effectiveness in various scenarios. The algorithm robustness is validated through rigorous theoretical analysis and numerical simulations across diverse uncertain scenarios. Results demonstrate that the distributed planning architecture outperforms centralized and decentralized approaches in terms of coverage rate, computational efficiency, and robustness to environmental disturbances. This makes it particularly well-suited for real-world applications requiring rapid and reliable multi-vehicle coordination. This work provides a scalable and adaptive framework for multi-vehicle systems operating in complex and uncertain environments, offering significant advancements in real-time planning and decision-making capabilities.