Optimization of Targeted Aerosol Disinfection via Integrated Physical Modeling and B-Spline Trajectory Planning

Abstract Effective autonomous disinfection in healthcare and office environments requires a precise balance between disinfectant coverage uniformity and operational efficiency. Unlike traditional full-coverage methods that rely on room saturation, this paper introduces a system-integrated framework that leverages a physical deposition model to drive an optimized trajectory-planning strategy. The methodology identifies critical disinfection keypoints via spatial analysis and sequences them using a constrained Traveling Salesman Problem (TSP) tailored to the robot's navigable space. The core innovation lies in a unified mathematical architecture that decouples the robot's navigation path from the disinfection target trajectory through B-Spline representations. By integrating an anisotropic Gaussian model to account for the disinfectant's spatiotemporal behavior, the framework enables independent optimization of robot mobility and spray orientation. This architecture provides real-time adaptability to environmental constraints by allowing independent control of navigation and nebulization without the need for computationally expensive replanning. The framework was validated through experimental trials in real-world indoor scenarios using a custom-built robotic platform. Results demonstrate a markedly more uniform disinfectant distribution on target surfaces, with concentration variability reduced from approximately 33% to 6%, underscoring that room saturation does not inherently ensure concentration homogeneity across specific points of interest. Furthermore, the methodology reduces execution time by nearly 50% and decreases disinfectant consumption by more than 40% compared to a brute-force strategy.