Speaker
Description
In modern construction, precision surface manipulation is increasingly required in areas that are difficult to access or sensitive to excessive pressure, such as fine-tuning plaster or leveling floor surfaces. However, current methods often rely on heavy machinery that can damage delicate surfaces or struggle to reach constrained spaces, leading to inefficiencies and suboptimal results. This paper presents a novel drone-based platform for non-contact precision surface operations, designed to minimize direct surface contact while maintaining high levels of accuracy. The system features a hexacopter drone equipped with a lightweight aluminum rotating disc designed to uniformly distribute mechanical energy across the surface, facilitating fine adjustments and leveling. Stability and precision during operation are achieved through a two-layer flight control system. The first layer employs a Proportional-Integral-Derivative (PID) control algorithm to maintain consistent altitude, using real-time data from an Inertial Measurement Unit (IMU) and perception sensors. The second layer integrates an adaptive trajectory refinement algorithm informed by sensor feedback to optimize the planned path and posture, ensuring accurate compensation for environmental disturbances and internal system vibrations. To enhance coverage efficiency, the platform employs a predictive surface attention model, which prioritizes high-importance zones by analyzing fused perception data. Experiment in simulation environment demonstrated refinement completeness of 99.2% on smooth surfaces, 95.6% on moderately rough surfaces, and 91.3% on rough surfaces, with average height deviations below 0.05 meters and processing latencies under 152 milliseconds. These results highlight the system’s potential for scalable, precise, and efficient surface refinement in construction applications.
