Virtual Painting Teaching Points for Reproducible Robot Trajectories

The patent changes the decimation criterion from angle-based to curvature-based. By calculating curvature values at each position along the operation trajectory and comparing them against a threshold, the system adapts the decimation process to preserve positions with significant trajectory changes (high curvature) while removing redundant positions in gentle arcs (low curvature). This parameter change resolves the contradiction by making decimation sensitive to the actual geometric characteristics of the trajectory rather than using a fixed angular criterion.

The patent introduces dynamic adaptation in the decimation process by continuously evaluating curvature values along the trajectory. The system dynamically adjusts which positions are retained based on local trajectory characteristics, rather than applying a static decimation rule. This dynamic approach allows the system to maintain high reproducibility for both turn-around points and gentle arc shapes by adapting to the local geometry at each position.

The patent extracts only the essential information needed for trajectory reproduction by applying curvature-based decimation to the continuous position data. Instead of processing all collected position information, the system identifies and retains only those positions that contribute significantly to trajectory accuracy (where curvature exceeds the threshold). This extraction principle reduces data volume and processing complexity while maintaining the necessary measurement precision.

The patent performs preliminary processing of the continuous position data by calculating curvature values and identifying key positions before generating the final operation program. This preliminary action of filtering and selecting critical positions reduces the complexity of subsequent processing steps while preserving trajectory accuracy, as the decimation is performed in advance based on geometric criteria.