High-precision dynamic measurement based on trajectory jump analysis and coordinate sequence processing

Aiming at the challenges that traditional dynamic measurement methods are insensitive to complex trajectory anomalies and conventional filter fitting easily loses motion details, this study proposes a high-precision dynamic measurement method that combines trajectory jump analysis and coordinate sequence processing. This method uses the isolation forest algorithm to automatically identify trajectory jump abnormal points and quantify trajectory deviations. The Transformer model is used to capture long-term motion trends. Finally, the two types of information are fused to generate high-precision measurement results. When the data volume reached 5,000 points, the jump analysis accuracy was 0.97. When the sample batch size was 16, the training time reached 122 s. In actual application, when the global timestamp was 100 s, the highest trajectory error reached 0.14m. At 30 dB, the error increase reached 5.2%. When the rise time reached 0.5 s, the optimal overshoot was 5.1%. The research has higher jump analysis accuracy and calculation speed, as well as better robustness and dynamic response capabilities, providing reliable technical support for industrial automation, intelligent manufacturing and other fields.