Time Data Sequence Smoothing via Global Local Trajectory Integration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing smoothing technologies for time data sequences face challenges in achieving real-time smoothing with minimal time delay and computation overhead, particularly in applications like mouse cursor tracking, where they struggle to balance global and local trajectory trends and often result in jitter or significant errors at turning points.
Innovation Solution
A method and apparatus that smooths time data sequences by determining reference points for global and local trajectory trends, calculating trajectory directions, and adaptively calculating new data points using these directions, thereby integrating global and local trends to reduce computation and maintain smoothness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If complex smoothing algorithms and iterative techniques (curve fitting and regression error analysis) are used to smooth data points, then smoothing effect is improved, but computation amount and memory requirements increase, causing time delay
Solution Approach 1:
The patent divides the smoothing process into two independent stages: a fast first-stage smoothing using simple moving average to reduce computation time, followed by a second-stage refinement using curve fitting only where necessary. This segmentation allows the system to achieve good smoothing effects without being bottlenecked by complex calculations on all data points simultaneously.
Solution Approach 2:
The patent applies a preliminary simple moving average smoothing operation before the more complex curve fitting process. This preliminary action pre-processes the data to reduce jitter and stabilize the trajectory, so that subsequent curve fitting operations require less computational effort and can be performed more quickly, reducing overall time delay.
2Manufacturing precision
If curve fitting and regression analysis are performed to obtain best curve and orientation prediction, then smoothing effect is improved, but computation amount increases
Solution Approach 1:
The patent applies partial curve fitting only to selected segments of the trajectory data rather than performing comprehensive regression analysis on the entire dataset. By applying curve fitting selectively to portions of the data where it is most needed, the system achieves sufficient trajectory accuracy while significantly reducing the overall computation amount required.
Solution Approach 2:
The patent segments the trajectory data into multiple segments and applies different processing strategies to each segment. Simple moving average is applied to all segments for initial smoothing, while curve fitting is applied only to selected segments that require refinement, reducing the total computation amount while maintaining trajectory accuracy.
3Speed
If prediction methods are used to control mouse cursor with current speed and direction, then response speed is improved, but error increases at turning points
Solution Approach 1:
The patent incorporates feedback mechanisms that continuously monitor the actual trajectory and compare it with the predicted trajectory. When deviations are detected, particularly at turning points, the system adjusts the prediction based on actual user input, correcting position accuracy while maintaining responsive speed through iterative refinement.
Solution Approach 2:
The patent dynamically adjusts the smoothing and prediction parameters based on the detected motion characteristics. At turning points or when high precision is needed, the system increases the weight of recent data points and adjusts the prediction model to account for changes in direction, thereby maintaining both response speed and position accuracy under varying conditions.
Data Source
AI summary
According to one exemplary embodiment, a smoothing method for time data sequences is performed, which includes: smoothing original data points of a time data sequence, then from smoothed data points, determining reference points for a global trajectory and a local trajectory; according to the reference points, calculating a direction of the global trajectory and a direction of the local trajectory; and adaptively calculating new data point(s) by integrating the two directions. The trajectory of new data points maintains the properties of smoothness and real-time perception.