Trajectory Candidate Grouping for Accurate Low-Load Object Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing tracking systems face challenges in balancing low computational load with high accuracy, particularly when sensor data errors occur, leading to decreased accuracy in target object tracking.
Innovation Solution
An information processing system that generates trajectory candidates from past sensor data, selects relevant sensor data based on correlation, and generates trajectory groups to enhance accuracy while reducing computational load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If many pieces of sensor data are referred to for tracking, then tracking accuracy is improved, but calculation load increases
Solution Approach 1:
The patent segments sensor data into multiple trajectories detected by different sensors, then processes each trajectory separately through generation, selection, and grouping stages. This segmentation allows the system to handle large amounts of sensor data in manageable units, reducing overall calculation load while maintaining tracking accuracy through comprehensive multi-sensor analysis.
Solution Approach 2:
The patent extracts only the necessary trajectory information from sensor data by generating trajectory candidates from past sensor data and selecting relevant candidates based on correlation. This extraction process filters out unnecessary data elements, reducing calculation load while preserving the essential information needed for accurate tracking.
2Device complexity
If trajectories detected by sensors are referred to instead of sensor data, then calculation load decreases, but tracking accuracy decreases due to undetected errors
Solution Approach 1:
The patent performs preliminary action by generating multiple trajectory candidates from past sensor data before final trajectory determination. This preliminary generation of candidates from raw sensor data allows the system to preserve original measurement information while reducing the complexity of real-time processing, as candidate generation can be performed with lower urgency.
Solution Approach 2:
The patent implements feedback mechanisms by selecting trajectory candidates based on correlation analysis and using this selection to inform subsequent tracking decisions. The feedback loop allows the system to correct errors by comparing multiple candidate trajectories and selecting the most consistent ones, thereby maintaining high tracking accuracy while working with processed trajectory information rather than raw sensor data.
3Measurement precision
If trajectory candidates from multiple sensors are correlated and grouped, then tracking accuracy is improved, but processing complexity increases
Solution Approach 1:
The patent merges trajectory candidates from multiple sensors into trajectory groups based on correlation analysis. By combining trajectories that show consistent patterns across different sensors into groups, the system achieves more accurate tracking through multi-sensor validation while managing processing complexity through systematic grouping rather than exhaustive pairwise comparison of all trajectory candidates.
Data Source
AI summary
At least one processor included in an information processing system carries out: an acquisition process for acquiring a plurality of pieces of sensor data; a trajectory candidate generation process for generating trajectory candidates related to a target object with reference to pieces of sensor data that have been acquired in the past; a selection process for selecting pieces of sensor data indicating a position of the target object from among the acquired pieces of sensor data; a trajectory group generation process for generating a trajectory group of the target object in accordance with a correlation between the trajectory candidates; and an output process for outputting the selected pieces of sensor data and the trajectory group.