Radar Object Tracking via Multi-Frame Tracklet Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Radar systems face challenges in tracking objects due to the difficulty in semantic segmentation of radar frames, where reflection points are not continuous and vary between frames, making it hard to accurately identify and follow objects over time.
Innovation Solution
A method that detects detection points in radar frames, forms tracklets by associating these points over multiple frames, and groups them into object-tracks using feature parameters, employing a dynamical system model for prediction and correction to improve tracking accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If semantic segmentation is performed on radar frames to identify objects, then object identification capability is improved, but the difficulty of detecting and measuring increases due to discontinuous reflection points
Solution Approach 1:
The patent segments the semantic segmentation task into two stages: first detecting reflection points (discrete elements), then associating them into tracklets and object tracks (continuous representations). This segmentation allows handling the discontinuous nature of radar reflections while achieving continuous object representation through temporal association.
Solution Approach 2:
The patent performs preliminary detection of reflection points before performing semantic segmentation. By first identifying discrete reflection points and then associating them across frames to form tracklets, the system prepares the data in advance to facilitate more reliable semantic segmentation despite the discontinuous nature of radar reflections.
2Measurement precision
If reflection points are detected in each radar frame, then detection precision is improved, but reliability decreases due to variation of reflection points between frames
Solution Approach 1:
The patent merges detection points across multiple radar frames by associating them into tracklets. Individual detection points are combined with their temporal counterparts to form consistent object representations, thereby improving reliability while maintaining the precision of individual detections.
Solution Approach 2:
The patent establishes continuous object tracks by maintaining association of reflection points across multiple frames. This continuity ensures that even though individual reflection points may vary or disappear between frames, the object representation remains consistent and reliable over time.
3Measurement precision
If multiple detection points are associated to form tracklets, then tracking accuracy is improved, but device complexity increases
Solution Approach 1:
The patent segments the tracking process into distinct stages: detection point association to form tracklets, then tracklet association to form object tracks. This segmentation reduces overall complexity by breaking down the complex tracking task into manageable sub-tasks with well-defined algorithms.
Solution Approach 2:
The patent employs dynamic association criteria that adapt to the tracking context. The system dynamically determines which detection points to associate based on their temporal and spatial relationships, allowing flexible and accurate tracking without requiring complex fixed rules.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enables robust and accurate tracking of objects in radar frames by associating detection points to tracklets and object-tracks, reducing false associations and improving the performance of the tracking solution, even in environments with specular reflections and varying reflection points.
Implementation Method 1
Radar systems, basically, transmit (emit) radar signals into the radar system's field of view, wherein the radar signals are reflected off of objects that are present in the radar system's field of view and received by the radar system
Implementation Method 2
radial velocities are measured by utilizing the frequency shift caused by the doppler effect
Data Source
AI summary
Detection points may be first observed over multiple radar frames. The observed detection points, which form the tracklets, can be used to form a segmentation of the present radar frame by associating the tracklats to at least one object-track, which represents at least one object based on at least one feature-parameter. Segmentation results from tracking of detection points over multiple radar frames (viz. utilizing tracking information) which is used for associating detection points to objects (segmentation-by-tracking). A two-level tracking approach can be implemented, in which for a present radar frame, (new) detection points are associated to tracklets, which may be seen as a first level, and then the tracklets are associated to object-tracks, which may be seen as a second level.