Extended Object Tracking With Center-Truncated Radar Measurement Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automotive radar systems face challenges in accurately tracking the kinematic and extended states of objects due to complex real-world measurement distributions, which are not effectively captured by existing measurement models, leading to inaccurate object tracking.
Innovation Solution
A system and method utilizing a center-truncated distribution and underlying Gaussian distribution pair to estimate the expanded state of objects, incorporating a probabilistic filter with a hierarchical measurement model that accounts for spatial distribution and sensor noise, allowing for accurate tracking of both kinematic and extended states.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If contour model is used to capture measurement distribution along object contour, then measurement precision is improved, but device complexity increases due to requiring much more degrees of freedom and complex data association
Solution Approach 1:
The patent changes the parameterization approach from complex contour models with many degrees of freedom to simple geometric primitives (rectangle, triangle, circle) with few parameters. This transforms the measurement model from requiring detailed contour fitting to using basic shape parameters (position, orientation, dimensions), thereby reducing complexity while maintaining measurement precision through the probabilistic filtering framework
Solution Approach 2:
The patent employs simple, computationally inexpensive geometric primitive models instead of complex contour models. These simple models act as disposable approximations that can be quickly computed and updated through probabilistic filtering, sacrificing the detailed accuracy of contour models for computational efficiency and reduced complexity
2Device complexity
If surface model such as Gaussian-based ellipse is used, then computational complexity is reduced, but measurement precision deteriorates because the model is away from real world automotive radar measurements
Solution Approach 1:
The patent applies local quality by using different geometric primitive models (rectangle, triangle, circle) that are locally appropriate for different types of objects. Each primitive model captures the local measurement distribution characteristics of specific object types, providing better accuracy than a single Gaussian-based ellipse model while maintaining computational simplicity
Solution Approach 2:
The patent changes from continuous Gaussian-based surface models to discrete geometric primitive models with specific parameter sets. This parameterization change allows the model to better match the actual distribution of radar measurements around object surfaces while keeping the computational framework simple and manageable
3Measurement precision
If fixed set of points on rigid body is used for extended object tracking, then object state estimation is improved, but data association complexity increases and scalability is reduced
Solution Approach 1:
The patent extracts the essential object state information (position, orientation, dimensions) from the measurements and represents it using simple geometric primitives. This extraction approach eliminates the need for complex data association between fixed points and measurements, as the geometric primitive parameters directly encode the object state without requiring point-to-point correspondence
Solution Approach 2:
The patent changes the state representation from a fixed set of points requiring complex association to a small set of geometric primitive parameters. This parameter transformation simplifies the data association problem by directly mapping measurements to primitive parameters through probabilistic filtering, thereby improving scalability and reducing computational complexity
Data Source
AI summary
A system and a method for tracking an expanded state of an object including a kinematic state indicative of a position of the object and an extended state indicative of one or combination of a dimension and an orientation of the object is provided herein. The system comprises at least one sensor configured to probe a scene including a moving object with one or multiple signal transmissions to produce one or multiple measurements of the object per the transmission, and a processor configured to execute a probabilistic filter tracking a joint probability of the expanded state of the object estimated by a motion model of the object and a measurement model of the object, wherein the measurement model includes a center-truncated distribution having predetermined truncation intervals. The system further comprises an output interface configured to output the expanded state of the object.