Extended Object Tracking via Component Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current driving assistance systems for automated driving face challenges in tracking extended objects with low-resolution sensors, as they often struggle to accurately assign point measurements to their origins due to insufficient data, leading to unreliable object state estimation and high computational intensity.
Innovation Solution
A method is developed to track extended objects by modeling multiple components of an object, such as wheels and vehicle sides, and determining association probabilities for these components using a probabilistic data association filter, which reduces computational resources and time required for tracking.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a radar measurement model with Bayesian filter is used to track extended objects, then measurement precision and object state estimation reliability are improved, but computational resources and processing time increase significantly
Solution Approach 1:
The patent segments the extended object into multiple discrete components (e.g., vehicle corners, wheels, sides). Each component is tracked independently with its own state vector, allowing the system to process objects through individual component associations rather than treating the entire object as a single complex entity. This segmentation reduces the computational burden of the Bayesian filter while maintaining tracking precision.
Solution Approach 2:
The patent introduces an intermediary measurement model that maps object components to sensor detections through probabilistic associations. Instead of directly applying complex Bayesian filtering to raw sensor data, the system uses component-based measurements as an intermediary layer, simplifying the filter's computational tasks while preserving measurement precision.
2Device complexity
If point measurements from low-resolution sensors are used to track extended objects, then device complexity is reduced, but measurement precision and origin assignment reliability deteriorate
Solution Approach 1:
The patent divides the extended object into multiple trackable components (corners, wheels, sides), allowing low-resolution sensors to capture and associate measurements with specific components rather than treating the object as an indistinct whole. This segmentation enables reliable origin assignment even with limited sensor resolution.
Solution Approach 2:
The patent changes the measurement representation from raw sensor data to component-based measurements with associated probabilities. By transforming the measurement parameters to include component-specific information and association probabilities, the system improves origin assignment reliability while maintaining the simplicity of low-resolution sensors.
3Measurement precision
If multiple detections are required to reliably estimate object state, then measurement precision improves, but tracking time and processing duration increase
Solution Approach 1:
The patent segments the object into multiple components that can be tracked independently. This allows the system to accumulate measurements for each component separately, achieving reliable state estimation more quickly by processing component-level data in parallel rather than waiting for sufficient detections of the entire object.
Solution Approach 2:
The patent applies partial action by tracking only the most significant components (e.g., corners, wheels) rather than requiring complete coverage of all object features. This selective approach achieves sufficient measurement precision for safe driving assistance while reducing processing time compared to comprehensive multi-detection requirements.
Data Source
AI summary
A method for tracking extended objects in the surroundings of a vehicle includes the steps of: (a) providing modeled components for an object; (b) determining a detection from sensor data; and (c) determining the origin of the detection on the object using the modeled components. The process of determining the origin of the detection on the object includes determining association probabilities for the plurality of modeled components, in which the association probabilities indicate the degree of probability with which the detection is associated with the individual components of the modeled components.