Autonomous Vehicle Object Tracking With Hierarchical Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous vehicle systems face challenges in managing and tracking the classification information of multiple objects due to the increasing number of classes, leading to difficulties in maintaining accurate object classification and communication among different modules.
Innovation Solution
Implementing a hierarchical classification structure using a Dempster-Shafer theory-based evidence-based classification tracking and fusion system, which organizes classification information in a tree structure, incorporates class similarity into distance metrics, and maintains a centralized hierarchical representation to facilitate efficient tracking and communication.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a flat vector representation is used for object classification with increasing number of classes, then the classification coverage is improved, but the difficulty of managing and tracking class information increases
Solution Approach 1:
The patent segments the flat vector representation into a hierarchical tree structure where classes are organized in parent-child relationships. This segmentation allows the system to handle hundreds of classes by breaking them down into manageable hierarchical levels, reducing the complexity of tracking while maintaining comprehensive classification coverage.
Solution Approach 2:
The patent transforms the one-dimensional flat vector representation into a multi-dimensional hierarchical tree structure. This dimensional change organizes classification information across multiple levels (parent classes, child classes, leaf nodes), making it more tractable to manage and track numerous object classes while preserving the complete classification space.
2Ease of operation
If hierarchical classification structure is implemented, then the tractability of tracking class information is improved, but the system complexity increases
Solution Approach 1:
The patent segments the classification space into hierarchical levels with clear parent-child relationships. This segmentation makes tracking tractable by allowing the system to focus on relevant subclasses based on parent class predictions, reducing the effective search space at each tracking step while maintaining comprehensive classification capability.
Solution Approach 2:
The patent applies local quality by allowing different parts of the hierarchical tree to have different levels of detail and importance. Not all classes require the same tracking intensity - leaf nodes represent specific objects requiring precise tracking, while parent nodes provide broader contextual classification. This localized approach improves tractability without uniformly increasing system complexity.
3Measurement precision
If confidence-based state updating is implemented, then the accuracy of object classification is improved, but the computational requirements increase
Solution Approach 1:
The patent implements partial action by updating only the necessary portions of the hierarchical tree based on confidence thresholds. Instead of re-evaluating the entire classification tree, the system performs partial updates on affected nodes and their descendants, reducing computational requirements while maintaining classification accuracy through targeted confidence-based filtering.
Data Source
AI summary
A method executed by a computing device of an autonomous vehicle includes (i) receiving first sensor data from a network of one or more sensors; (ii) generating a first prediction, based at least in part on the received first sensor data, for an object being associated with a first state with a first confidence value; (iii) receiving second sensor data from the network of one or more sensors after elapsing of time t after receiving the first sensor data; (iv) generating a second prediction, based at least in part on the received second sensor data, for the object being associated with a second state with a second confidence value; and (v) updating a state of the object in a hierarchical representation tree to be the first state or the second state based upon a higher confidence value between the first confidence value and the second confidence value.