Vehicle Object Relevance Scoring for Trajectory-Based Planning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Perception and prediction systems in vehicles require significant computing resources and time to determine actions due to the large number of objects in the environment, leading to potential inefficiencies and safety concerns.
Innovation Solution
A method is introduced where vehicles analyze sensor data to identify relevant objects by determining relevance scores based on potential interactions, using threshold scores and time differences to prioritize objects for the planner component, thereby reducing the computational load and increasing safety.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the planning system analyzes all objects in the environment, then the completeness of object detection is improved, but the computing resources and processing time increase significantly
Solution Approach 1:
The patent segments the set of all objects into two distinct groups: relevant objects that require detailed planning analysis and irrelevant objects that can be processed with simplified methods. This segmentation is achieved through a relevance determination component that evaluates each object's importance based on predefined criteria, allowing the system to allocate computing resources efficiently while maintaining detection completeness for critical objects
Solution Approach 2:
The patent extracts and identifies a subset of relevant objects from the complete set of detected objects using a relevance determination component. This extraction process separates objects that have significant impact on vehicle operation from those with minimal impact, enabling the planning system to focus computational efforts on the extracted relevant subset while still maintaining awareness of all objects in the environment
2Measurement precision
If the planning system analyzes all objects in the environment, then the accuracy of action determination is improved, but the processing time increases
Solution Approach 1:
The patent applies partial action by analyzing all objects for detection purposes but applying detailed accuracy-based planning analysis only to the subset of relevant objects. The relevance determination component identifies which objects warrant full analytical treatment, allowing the system to achieve sufficient accuracy for action determination while processing fewer objects in detail, thereby reducing overall processing time without sacrificing critical decision-making accuracy
3Loss of information
If the planning system processes a large number of objects, then the comprehensiveness of environmental awareness is improved, but the productivity of the system decreases
Solution Approach 1:
The patent segments object processing into two parallel tracks: a comprehensive detection phase that maintains environmental awareness of all objects, and a focused planning phase that processes only relevant objects in detail. The relevance determination component enables this segmentation by categorizing objects based on their significance to vehicle operation, allowing the system to maintain comprehensive environmental awareness while improving productivity through selective detailed analysis
Solution Approach 2:
The relevance determination component acts as an intermediary between the perception system that detects all objects and the planning system that requires detailed analysis. This intermediary evaluates each detected object and determines its relevance, filtering the complete set of objects into relevant and irrelevant categories, thereby enabling the planning system to operate efficiently on a reduced subset while the perception system maintains comprehensive environmental awareness
Data Source
AI summary
This disclosure is directed to techniques for identifying relevant objects within an environment. For instance, a vehicle may use sensor data to determine a candidate trajectory associated with the vehicle and a predicted trajectory associated with an object. The vehicle may then use the candidate trajectory and the predicted trajectory to determine an interaction between the vehicle and the object. Based on the interaction, the vehicle may determine a time difference between when the vehicle is predicted to arrive at a location and when the object is predicted to arrive at the location. The vehicle may then determine a relevance score associated with the object using the time difference. Additionally, the vehicle may determine whether to input object data associated with the object into a planner component based on the relevance score. The planner component determines one or more actions for the vehicle to perform.