Autonomous Object Track Prioritization Under Compute Limits
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object tracking systems for autonomous and semi-autonomous vehicles face challenges in managing a large number of tracked objects, leading to increased computing resources and latency, which can affect safety-critical functions by failing to track important objects.
Innovation Solution
Implement a system that prioritizes object tracking based on criteria such as proximity, classification, and time-to-collision, limiting the number of tracked objects to a threshold while ensuring safety-critical objects are maintained, using a track-management component to determine priority scores and manage resource usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the number of tracked objects is increased to track all detected objects, then tracking completeness is improved, but computing resources and latency increase
Solution Approach 1:
The patent applies local quality by assigning different tracking priorities to different objects based on their characteristics. Safety-critical objects (e.g., pedestrians, cyclists, animals) receive high priority and are always tracked, while non-critical objects (e.g., stationary signs, buildings) receive low priority and are tracked only when resources permit. This differentiated approach ensures critical tracking completeness while optimizing resource usage.
Solution Approach 2:
The system dynamically changes the parameter of track limit based on available computing resources. When resources are abundant, the track limit is increased to track more objects. When resources are constrained, the limit is reduced while maintaining priority tracking of safety-critical objects. This adaptive parameter adjustment resolves the contradiction between tracking completeness and resource efficiency.
2Productivity
If the number of tracked objects is limited to a threshold, then computing resources are conserved, but safety-critical objects may be missed
Solution Approach 1:
The patent segments the tracked objects into different priority levels: high-priority safety-critical objects (pedestrians, cyclists, animals, emergency vehicles), medium-priority moving objects (other vehicles, motorcycles), and low-priority stationary objects (signs, buildings, fences). This segmentation ensures that even when the total track number is limited, high-priority objects are always preserved in the tracking list, preventing safety-critical objects from being missed.
Solution Approach 2:
The system performs preliminary classification of detected objects into priority categories before applying the track limit. By pre-identifying safety-critical objects and assigning them high priority, the system ensures these objects are protected from being dropped when the track threshold is reached. This preliminary action guarantees safety-critical object detection regardless of resource constraints.
3Speed
If tracking latency is reduced to improve real-time performance, then planning and control functions are enhanced, but tracking accuracy may deteriorate
Solution Approach 1:
The patent implements dynamic tracking update frequencies based on object priority and motion characteristics. High-priority safety-critical objects are tracked at every frame with full accuracy to minimize latency. Medium-priority objects are tracked at reduced frequencies when resources are constrained. This dynamic approach ensures that critical tracking functions maintain high accuracy and low latency, while non-critical tracking can tolerate higher latency.
Data Source
AI summary
In various examples, object track management for autonomous or semi-autonomous systems and applications is described herein. Systems and methods are disclosed that may limit the number of objects that are tracked based on one or more criteria. For instance, the number of objects that are tracked may be limited to a threshold number of objects when a number of detected objects exceeds a threshold. The systems and methods may use parameters associated with the detected objects to determine priority scores associated with the detected objects, and may then determine to only track the detected objects with the highest scores (e.g., high priority objects). As a result, latency and compute of the system may be reduced while still maintaining tracking with respect to safety-critical objects.