Map-Anchored Object Detection Using Travel Way Marker Offsets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in accurately detecting objects at long ranges using limited sensor data, particularly with sparse LIDAR returns and uncertainty in range estimation.
Innovation Solution
The system anchors object detections to map data, processing sensor data fused with map data to determine the position of detected objects in the mapped environment, thereby constraining the detection solution space and improving accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Length of stationary object
If LIDAR sensors are used for long-range object detection, then detection range is extended, but sensor cost and data sparsity increase
Solution Approach 1:
The patent merges map data with sensor data to create a fused representation of the environment. By combining the structured spatial information from maps with the real-time observations from sensors, the system achieves denser environmental representation at long ranges without requiring additional expensive sensors.
Solution Approach 2:
The patent introduces map data as an intermediary to bridge the gap between sparse sensor observations and the complete environmental model. The map data serves as a mediator that provides missing spatial information, allowing the system to achieve dense environmental understanding even when sensor returns are sparse at long ranges.
2Measurement precision
If expensive LIDAR sensors are used, then detection accuracy improves, but vehicle cost increases
Solution Approach 1:
The patent uses map data as a copy or prior representation of the environment to supplement real-time sensor data. By leveraging the pre-existing detailed environmental model in the map, the system achieves accurate object detection without relying solely on expensive high-performance sensors.
Solution Approach 2:
The patent makes the map data serve multiple functions: providing spatial constraints for detection, filling in gaps in sensor observations, and enabling accurate localization. This multi-functional use of map data reduces dependence on expensive specialized sensors.
3Speed
If sensor data is used alone for detection, then real-time performance is maintained, but detection accuracy at long ranges decreases
Solution Approach 1:
The patent performs preliminary processing of map data to create a ready-to-use environmental model before real-time detection. By pre-processing the map to extract relevant spatial information and structures, the system can quickly fuse this with incoming sensor data during real-time operation, maintaining speed while improving accuracy.
Data Source
AI summary
An example method includes (a) obtaining sensor data descriptive of an environment of an autonomous vehicle; (b) obtaining a plurality of travel way markers from map data descriptive of the environment; (c) determining, using a machine-learned object detection model and based on the sensor data, an association between one or more travel way markers of the plurality of travel way markers and an object in the environment; and (d) generating, using the machine-learned object detection model, an offset with respect to the one or more travel way markers of a spatial region of the environment associated with the object.