Crowdsourced Dynamic Map for Autonomous Vehicle Perception
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in detecting environmental objects due to line-of-sight limitations of sensors and adverse conditions, leading to impaired driving performance.
Innovation Solution
A system and method that aggregates and distributes dynamic information from multiple vehicles using a remote computing device to create a crowdsourced dynamic map, allowing vehicles to share detected object data and enhance environmental awareness beyond individual sensor capabilities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If autonomous vehicles rely on onboard sensors (cameras, LiDAR) for environment perception, then they can detect objects within line-of-sight, but they cannot detect objects blocked by obstacles or in adverse conditions
Solution Approach 1:
The patent merges sensor data from multiple autonomous vehicles to create a collective environmental perception system. By combining detections from vehicles at different positions and angles, the system overcomes individual line-of-sight limitations and adverse conditions, achieving more reliable environment detection than any single vehicle could accomplish alone.
Solution Approach 2:
The patent introduces a communication network and data processing system as intermediaries between individual vehicle sensors and the autonomous driving decision-making process. This intermediary infrastructure aggregates, validates, and distributes environmental information from multiple sources, enabling vehicles to access感知 data beyond their own sensor capabilities.
2Measurement precision
If autonomous vehicles use HD maps with multiple layers (geometric, semantic, dynamic) for navigation, then driving performance is enhanced, but incomplete or inaccurate sensor data impairs the accuracy of these maps
Solution Approach 1:
The patent combines sensor data from multiple vehicles to create a more complete and accurate representation of the environment. By merging observations from different perspectives and conditions, the system reduces information loss and creates more reliable HD map data, particularly for objects that may be obscured from any single vehicle's viewpoint.
3Speed
If a single vehicle uses its own sensors to detect objects, then it has real-time detection capability, but it cannot detect objects outside its field of view
Solution Approach 1:
The patent merges real-time sensor data from multiple vehicles moving through the environment, creating a collectively comprehensive view that covers areas outside any single vehicle's field of view. This distributed sensing approach maintains real-time detection capability while eliminating blind spots that would exist for any individual vehicle.
Data Source
AI summary
A method includes receiving sensor data, with one or more sensor of a vehicle, wherein the sensor data is associated with an environment of the vehicle, detecting an object based on the sensor data, determining pixel coordinates of the object based on the sensor data, converting the pixel coordinates of the object to world coordinates of the object, extracting a set of features associated with the object, and transmitting the set of features and the world coordinates to a remote computing device.