Autonomous Vehicle Lidar Point Cloud Ephemeral Noise Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional maps for autonomous vehicles lack the accuracy and timeliness required for safe navigation due to limitations in sensor data processing and the expense of maintaining high-definition maps, with dynamic objects like moving vehicles and pedestrians not being effectively filtered out.
Innovation Solution
A system using lidar sensors to capture 3D maps by identifying and removing ephemeral points associated with dynamic objects through voxel cell filtering, ensuring accurate and up-to-date map generation and updates for autonomous vehicles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional maps are created using survey teams with high resolution sensors, then map accuracy is improved, but the cost and time required for map creation increases
Solution Approach 1:
The patent uses standard-resolution sensor data from commercially available autonomous vehicles to create high-definition maps, copying the mapping capability from expensive survey equipment to ordinary vehicles. This allows any autonomous vehicle to contribute to map creation and updates without requiring specialized survey teams.
Solution Approach 2:
The system enables autonomous vehicles to self-update maps using their own sensor data. Vehicles automatically capture, process, and contribute their environmental observations to the map database, eliminating the need for manual survey teams while maintaining high accuracy through continuous self-updating.
2Reliability
If survey fleets are expanded to capture more road updates, then map freshness is improved, but the cost increases
Solution Approach 1:
The patent makes every autonomous vehicle a potential mapping asset. Instead of relying on a specialized survey fleet, any autonomous vehicle on the road can capture and contribute map data, universalizing the mapping function across the entire vehicle fleet.
Solution Approach 2:
Vehicles automatically detect road changes and updates using their sensors, process this data, and contribute it to the map database without requiring manual intervention or expansion of survey teams. This self-service approach maintains map freshness at low cost.
3Measurement precision
If machine learning techniques are used to identify dynamic objects, then filtering accuracy is improved, but computational cost increases
Solution Approach 1:
The patent divides the point cloud data into multiple layers or segments, processing different portions of the data with different levels of computational intensity. This segmentation allows accurate dynamic object detection in critical areas while using simpler methods for less critical regions, reducing overall computational cost.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The system provides high-definition maps with precise location accuracy (within 10 cm) and timely updates, enabling safe navigation by filtering out noise from dynamic objects, thus enhancing the reliability of autonomous vehicle systems.
Implementation Method 1
An autonomous vehicle system comprises a light detection and ranging (lidar) sensor that captures lidar samples for use in generating three-dimensional (3D) maps
Data Source
AI summary
An autonomous vehicle system removes ephemeral points from lidar samples. The system receives a plurality of light detection and ranging (lidar) samples captured by a lidar sensor. Along with the lidar samples, the system receives an aligned pose and an unwinding transform for each of the lidar samples. The system determines one or more occupied voxel cells in a three-dimensional (3D) space using the lidar samples, their aligned poses, and their unwinding transforms. The system identifies occupied voxel cells representative of noise associated with motion of an object relative to the lidar sensor. The system filters the occupied voxel cells by removing the cells representative of noise. The system inputs the filtered occupied voxel cells in a 3D map comprising voxel cells, e.g., during the map generation and/or a map update.