LiDAR Data Encoding for HD Maps With Lower Bandwidth Load
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional maps for autonomous vehicles lack the precision and timeliness required for safe navigation due to limitations in sensor data accuracy and the expense and inefficiency of traditional mapping methods, which struggle to keep up with frequent road updates.
Innovation Solution
The method involves encoding and compressing sensor data, particularly LiDAR data, to generate high-definition maps with sub-2 cm resolution, allowing autonomous vehicles to accurately navigate by representing data in relation to a reference centerline, reducing storage and transmission requirements, and enabling efficient data transmission and storage.
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 to create and update maps increases significantly
Solution Approach 1:
The patent uses autonomous vehicles to collect sensor data that copies the environment's geometric features, which are then processed to create HD maps. This replaces the need for expensive survey teams while maintaining map accuracy through automated data collection and processing pipelines.
Solution Approach 2:
The system enables autonomous vehicles to self-collect mapping data during normal operation, eliminating the need for dedicated survey missions. The vehicles autonomously navigate, collect LiDAR and image data, and contribute to map creation and updates without human intervention.
2Loss of time
If survey fleets are expanded to capture frequent road updates, then map freshness is improved, but the cost and logistical complexity increases
Solution Approach 1:
The patent makes autonomous vehicles serve dual purposes: both performing their primary delivery function and simultaneously collecting mapping data. This eliminates the need for separate survey fleets, as the same vehicles perform both transportation and mapping tasks, reducing overall system complexity.
Solution Approach 2:
The system enables continuous map updates by having autonomous vehicles constantly collect and transmit sensor data during their regular operations. This creates an ongoing feedback loop that continuously refreshes the HD maps without requiring periodic dedicated survey missions.
3Measurement precision
If high resolution sensor data is collected continuously, then navigation precision is improved, but data storage and transmission requirements increase
Solution Approach 1:
The patent extracts only the essential geometric features and semantic information from raw sensor data that are necessary for navigation, discarding redundant information. This extraction process retains navigation precision while significantly reducing the volume of data that needs to be stored and transmitted.
Solution Approach 2:
The system segments the continuous sensor data stream into discrete, manageable units representing specific environmental features (lanes, curbs, signs, etc.). This segmentation allows for efficient compression and selective transmission of only the most critical navigation-relevant information.
Data Source
AI summary
Embodiments relate to methods for efficiently encoding sensor data captured by an autonomous vehicle and building a high definition map using the encoded sensor data. The sensor data can be LiDAR data which is expressed as multiple image representations. Image representations that include important LiDAR data undergo a lossless compression while image representations that include LiDAR data that is more error-tolerant undergo a lossy compression. Therefore, the compressed sensor data can be transmitted to an online system for building a high definition map. When building a high definition map, entities, such as road signs and road lines, are constructed such that when encoded and compressed, the high definition map consumes less storage space. The positions of entities are expressed in relation to a reference centerline in the high definition map. Therefore, each position of an entity can be expressed in fewer numerical digits in comparison to conventional methods.