Crowdsourced Position-Orientation Messages for HD Map Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for generating and maintaining high-definition maps for autonomous vehicles, such as using dedicated sensor fleets, are not scalable for wide geographic coverage and frequent updates, and rely on the quality of sensors deployed in vehicles, which limits the accuracy of position estimation and mapping.
Innovation Solution
A method for configuring, triggering, and transmitting relative and global position-orientation messages using user equipment like autonomous vehicles, which collect sensor data from GPS, IMU, wheel encoders, or cameras to estimate their position and orientation, and transmit these messages to a remote server for map generation and updates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a dedicated fleet of vehicles with sensors is used to generate HD maps, then map quality and accuracy are improved, but the solution is not scalable for wide geographic coverage and frequent updates
Solution Approach 1:
The system enables ordinary vehicles (not a dedicated fleet) to contribute to HD map generation by automatically collecting sensor data and transmitting position-orientation messages. Each vehicle serves dual purposes: its primary function and map contribution, eliminating the need for a specialized dedicated fleet while achieving scalable geographic coverage.
Solution Approach 2:
The patent makes standard vehicle sensors (cameras, LIDAR, radar, GPS) serve multiple functions: both their original purposes and HD map generation. This multi-functionality allows any vehicle in the network to participate in map crowdsourcing, dramatically increasing scalability without requiring specialized hardware.
2Adaptability or versatility
If map crowdsourcing is used to enable scalable coverage, then geographic coverage and update frequency are improved, but the quality of position estimation and mapping depends on sensor quality in vehicles
Solution Approach 1:
The system combines multiple sensor types (camera, LIDAR, radar, wheel encoders, GPS) and multiple data sources (vehicle sensors, remote server HD maps) to create redundant measurement systems. This merging allows cross-validation and compensation for individual sensor limitations, maintaining position estimation accuracy across diverse vehicle fleets.
Solution Approach 2:
The remote server receives position-orientation messages from vehicles, compares them with existing HD maps, and provides feedback for map updates. This feedback mechanism allows continuous refinement of map accuracy using real-time vehicle data, ensuring that crowdsourced maps maintain high quality despite varying sensor capabilities.
3Productivity
If vehicles transmit raw or processed sensor data to remote servers for map generation, then map updates can be frequent and coverage wide, but the system complexity and data processing requirements increase
Solution Approach 1:
The system segments map data into multiple layers (e.g., 2D maps, 3D models, point clouds, road network data) and processes them separately at different locations. Vehicle sensors first process data locally to extract position-orientation messages, then transmit only essential information to the remote server for integration into the appropriate map layer, reducing overall system complexity.
Solution Approach 2:
Vehicles perform preliminary processing of sensor data locally before transmission to the remote server. By pre-processing and filtering data at the vehicle level, the system reduces the volume and complexity of data that needs to be transmitted and processed centrally, enabling frequent updates without overwhelming system resources.
Data Source
AI summary
Disclosed are techniques for wireless communication. In particular, aspects relate to configuring, triggering and/or transmitting relative and global position-orientation messages (e.g., from a vehicle equipped with sensors to enable estimation of relative and global position-orientation to a network entity).