Radar Point Cloud Correction for Multipath Reflection Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in accurately interpreting multipath reflections in radar and lidar point clouds, which can degrade data resolution and limit usable information, often leading to incorrect navigation decisions due to the indistinguishability of multipath and single path data.
Innovation Solution
A computing device equipped with a machine learning algorithm, such as a graph neural network, differentiates between single path and multipath reflections in point cloud data, correcting the location and orientation of multipath reflections by training on labeled sensor and mapping data, thereby generating accurate corrected point cloud data for navigation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multipath reflection data is discarded to maintain data quality, then measurement precision is improved, but loss of information increases
Solution Approach 1:
The patent applies this principle by transforming multipath reflection data, which was traditionally considered harmful noise, into beneficial information. The system identifies multipath reflections and reinterprets them to extract valid object location data, converting what was previously discarded harmful data into useful navigation information.
Solution Approach 2:
The patent changes the interpretation parameters of multipath reflection data. Instead of discarding these reflections based on traditional parameters, the system modifies how these reflections are analyzed and interpreted, allowing the extraction of meaningful object position information from previously unusable data.
2Productivity
If multipath reflection data is used without differentiation, then productivity is improved by utilizing more data, but measurement precision deteriorates due to incorrect interpretation
Solution Approach 1:
The patent segments point cloud data into different categories based on reflection paths. By separating single-path reflections from multipath reflections and processing them differently, the system maintains high measurement precision while utilizing all available data for navigation decisions.
Solution Approach 2:
The patent introduces an intermediary classification process that acts as a mediator between raw point cloud data and navigation decisions. This intermediary step identifies and differentiates multipath reflections, allowing the system to use the data appropriately without compromising accuracy.
3Measurement precision
If machine learning algorithms are implemented to differentiate reflection types, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent replaces complex mechanical or geometric processing systems with machine learning algorithms. The ML-based approach simplifies the overall system architecture by using learned patterns to identify multipath reflections, rather than requiring complex rule-based or geometric reasoning systems.
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
Enhances the accuracy and resolution of point cloud data by correctly interpreting multipath reflections, improving the navigation capabilities of autonomous vehicles by providing reliable positional and maneuvering information.
Implementation Method 1
each point in the cloud results from a reflection of a signal emitted by the autonomous vehicle's radar or lidar system 101 into the environment 103
Implementation Method 2
a radio detection and ranging (radar) sensor system
Implementation Method 3
a light detection and ranging (lidar) sensor system
Implementation Method 4
a light detection and ranging (lidar) sensor system
Data Source
AI summary
A method and apparatus for implementing a method are disclosed. The method includes providing point cloud data to a machine learning algorithm, the point cloud data detected in the vicinity of an autonomous vehicle. The method further includes differentiating, via the machine learning algorithm, in the point cloud, data directly representing a location of a first object and data indirectly representing a location of a second object. The method includes transforming the data indirectly representing the location of the second object into data directly representing the location of the second object and generating corrected point cloud data based on the data directly representing the location of the first object and the data directly representing the location of the second object. The method includes outputting the corrected point cloud data to the autonomous vehicle.