Vehicle Position Estimation Using Asynchronous Sensor Graph Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous driving systems face challenges in accurately estimating vehicle position when dealing with asynchronous sensing data, as the node creation time of the graph structure and sensing data creation times differ, leading to potential inaccuracies and latency in real-time position information.
Innovation Solution
The method modifies the graph structure by selectively applying sensor data interpolation and graph node interpolation to match asynchronous sensing data, minimizing the number of nodes added and optimizing computation, thereby ensuring accurate and timely position estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the graph structure is modified to match asynchronous sensing data by adding nodes, then the position estimation accuracy is improved, but the computational time and complexity increase
Solution Approach 1:
The system performs preliminary actions by predicting vehicle position using motion information before sensing data arrives, and pre-processes graph structure modifications. This allows the system to prepare node additions and graph updates in advance, reducing the computational time required when actual position estimation is performed, while still achieving accurate results by incorporating the asynchronous sensing data when it becomes available.
Solution Approach 2:
The graph structure is made dynamic by selectively adding nodes only when necessary to represent new sensing data timestamps. The system dynamically adjusts the graph structure based on the arrival of asynchronous sensing data, adding nodes only for timestamps that don't exist in the current graph structure. This dynamic approach improves position estimation accuracy by incorporating all available sensing data while minimizing unnecessary computational overhead from adding redundant nodes.
2Measurement precision
If the graph structure is modified to incorporate asynchronous sensing data, then the position estimation accuracy is improved, but the number of nodes to be added increases
Solution Approach 1:
The system applies partial action by selectively adding only the minimum necessary nodes to the graph structure. Instead of adding nodes for every possible timestamp or creating a complete graph for all sensing data, the system adds nodes only for the specific timestamps where asynchronous sensing data needs to be incorporated. This partial approach maintains position estimation accuracy by including necessary data points while avoiding the complexity of adding excessive nodes that would not contribute to improving accuracy.
Solution Approach 2:
The graph structure modification is applied locally rather than globally. The system identifies specific timestamps and locations in the graph structure where asynchronous sensing data needs to be incorporated, and modifies only those local portions of the graph. This local quality approach ensures that nodes are added only where necessary to represent specific sensing data, rather than uniformly increasing the number of nodes throughout the entire graph structure, thus maintaining accuracy while minimizing complexity.
3Measurement precision
If sensor data interpolation is applied to match asynchronous sensing data, then the position estimation accuracy is improved, but the computation amount increases
Solution Approach 1:
Motion information is used to predict vehicle position in advance, serving as a preliminary estimation before actual sensing data is processed. This preliminary action provides a starting point for position estimation that reduces the computational burden of interpolation, as the system only needs to make adjustments from this predicted position rather than performing extensive interpolation calculations from scratch, thereby reducing computation amount while maintaining accuracy.
4Measurement precision
If graph node interpolation is applied to match asynchronous sensing data, then the position estimation accuracy is improved, but the computation amount increases
Solution Approach 1:
Graph node interpolation is applied partially and selectively only when necessary to incorporate asynchronous sensing data, rather than being applied uniformly to all graph nodes. The system identifies specific nodes and timepoints where interpolation is needed to match the asynchronous data timestamps, and performs interpolation only in those localized areas. This partial application maintains position estimation accuracy by ensuring necessary interpolations are performed while significantly reducing the overall computation amount by avoiding unnecessary interpolations across the entire graph structure.
Data Source
Figure 1
Figure 2(a)~2(b)
Figure 3(a)~3(b)
AI summary
Disclosed are a method for position estimation of a vehicle based on a graph structure and a vehicle using the same. The method for position estimation of a vehicle based on a graph structure according to an exemplary embodiment of the present invention may include a sensing step of creating sensing data according to translation of the vehicle using a plurality of sensors; a graph structure creating step of creating a node according to a reference period and setting a constraint condition between the nodes to the sensing data synchronized with the reference period to create a graph structure; a graph structure modifying step of modifying the graph structure using the asynchronous sensing data when the asynchronous sensing data which is not synchronized with the reference period among the sensing data is input; and a position estimating step of creating position information of the vehicle using the graph structure.