GPS Trajectory Map Matching With Path Search and Backtracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing map matching algorithms for GPS trajectories suffer from low matching efficiency, high error rates, and inability to simultaneously consider matching accuracy and efficiency, particularly in complex traffic networks.
Innovation Solution
The proposed method involves a path search guided GPS point matching logic, which replaces conventional GPS point guided path search logic, avoiding endless loops and inefficient map data calling processes, and improving matching efficiency significantly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional stepwise matching algorithms (HMM, MDP, LBMM) are used for map matching, then matching accuracy can be maintained, but matching speed becomes significantly slower and application effectiveness is restricted
Solution Approach 1:
The patent segments the map matching process into two distinct phases: forward iteration (path search) and backtracking verification. The forward iteration efficiently searches for candidate paths using path search algorithms, while the backtracking phase verifies and refines the matching by traversing GPS points backward. This segmentation allows the system to achieve both high speed in path search and high accuracy in verification, resolving the contradiction between matching speed and accuracy.
2Measurement precision
If path search algorithms are used for accurate matching, then matching accuracy improves, but overall matching proportion decreases to below 50% due to network topology errors
Solution Approach 1:
The patent implements a feedback mechanism through the backtracking verification phase. After the forward iteration produces candidate paths, the backtracking phase reviews and corrects matching results by traversing GPS points backward from the end point. This feedback loop identifies and corrects local matching errors caused by network topology issues, ensuring that the final matching proportion reaches above 90% while maintaining high accuracy through the combined forward-search and backward-verify approach.
3Productivity
If conventional matching algorithms are used, then processing can be completed, but abnormal GPS points caused by equipment faults seriously affect matching quality
Solution Approach 1:
The patent applies preliminary action by initializing a valid GPS point list before the main matching process. Abnormal GPS points are identified and removed from the trajectory data in advance, creating a cleaned list of valid points for matching. This preliminary data preparation ensures that the subsequent forward iteration and backtracking verification processes work only with reliable GPS points, preventing equipment faults from degrading matching quality while maintaining processing efficiency.
4Productivity
If existing matching methods are used for complex traffic networks, then matching can be performed, but local matching errors are unavoidable and uncontrollable
Solution Approach 1:
The patent introduces dynamic adaptability through the backtracking verification phase, which dynamically adjusts and corrects local matching results based on the actual GPS trajectory. The system can identify when a local matching error occurs in complex network regions and automatically correct it by reviewing the trajectory backward and comparing against road network topology. This dynamic correction capability allows the system to handle complex traffic networks effectively, achieving above 90% matching proportion while controlling local matching errors through adaptive verification.
Data Source
AI summary
An efficient map matching method for GPS trajectory includes: obtaining a valid GPS point list; obtaining a current searched end point; backtracking a best map matching path from the current searched end point, traversing GPS points along the path to obtain best matching results, and summarizing the best matching results to output a matching result statistical table; and determining whether the current searched end point is a final point.