Traffic State Determination Using Hidden Markov Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Determining the traffic state of a vehicle based solely on location and speed data is challenging, as a vehicle's speed of zero can indicate being parked, at a temporary stop, or in heavy congestion, making it difficult to accurately assess traffic conditions.
Innovation Solution
A hidden Markov model is used to analyze a series of probe data instances, determining the current traffic state by building a model with hidden states representing traffic conditions such as free flow, light congestion, heavy congestion, temporary stop, and parked states, using observational and transition probabilities based on speed and distance parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If only location and speed data are used to determine traffic state, then the system is simple to operate, but the measurement precision of traffic state determination deteriorates
Solution Approach 1:
The patent transforms the single parameter (speed) into multiple parameters by introducing temporal dimension (series of speed measurements over time) and statistical parameters (standard deviation, trend analysis). This allows distinguishing between different traffic states that share the same instantaneous speed value, thereby improving measurement precision while maintaining ease of operation through automated analysis.
Solution Approach 2:
The patent applies partial action by using a limited set of probe data instances rather than analyzing all possible data. By selecting representative samples and applying statistical thresholds, the system achieves sufficient measurement precision without overwhelming computational complexity, balancing ease of operation with accurate traffic state determination.
2Measurement precision
If a series of probe data instances are analyzed using hidden Markov model, then the measurement precision of traffic state determination is improved, but the device complexity increases
Solution Approach 1:
The patent introduces a hidden Markov model as an intermediary computational layer between raw probe data and traffic state determination. This intermediary processes the complex temporal patterns in speed data, transforming them into reliable traffic state classifications. The model acts as a mediator that handles the complexity internally while presenting simple, accurate results to the user system.
Solution Approach 2:
The patent uses probabilistic copying through the hidden Markov model, where multiple possible traffic state sequences are evaluated and the most likely sequence is selected. This copying approach allows the system to explore multiple hypotheses simultaneously and converge on the correct traffic state determination, improving measurement precision without requiring direct complex physical measurements.
3Measurement precision
If hidden Markov model with multiple traffic states is built, then the measurement precision is improved, but the difficulty of detecting and measuring traffic state deteriorates
Solution Approach 1:
The patent segments the continuous traffic state space into discrete, manageable categories (e.g., free flow, light congestion, heavy congestion, temporary stop, parked). Each segment corresponds to a specific traffic state with characteristic speed patterns. This segmentation makes detection and measurement feasible by providing clear classification boundaries and reducing the complexity of identifying the correct state among multiple possibilities.
Data Source
AI summary
An instance of probe data that was captured by sensors of a probe apparatus onboard a vehicle is received. Previous instances of probe data captured by the probe apparatus onboard the vehicle are accessed and the instance of probe is analyzed based on the previous instances of probe data. A current traffic state is determined for the vehicle based on the analysis. In an example embodiment, the analysis comprises generating a hidden Markov model based on speed data of the probe data. A Viterbi-path is obtained corresponding to the instance of probe data and the previous instances of probe data. The current traffic state is defined based on the Viterbi-path. The current traffic state may be used to determine traffic information/data for a road segment and/or predict a future traffic state for the vehicle. Traffic management decisions and/or routing decisions for the vehicle may be made based thereon.