Transport Mode Determination via Mobile Network Trajectory Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for determining modes of transport based on smartphone data are inefficient in accounting for short stops and require extensive calibration and population surveys, making them difficult to update and maintain, while also being resource-intensive in terms of memory and computing power.
Innovation Solution
A method that uses spatial and temporal data from mobile telephone networks to identify modes of transport by forming transport graphs, determining sub-routes, and calculating correlation indices to optimize the determination of transport modes, which can be updated regularly without the need for frequent population surveys.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If mobility modeling approaches are used to measure transport mode choices, then analytical tools are available for public authorities, but the models are difficult to implement, maintain, and update due to expensive population survey data requirements and significant calibration effort
Solution Approach 1:
The patent replaces traditional mechanical survey-based mobility modeling with an electronic signal processing approach. Instead of using expensive population survey data and complex calibration procedures, the system uses smartphone sensor data (accelerometer, gyroscope, magnetometer, barometer) and GPS information processed through signal filtering and pattern recognition algorithms to automatically determine transport modes, thereby substituting a resource-intensive mechanical system with a more efficient electronic one
Solution Approach 2:
The patent creates a virtual replica of transport mode characteristics by analyzing patterns in smartphone sensor data. Instead of directly observing actual transport usage through surveys, the system copies the characteristic motion patterns, acceleration profiles, and spatial-temporal features of different transport modes from smartphone sensors to identify and classify them, enabling indirect measurement without physical surveys
2Reliability
If population surveys are conducted every 5 to 10 years to update mobility data, then statistical data is available, but the data cannot capture rapid changes in mobility choices over time
Solution Approach 1:
The patent enables continuous measurement of transport mode choices by leveraging smartphones that users carry continuously in daily life. The system processes sensor data and GPS information in real-time as users move, providing ongoing updates on mobility patterns rather than periodic snapshots. This continuous data collection and processing allows public authorities to monitor evolving transport choices immediately as they occur, eliminating the 5-10 year update cycle of traditional surveys
3Measurement precision
If traditional methods account for short stops in trajectory analysis, then more accurate transport mode identification is achieved, but the computational resources and memory required increase significantly
Solution Approach 1:
The patent segments the trajectory analysis into distinct phases by detecting and isolating short stops as separate events. Instead of analyzing the entire continuous trajectory with high computational complexity, the system divides the path into moving segments and stationary segments, applying simplified analysis to each. This segmentation allows accurate identification of transport modes while reducing overall computational burden by treating stops as discrete, manageable units rather than continuous complexity
Data Source
Figure 1~2
Figure 3
Figure 4~5
AI summary
The invention relates to a method for determining a mode of transport (MDP) of a path (traj) by means of a telephone network (Res), and transport graphs (G_MDP) for each mode of transport in which the following steps are carried out: a) spatial and temporal data (DSP) of the connection of a telephone (Tel) to the network (Res) are acquired and a path (traj) is determined by defined location-times of the positions of network antennas connected to the telephone; b) For each transport graph (G_MDP), identified nodes (NIS) are determined (Det) from the location-time pairs; c) Then, possible paths (CP) are determined (Opt) by shortest path optimizations; d) we determine (Corr) a correlation index (Ind) between each path (CP) and the route (traj) and e) we determine (Det_MDP) the mode of transport (MDP) by optimizing the correlation index (Ind).