In4Mo. Advanced information system for the mobility of people and vehicles
The design and development of In4Mo takes into account potential short- and medium-term technological scenarios, in which traditional traffic detection technologies (for example, magnetic induction loops) coexist with emerging technologies (such as magnetometers). ICT (including Bluetooth and GPS) play a particularly relevant role, and their progressive introduction into society means they can be used to develop more efficient systems.
The basic thesis of In4Mo is that technology, that is, the introduction of sensors in cities, is essential but not enough on its own to generate reliable, precise, value-added information that is available at the required time and place. In other words, the degree of "smartness" is the result of efficient data combination and processing.
However, data from sensors that use different technologies are heterogeneous. Consequently, the main innovation in In4Mo is the proposed method for filtering, merging and collecting data that incorporates a range of flexible, efficient data processing and analysis methods. The final product is comprised of a basic platform providing homogeneous, consistent data that can then be fed into traffic information and management models in any control centre. The following techniques are applied:
Filtering techniques (nonlinear and Kalman filter), which generate complete, consistent, solid sequences, delete outliers, and obtain data completeness.
Fusion techniques (Bayesian, neural networks and traffic models, among others), which logically combine data from heterogeneous sources to obtain better quality homogeneous information.
State forecasting techniques (state estimation models) to estimate the state of the traffic system and predict its potential short-term development.
The availability of traffic data from ICT sensors considerably improves the quality of traffic information, compared to that generated by current systems. However, the implementation of ICT alone is not enough to provide a complete picture of the state of traffic in the entire road network, particularly in mid- to large-sized urban areas. This is particularly true when the information system must support applications that require real-time computing of paths between any origin-destination pair, as in complete navigation systems or journey planning systems that that should not be limited to sets of predetermined paths (however important they may be). Consequently, the generation of complete, consistent information requires the use of dynamic traffic models that provide an overall estimation of the state of the road network and predict its short-term development in the absence of incidents. In4Mo proposes two complementary dynamic traffic models:
Models for the estimation of time-dependent origin-destination matrices that use measurements of traffic variables from ICT applications, through ad hoc versions of Kalman filters.
Models that describe the dynamics of the propagation of traffic flow in the road network, according to demand patterns defined by time-dependent origin destination matrices.
Those responsible for a road network’s traffic management need access to information generated by combined data and traffic models. In In4Mo, this finding has been transformed into an innovative Smart Mobility proposal: active management based on the Macro Fundamental Diagram (MFD).
- PostGIS, JQuery, Hibernate, AMPL, alquilar CPLEX, JBoss, PostgreSQL, HTML, JPA, Java, C + +, Matlab, R, Aimsun simulator, Dynameq simulator
Areas of expertise involved in the projects
- Optimization models for situating traffic sensors (Detection Layout)
- Dynamic traffic models
- Discrete event simulation applied to transport, manufacture, logistics and other services
- Applications for active management and traffic information
- Smart Mobility
- Data merging
- Fusion of traffic data from heterogeneous sources
- Data and information analysis