Modelit creates scientific software and carries out research. We develop scientific software for disciplines that require highly interactive visualization tools and lend itself to the application of mathematical tools like statistics, optimization and linear algebra.
Our main activities are:
- Creating and maintaining turn-key applications for the visualization, validation and analysis of data obtained from sensor networks;
- Creating data driven web portals and dashboards;
- Development and optimization of algorithms;
- Carrying out impact analysis and evaluation studies;
- Taking care of periodic reporting of accessibility indicators;
- Creation of ready to use software components for the Matlab community.
Traffic engineering
- Validation of traffic data
- Calculation and publication of accessibility indicators
- Estimation of (dynamic) OD matrices
- Development of incident detection algorithms
- Single and Multiuserclass traffic assignment models
- Traveltime prediction
- Processing of Floating Car Data
- Visualisation, analysis and publication of traffic data
- Discrete choice modelling
Coastal protection
- Validation of waterlevel measurements
- Modelling of waterlevel in rivers
- Validation and processing of coastal morphology data
- Validation of wave data
- Publication of yearly wave measurement reports and storm reports for the IJsselmeer and Waddenzee
- Prediction of stormsurges with neural networks
- Provide access to the DONAR and LOL (National Database for depth soundings) databases
- Composing of Bathymetric charts
In addition to desktop applications Modelit also creates and hosts web applications and inspection tools. Examples are:
- Tripcast, route planner with actual and predicted traveltimes
- Hotspots, web portal for support of risk based speed enforcement
- Sensor data inspection tool, show real time and historic traffic sensor data
The front-end for these web applications is coded in frameworks such as Angular4, React or Dojo and communicates with the back-end through JSON messages. The coding for the back-end is done in Matlab. Matlab is often perceived as a prototyping tool because math computations and powerful graphics can be launched with intuitive commands written on the command line or collected in simple scripts. However, Matlab is also a fully fledged and powerful programming language.
For deploying Matlab programs as a webservice, Modelit has developed the Modelit Webserver Toolbox (2008) and more recent the Modelit Embedded Webserver Toolbox (2020). Using either of these toolboxes results in a compiled Matlab program that can be deployed (and scaled) at no extra cost. For additional ease of deployment, programs created with MEWT can be packed as Docker containers that will run on any Linux machine with Docker installed.
MWT and/or MEWT let a developer complete the full cycle of experimenting, coding, testing and deployment from within the Matlab ecosystem.
Most of our application development is done in Matlab. Matlab has become the standard language for numerical computation, mathematical modelling and visualisation of data in the last decades. Using the principle "Use the best commercial tools available and develop new tools when necessary" has lead to the development of various toolboxes which make it possible to use Matlab as a fully fledged programming language while preserving the typical advantages of Matlab. Modelit differs in this approach from other providers of tailor-made Matlab solutions. Examples of toolboxes that were developed by Modelit are:
- Matlab Application Framework; the framework for all our applications
- Matlab Layout Manager; a layout manager for complex GUI's
- User Interface Toolbox; integrate Java components in Matlab applicaties
- Webserver Toolbox;
- XML toolbox for Matlab;
- Network modelling toolbox; building blocks for all applications for traffic-engineers
- Distributed computing for Matlab; remote execution of Matlab functions
- In memory compression; make extremely large datasets accessible as arrays
- Docutool; automatic generation of sourcecode documentation
Java is an important component in our applications. Using Java enables us to quickly develop an attractive and useful software solution for every mathematical problem. If needed our applications can be deployed as compiled stand alone applications for which no Matlab license is required. Examples of applications developed by Modelit are:
- DelftOD (proprietary), visualisation and analysis of traffic data on corridor level
- TRIP (proprietary), visualisation and analysis of traffic data on network level
- Maria (RWS-RIKZ), visualisatie, validation and acceptation of depth soundings
- Melissa (RWS-RIKZ), validation waterlevel measurements in rivers and in coastal areas
- DiaInspector (RWS-RIKZ), visualisation van timeseries from the Donar database of Rijkswaterstaat
- WAVIX (RWS-RIKZ), model based validation of wave data based on neural networks.
- Stormnet (RWS-RIKZ), prediction of stormsurges with neural networks.
- SMARA (TNO/RPB), network model for traveltime prediction
- SMARTMUC (TNO), multi-userclass traffic assignment model
- Davinci (RWS-AVV), visualisation and validation of the traffic data
- Fix, Database and analysis tool for fermentation data
- FermBase, a datamining tool for the analysis of batch processes
- FermView, online tool for monitoring of batch processes