Data Fusion Algorithm Design

Sensor fusion development framework – BASELABS Create

BASELABS Create is a software framework that is designed for the fast development of complex data fusion algorithms and environment models. The framework can be used with field-tested, pre-implemented algorithms as well as for the development of fully custom algorithms. BASELABS Create provides numerous features for the development of data fusion applications. It can be used in connection with vADASdeveloper, or with other established development tools like Matlab and ADTF. Typical use cases for BASELABS Create include multiple object tracking (MOT), lane marking recognition, vehicle detection or dead reckoning applications.


Efficient data fusion development for ADAS and automated driving
BASELABS Create is designed for the fast development of complex data fusion algorithms. BASELABS Create can be used with field-tested, pre-implemented algorithms as well as for the development of fully custom algorithms. Key features and use cases are:

  • Framework and GUI for data fusion algorithm development
  • Ready-to-use system and sensor models
  • Suitable for arbitrary multiple sensor data fusion use cases
  • Interoperates with related products (e.g. vADASdeveloper, ADTF, ROS, Polysync, RTMaps, Matlab)
  • BASELABS Create is available for Windows and Linux

Workflow overview for BASELABS Create
BASELABS Create and Code as well as vADASdeveloper from Vector can be integrated in different workflows for the development of advanced driver assistance systems and automated vehicles. BASELABS Modules provide customer-specific functionality, if required. The software interops with established products to increase the flexibility in development. It and can be used as the main tool chain for the prototyping of systems or to add functionality to existing tools.

If you use…

  • … ADTF, you can develop your ADAS algorithms in BASELABS Create and integrate them into your ADTF system.
  • ... MATLAB for your function development, you can use BASELABS Create for the convenient design of environment perception systems.
  • ... SIMULINK for the design of control algorithms, then BASELABS Create is a tool for the model-based design of perception algorithms. The perception system can then directly be used in Simulink.
  • ... vADASdeveloper, you can directly use your algorithm in an application with real sensors.

Use Cases

BASELABS Create is a powerful software framework to implement data fusion applications
Typical use cases for BASELABS Create include lane marking recognition, vehicle detection, dead reckoning or multiple object tracking applications, among others.

Multiple Object Tracking

Sensor fusion for Multiple Object Tracking (MOT)
The identification and tracking of objects using uncertain sensor data is a common task during the design of driver assistance and autonomous systems. BASELABS Create is used to implement the required algorithms, which can be directly used within vADASdeveloper or other frameworks like Matlab Simulink. To get started very fast, BASELABS Create provides a template application for a multiple object tracker that can directly be used as a basis for the system development.



Loading the player ...

Tracking Template

Loading the player ...

Multiple Sensor Fusion

Loading the player ...

Matlab Interop

Loading the player ...

The Multi Object Tracking (MOT) template application and sample measurement data set (included in BASELABS Create) including overlay and bird's-eye view data visualization for tracked objects allows a quick start for setting up data fusion applications. The template application can be modified by the user to address specific challenges.

Central multiple sensor data fusion

Use Case for 360° Perception

Loading the player ...

360° Perception

Loading the player ...


Localization and navigation system development
Designing precise dead reckoning applications is a major challenge in the industry, as the requirements towards the localization increase. To give an example, urban or cooperative scenarios demand for very accurate and reliable gps localization. With BASELABS Create, the relevant localization applications can be designed in conjunction with vADASdeveloper or other frameworks:

  • Lane level positioning - e.g. by camera data fusion
  • Car-2-Car communication - e.g. with more accurate CAMs
  • Map matching - e.g. on 3D-maps
  • Autonomous driving - e.g. with 360° perception
  • Data logging and evaluation - e.g. for post-processing and test

Increase of integrity, accuracy and availability in localization applications
The sensor fusion applications used for localization have to fulfil though requirements towards integrity, accuracy and availability. BASELABS Create (in conjunction with vADASdeveloper) allows to:

Increase integrity by:

  • Tightly coupling of dead reckoning and GPS observations
  • Exclusion of erroneous GPS signals
  • Accounting of sensor latencies and bias errors
  • Multipath detection and mitigation

Increase accuracy by:

  • Using ionospheric corrections (IONEX, SBAS)
  • NLOS detection
  • Usage of multiple GNSSs (GPS, GLONASS, Galileo)

Increase availability by:

  • Tightly coupling in combination with bias-compensated dead reckoning and flexible vehicular motion models

Other industries

Sensor data fusion systems in other industries
BASELABS Create can be used in different industries. Exemplarily, the marine sector provides similar challenges to the developer as in the automotive industry. The recognition and tracking of other vessels on the water is required to realize safety or comfort functions. The data fusion for these applications can be implemented using BASELABS Create.

Use of BASELABS Create for marine environment

Loading the player ...


The right features for state-of-the-art data fusion development
BASELABS Create provides numerous features for the development of data fusion applications that can be called from the framework. It can be used in connection with vADASdeveloper, or with other establised development tools like Matlab or ADTF.

Sensor fusion

Data fusion of multiple sensors

BASELABS Create supports the design of data fusion systems with an arbitrary number of sensors. This includes systems with identical sensors (e.g. multiple radars) as well as systems with different sensors (e.g. radar and MobilEye).

Different data fusion levels

For each sensor setup, an appropriate level of data fusion can be chosen. To combine high level information from multiple sensors like track lists from radars and cameras, track-to-track fusion methods (e.g. covariance intersection) are provided by BASELABS Create. The fusion of raw sensor data is also possible.

Increased system reliability

BASELABS Create can significantly increase the reliability of perception systems by supporting redundant (e.g. a camera and a radar surveying the same area) as well as complementary (e.g. multiple radars for a 360° view) sensor setups through an integrated existence estimation and clutter handling.

Multiple Object Tracking (MOT)

Multiple field of view support and track management

The sensors that are used by BASELABS Create may observe overlapping or non-overlapping regions. For each region, an independent detection and persistence model can be applied. Tracks that are initialized by a sensor are reliable tracked while they pass other sensors (handover). See an example of a complex tracking system with eleven sensors in this video.

Measurement Association Algorithms and Integrated Gating

To associate sensor measurements to tracked objects, Create provides several association algorithms, e.g. single or multiple local nearest neighbor association. In addition, BASELABS Create integrates multiple gating mechanisms to cope with a high number of objects or measurements, e.g. Mahalanobis distance or probability gating. Beside that shipped methods, custom association and gating techniques can be implemented using the provided interfaces.

Integrated existence estimation and clutter handling

To reduce sensor effects like missed detections and false alarms, BASELABS Create contains algorithms that explicitly address these issues, e.g.

  • Sequential Probability Ratio Testing (SPRT)
  • Integrated Probabilistic Data Association (IPDA)
  • Generalized probabilistic Data association (GPDA)

    Detection models

    To describe sensor characteristics like its false alarm rate or occlusion behavior, Create provides several detection models. These models are used in an integrated way to improve the reliability of the estimation.
    If you are searching for sensor models for simulation and validation, please refer to BASELABS Models.

    State estimation

    Built-in system and measurement models

    Create is shipped with several system and sensor models which are suitable for different use cases like object tracking or localization using different sensors like radar or camera.

    • Constant Velocity (CV model)
    • Vector-wise Constant Velocity
    • Constant Acceleration (CA model)
    • Constant Turn Rate and Velocity (CTRV model)
    • Constant Turn Rate and Acceleration (CTRA model)

    Beside the shipped models, custom models (linear or non-linear) can be implemented by either extending existing models or by using the provided interfaces.

    Estimation algorithms / filters

    BASELABS Create contains estimation algorithms to cover a variety of use cases. Filters for estimation that are shipped with BASELABS Create are:

    • Kalman filter
    • Extended Kalman filter
    • Unscented Kalman filter
    • Particle filter

    Existing filter in different programming languages (e.g. C or C++) can directly be integrated into a data fusion system designed with BASELABS Create.

    Development process

    Design/compile time checks

    BASELABS Create is designed to reduce errors already at compile time. This prevents from taking undiscovered errors into the later stages of development.
    The BASELABS Create space validation is a compiler extension which checks that spaces fulfil the requirements of BASELABS Create. By that, algorithm development can be made even more safe as errors are detected already at compile time. Finally, this reduces the test and debugging effort of complex signal processing algorithms.
    Full Space Safety is included for

    • system, sensor, detection and persistence models,
    • estimators/filters,
    • association algorithms,
    • gating and
    • state representations.

      Multiple Object Tracking template application

      To get started very fast, BASELABS Create provides a template application for a multiple object tracker that can directly be used as a basis for the system development.


      Product improvements, new software features and bug fixing
      BASELABS is working on permanent product improvements including new features or bug fixing. We inform our users about software updates and releases regularly.


      This is a bug fix release of BASELABS Create which contains the following improvements:

      • Fixed a bug in the example code of the 'BASELABS Create Data Fusion Application' template project which is part of the Visual Studio integration of BASELABS Create.
      • Fixed a bug in the Visual Studio extension which caused Visual Studio to crash if multiple versions of BASELABS Create are used.
      • Fixed a bug in the 'Baselabs.Statistics' NuGet package which could lead to inconsistent assembly references after an update of BASELABS Create.
      • Removed the contract assemblies from the 'Baselabs.Statistics' NuGet package.


      This is a bug fix release of BASELABS Create which contains the following improvements:

      • Fixed a bug that could cause the BayesFilter.Predict() method to return wrong results.
      • Fixed a bug that could cause a biased calculation of the covariance in the SampleSet.ToGaussian() method.
      • Fixed a bug that could cause the PersistenceModel.Evaluate() method to return negative likelihood values in retrodiction scenarios.
      • Fixed a bug that could cause licensing issues on Linux.


      This software release includes the following product improvements:

      • Linux development support. Developers can now develop on Linux, e.g. Ubuntu with MonoDevelop.
      • Fixed a bug in the vADASdeveloper connection designer when GhostDoc is installed which caused a license exception.
      Top of page