From data to model
TOPAS Model Fitting
For intelligent systems to make informed decisions in dynamic scenarios, they require a reliable and well-coordinated digital representation of the scenario. Digital twins created for this purpose can interpret the current state of the environment, determine critical parameters for detection and control, and extrapolate the current state into the future. This enables 'what if' scenarios to be created.
TOPAS ModelFitting is a flexible tool that enables you to create automated models using your own data sets.
These models are created using the TOPAS Model Fitting GUI. This standalone application allows you to enter data, optimize settings and track the model fitting process in real time.
If you would like to build on the TOPAS ModelFitting algorithms, you can use our Developer Kit. This provides access to the TOPAS ModelFitting library, written in modern C++. This enables you to seamlessly integrate model-fitting functions into your existing code base.
Simulating complex systems using digital twins
The development of a digital twin starts with your system. This could be a mobile robot system, for example, or a distributed process involving multiple types of sensor and actuator. The aim is to create a digital model that can be used to predict how the system will behave under certain controls, enabling you to make optimal decisions.
Modelling the relationships between control systems and observed state changes is challenging. This is particularly true of complex systems, where the underlying mathematical models can be highly dimensional, making manual identification and parameterisation difficult and often requiring additional specialist knowledge.
The aim of TOPAS Model Fitting is to streamline this process. It applies methods developed for solving high-dimensional optimization problems to the identification and fitting of models. In doing so, it uses information contained in the data sets without the need to implement the optimization techniques themselves.
You can use either the TOPAS Model Fitting GUI or the C++ interface to import your data sets, initiate identification and fitting processes, and export the results to various file formats.
Use the digital twin to control your system optimally, detect real-time deviations, and make long-term forecasts.
Examples of use
Autonomous systems and robotics are key technologies that allow machines to carry out tasks independently and flexibly. These technologies are indispensable in industries such as automotive and aerospace, where they are used for autonomous driving, manufacturing and inspections in challenging environments.
Accurate models are necessary in order to calculate how to control technical systems in order to achieve a goal. For example:
Model - Movement of a robotic arm
Control - Torques in the joints
Objective - Planning pick-and-place jobs
Model – Flight behaviour of a satellite
Control – Propulsion systems
Target – Trajectory planning
Model – Driving dynamics of a vehicle
Control – Acceleration, steering
Goal – Autonomous driving manoeuvres