Python in the loop: AI-Driven simulation and validation for autonomous systems – Simcenter Prescan 2606 

What’s new

ADAS and autonomous system development increasingly relies on Python, machine learning frameworks, and large volumes of sensor data. Yet simulation often remains disconnected from these workflows, requiring engineers to switch between separate tools, scripting environments, and post-processing pipelines. 

Simcenter Prescan 2606 helps bridge this gap by bringing simulation directly into Python-based development workflows through Python-native execution, AI-assisted scenario creation, and support for large-scale digital twins. 

A new way of working with Python and AI assisted workflows 

Reducing the effort spent integrating simulation tools is becoming increasingly important as development workflows shift toward Python-based software engineering practices. 

Engineers can run complete simulations directly from Python, control scenario logic and actuators programmatically, and access sensor data such as camera, lidar, and radar in real time. This makes it possible to build custom simulation loops and user-defined models within development environments such as VS Code and PyCharm. 

Working directly in Python also makes it easier to take advantage of AI coding assistants. With SKILL files and improved API documentation, these tools can better understand Prescan APIs, helping users discover functionality, generate code examples, and create simulation scenarios with less manual effort. 

By bringing simulation, sensor data, and analysis into the same development environment, teams can spend less time connecting tools and more time developing, testing, and validating autonomous system functionality. 

Scalable environments and accelerated simulation performance 

High-fidelity validation requires realistic environments at scale, but building and simulating these environments can quickly become a performance bottleneck. With Simcenter Prescan 2606, new instancing capabilities allow efficient rendering of thousands of repeated objects, from dense vegetation and roadside elements to large urban environments. This makes it easier to create larger and more realistic digital twins without compromising performance. 

At the same time, Simcenter Prescan 2606 further advances simulation performance on modern hardware. Multi GPU support for Physics Based Point Cloud Lidar allows the workload of a single sensor to be distributed across multiple GPUs, while support for modern GPU architectures ensures optimal performance on the latest systems. 

Together, these improvements reduce simulation runtime for complex, high-resolution scenarios, allowing teams to validate larger environments and run more simulation campaigns within the same development cycle. 

Additional enhancements and a modern documentation experience 

Beyond its core innovations, Simcenter Prescan 2606 introduces a range of enhancements that further improve usability, realism, and workflow efficiency. Support for OpenSCENARIO 1.2 and 1.3 enables more flexible and standardized scenario development, while enhanced radar outputs provide high fidelity scene data that allow users to develop their own custom signal generation models. Dynamic resource actuation enables lighting conditions and display content to be updated during simulation, creating more interactive and realistic environments. In addition, several improvements have been introduced to streamline scenarios and path creation, making it easier and more efficient to design and modify simulation setups. These improvements are complemented by ongoing refinements to sensor robustness and overall simulation performance. 

To support these advanced workflows, Simcenter Prescan documentation has also been fully modernized. The transition to a searchable HTML based experience, together with dedicated Python and MATLAB API documentation, makes it easier to explore functionality, discover APIs, and accelerate development in a more intuitive way. 

Bringing it all together 

Simcenter Prescan 2606 introduces a new paradigm for simulation: 

• Python-first workflows that integrate seamlessly with modern development tools 

• AI-assisted engineering for faster scenario creation and innovation 

• Scalable digital twins with realistic environments and high performance 

With these capabilities, Simcenter Prescan continues to evolve into a platform that not only simulates reality but also integrates directly into how modern engineering software is developed. 

Find out more in the full release notes, or download the updated version now (links will only work for exisiting customers)

Past releases