RT-LAB™ is a distributed Real-Time platform that enables engineers to conduct Real-Time Simulation of Simulink models with Hardware-in-the-Loop, in a very short time, at a low cost. Its scalability allows the developer to add computing power where and when it is needed. It is flexible enough to be applied to the most complex simulation and control problem, whether it is a Real-Time Hardware-in-the-Loop application or for speeding up model execution, control and test.
RT-LAB provides tools for running simulations of highly complex models on a network of distributed run-time targets, communicating via ultra low-latency technologies, in order to achieve the required performance. In addition, RT-LAB's modular design enables the delivery of economical systems by supplying only the modules needed for the application in order to minimize computational requirements and meet customers price targets. This is essential for high-volume embedded applications.
How it Works
RT-LAB allows the user to readily convert Simulink models, via Real-Time Workshop (RTW), and then to conduct Real-Time Simulation of those models executed on multiple target computers equipped with multi-core PC processors. This is used particularly for Hardware-in-the-Loop (HIL) and rapid control prototyping applications. RT-LAB transparently handles synchronization, user interaction, real-world interfacing using I/O boards and data exchanges for seamless distributed execution.
Single Target Configuration
In this configuration, typically used for rapid control prototyping, a single computer runs the plant simulation or control logic. One or more hosts may connect to the target via an Ethernet link. The target uses QNX or Linux as the RTOS for fast simulation or for applications where real-time performance is required. RT-LAB used Red Hat ORT which is the standard Red Hat distribution package with an optimized set of parameters to reach real-time performance enabling to reach model time step as low as 10 micros on multi-core processors.
Distributed Target Configuration
The distributed configuration allows for complex models to be distributed over a cluster of multi-core PCs running in parallel. The target nodes in the cluster communicate between each other with low latency protocols such as FireWire, SignalWire or InfiniBand, fast enough to provide reliable communication for real-time applications. The real-time cluster is linked to one or more host stations through a TCP/IP network. Here again, the cluster of PCs can be used for Real-Time applications (using QNX or RedHawk Linux), or fast simulation of complex systems (using QNX, RedHawk or Windows). RT-LAB PC-cluster targets are designed for flexible and reconfigurable mega-simulation. The user can build and expand the PC-cluster as needed, then redeploy the PCs for other applications when the simulation is done. RT-LAB can accommodate up to 64 nodes running in parallel.
