Open Simulation Interface ASAM OSI Using RoadRunner for ADAS ECU Validation

Round open simulation interface. Our models extract all the objects that is there in the trace. It could be moving objects, stationary objects, or any traffic signs, environmental information. And what is done in sensor model is we do coordinate transformation from the world coordinates to the sensor coordinates. We will do filtering. We do parameter calculation like detection parameters, range rate, azimuth, et cetera.

And then that detection is given then to the logic model, where we run the tracking filters. And then the output of that is given to feature functions, which give out warning, alerts, and stuff. So here is an example where the white object here is the host. And we have a other object, green, which is a traffic object. And there are dots. You can see these are the detections created by the sensor model.

Now those detections are then given to the logic model. Again, you can see the logic model creating a white box, which is the object. And then this object is then given to the feature functions. For example, what you're seeing is a closing vehicle warning alert. And when it comes into the blind spot region, it gives a blind spot warning. All of this is done in a virtual environment using open simulation interface as the heart.

So another area where we have used open simulation interface is in validation. So we have vehicle logs. We collect vehicle logs using the real vehicles, using the real sensors on the vehicles. Then we create digital twin of that using RoadRunner of the actual log.

So then this digital twin data, again, as you saw before, we get open simulation interface output. And this is given to the sensor model and logic model giving out. For example, from the sensor model, we have detections. And this is compared to the actual detections that was collected during the actual drive. And similarly, we can also compare the actual objects, which are created by the perception algorithms with the logic model output.

So here, there are twofold advantage. One is we are validating our models, which we developed using the real vehicles. And then we can also actually-- so if we have an update in the software, in the actual software, we actually don't have to go again on the road to do this test. We can actually do the same scenario with the updated models and see what is the performance with respect to the previous run.

So open simulation interface is one of the standards of ASAM, and other standards will be then spoken by Anaga.

Yeah. ASAM stands for Association for Standardization of Automation and Measuring Systems. It's basically a standards body. What they have done is, especially for the [INAUDIBLE] AD world, they have come up with something called as an OpenX standards, mainly called as scenario-based testing with OpenX. And what is OpenX? Basically, OpenX depends on where exactly we are with respect to the development or verification and validation.

If you are talking about specifications or description of the test, you have something called ASAM open test. Similarly, ODD description, we have ASAM OpenODD. If you're talking about the scenes and scenarios that we just looked into, for scene, what we have is called as an OpenDRIVE. So OpenDRIVE. Is nothing but how the road networks are, right?

And similarly, OpenCRG is a texture. I'll talk about this a little bit more. And similarly, OpenSCENARIO is how your dynamic elements move. Open Simulation Interface is during the simulation. Where you are seeing here is during the simulation interface, OSI comes into the picture.

So how exactly MathWorks supports this ASAM Open standards. We start with OpenX, OpenDRIVE. OpenDRIVE, as I said, everything talks about lanes, roads. Anything on those, it's all about the OpenDRIVE, how about the junctions, right? So OpenDRIVE is a standard that helps us give the description of the roads, networks, and all.

And then you can export into OpenDRIVE format. The beauty is this OpenDRIVE format file can also be imported into any of the other simulator tools. So in MathWorks, we can actually import OpenDRIVE, as well as we also can export to OpenDRIVE.

Now talking about the CRG. Now we talked about the roads. But how exactly the roads are, like the texture of the roads, the potholes on the roads, everything is defined in the ASAM OpenCRG format. That's the other standard. Again, MathWorks supports that.

Similarly, we have ASAM OpenSCENARIO again. Now we're talking about the dynamics. So far, we talked about the static elements. When we're talking about the dynamics, how exactly the vehicle moves? Is the vehicle taking a left turn or a right turn? Is it overtaking or cutting into another lane? All that piece of information will be in the form of ASAM OpenSCENARIO format.

And OpenSCENARIO is into two main standards, two main formats. One is OpenSCENARIO 1.x and OpenSCENARIO 2.0. And MathWorks supports both the standards. If you are working in any of these areas, feel free to reach out to us. We'll be happy to support you in supporting these standards.

And last but not the least is ASAM OSI Open Simulation Interface. That's what the stock is all about, how you take the simulation data from one particular simulator to a different simulator, agnostic of environment, basically. With that, we'll come to summary. Natij will close the summary.

Now, looping back to the same slide that we started with, we can see we talked about what is virtual simulation, where we compare the real flow against the virtual flow. And the heart of all of this is Open Simulation Interface. It is an interface between the environmental simulators and the models. And why it is advantageous, because we can then use any environmental simulator with any model without making any change in the interface.

Aptu has built the models as a platform-based model using Open Simulation Interface, whereas I had mentioned previously, we can use different environmental simulators without making any change to the models. RoadRunner has been really helpful in creating scenarios. The use of the UI is really great. And it is faster in creating scenarios. And it is really easy in the usage.

And recently, we had collaboration with MathWorks to get this open simulation interface to be an output from the RoadRunner scenario. And that is it for today. Thank you.