In episode 24 we take on realistic simulations through real-time data exchange. We’re joined by simulation expert Rob Proctor, Field Application Engineer for RTI. Rob’s extensive knowledge of the Training and Simulation space helps us understand the latest trends and challenges in simulated training and testing environments. We’ll also discuss how other industries, such as healthcare and autonomous systems, can benefit from this type of technology.
In Episode 24 of The Connext Podcast:
- [1:32] “We don’t make simulations, we make them better”
- [2:19] Current trends in the Training and Simulation market
- [4:52] How simulations can help test autonomous vehicles
- [9:47] How does Connext DDS compare to legacy solutions used in the industry today?
- [12:03] Rob tells a story of the importance of interoperability
- [18:56] RTI’s experience in the training and simulation market to date
- [Blog] Can DDS Help Solve the Distributed Simulation Integration Challenge?
- [Capability Brief] RTI in Training and Simulation
- [Press Release] Real-Time Innovations Announces Partnership with VT MAK for the Training and Simulation Market
- [Press Release] Real-Time Innovations Connext DDS Selected as Part of Winning Proposals for the U.S. Army Synthetic Training Environment (STE) Program
Steven Onzo: Hello everybody. Thanks for joining us for another episode on the Connext Podcast. Today we'll be interviewing Rob Proctor, field applications engineer here at RTI. Rob provides insight on the current training and simulation market and how leveraging real time connectivity can offer a much more realistic simulation experience. We'll also discuss how we can test autonomous systems much faster using simulation modeling. Thanks and we hope you enjoy this episode.
David Barnett: Welcome to another episode of Connext Podcast. I'm David Barnett, the Vice President of Products and Markets at RTI. I am here today with Rob Proctor, Senior Field Applications Engineer with RTI, to talk about the training and simulation market. Rob, thank you for being here today. Do you wanna introduce yourself and tell us a little bit about your background?
Rob Proctor: Sure. I'm Rob Proctor. I'm a Senior Field Applications Engineer for Real-Time Innovations. I say, Real-Time Innovations, not RTI, because of the segment we're gonna talk about today. That's modeling and simulation. I've been an FAE for about 12 years now. What that really means, is I'm the first interface between a customer and our products. As folks come up with requirements, they start doing Google searches, or they see us at conventions and shows, they generally find and talk to me. We are geographically spaced. I cover Florida, Georgia and South Carolina.
[1:32] “We don’t make simulations, we make them better”
David Barnett: Okay. Thanks. Well, now that you've educated me on the proper use of the RTI acronym, let's start with an overview of where Real-Time Innovations fits in the training and simulation market.
Rob Proctor: Right. We don't make simulations. We don't really make anything, do we. We make things better. I always think about the old BASF commercials. We don't make the ball, we make it shiny. Well, we don't make the simulations, we make them interoperate better. Clearly we are focused on that. That's something that's kind of strange to folks, when they see us at these training simulation events. They're like, "Oh, what kind of simulations do you run? What kind of training do you provide?" I say, "We don't provide any simulations or training. We are a networking interface for real time communications."
[2:19] Current trends in the Training and Simulation market
David Barnett: What are some of, the current trends in the training and simulation market and some of the challenges that result from those trends?
Rob Proctor: Okay. Let's break it down by segments. In aerospace and defense segment, some of, the big nuggets, big rocks that are out there, these major all encompassing programs. For the army, there is something called STE, which is S-T-E or synthetic training environment. The STE is an all encompassing, live virtual and constructive simulation space, which is going to be all things to all people. It's gonna have a one world terrain, where you can go actually ... Let's say you're in San Francisco and you wanna quickly go over to the middle east to zoom in, zoom in on little townships, and actually go into buildings and effect scenarios and change things around.
Rob Proctor: The point here, is that there's gonna be distributed locations, with interfacing interoperability between all these different components. You can have folks in Parris Island, interfacing with folks out here in Moffett Field and San Francisco. You can have folks at Eglin Air force Base, interfacing with folks at Fort Leavenworth Kansas. That's kind of the big overall theme in aerospace and defense. The air force has one as well. It's called, SCAR. Navy is just now looking at interoperability.
Rob Proctor: Of course, since it's aerospace and defense, everything needs to be secure. That's definitely interoperability, dispersed locations, with security. That's the big trend in aerospace and defense. There's plenty of other fields that when you talk about simulation, there's medical. There's autonomous cars.
Rob Proctor: Medical in particular, is an interesting one, because, when we talk medical simulations, we're really talking mannequins with thousands of sensors on them. Trying to learn, train folks to have better accuracy. What you're looking to avoid in that case, since you're not dealing with security, dispersed locations, maybe some interoperability. What you're really looking for there is to avoid something called, negative training. Negative training is almost exactly what it says. You train negatively, and then when you go to operate live, you have improper and incorrect training. We're trying to avoid that. We want integrity and fidelity with the data that we're dispersing around in our simulations.
[4:52] How simulations can help test autonomous Vehicles
Rob Proctor: Finally, the last major front in simulation market today is the need for testing within autonomous systems. Obviously, autonomous cars is a very important subject. Sooner rather than later, we're all gonna be sitting back and relaxing as the car drives us around. These things need to be tested out. Lidars need to be tested out. Swarms of autonomous aerial vehicles need to be tested out. What is the need there? What is the compelling event there?
Rob Proctor: For example. An autonomous car can't drive a million miles and do a million miles of testing, actual physical road miles. I mean, it can, just takes months, possibly years. Well, if you can run things faster than wall time. You can easily speed up those simulations and find issues and problems much more quickly. The same arises with swarms of autonomous vehicles. We're talking about 30 minute or less battery life for a swarm of a thousand different UAV's, doing all sorts of different applications, whether it be inspecting lines for power grids, or doing reconnaissance and surveillance, or checking out the health of environmental wetlands.
Rob Proctor: You need to be able to test those things out in a way that you do with real hardware. Real hardware, you might run into these again, 30 minute battery life requirements. That kind of defeats the purpose. If you can simulate that and then test out your network to do things in that simulated environment, that's clearly a direction that's happening in this space.
David Barnett: It seems then, common requirement across these different simulators is that they all operate in a very responsive manner that is really fast in a, faster than human perceived real time, so that to the human it actually is behaving in real time and is very realistic, this synthetic environment to the real environment.
Rob Proctor: It's funny you mentioned that. The thing I keep hearing in this industry over and over again is that a simulation is an abstract of a real system. Alright? It doesn't have to have definitive real time speeds, like a real time operating system might, like DDS is used in some fairly, heavy duty real time requirement stuff. It does need to be appropriate enough, so that it's usable by the end user in this case, generally speaking war fighters or folks testing of systems. I'd say, yeah.
David Barnett: It also seems like support for a lot of different types of networks, both local and wide area network is important, given it sounds like in the UAV example, you could be talking about line of sight-type networks. You could be talking about, global wide area networks, satellite networks perhaps in some cases.
Rob Proctor: Or all three, in some cases, absolutely. Thinking about STE, one of the questions that they're looking to solve right now, is the army going to use, Army Cloud or are they gonna use public cloud and use the police band and the bandwidth range for that. These pose security questions.
Rob Proctor: Normally, nominally a simulation environment is, there's an instructor. You're in a room. It's a local area network. There could be 30 to 50 people running through exercises. What happens now, when you have that simulation with another simulation of, exactly the same, another different local area network on a different set of interoperability standings. How do you make these things talk to each other over a WAN, or a secure WAN? That's clearly what we're facing right now in the industry.
Rob Proctor: What it's done, at least for me and Florida and especially Orlando, is the world has kind of come to RTI. I owe someone a nickel now, because I said RTI and not Real-Time Innovations. The world has come to Real-Time Innovations, because we meet these gaps.
David Barnett: How does Real-Time Innovations address these challenges?
Rob Proctor: You get to save your nickel. Right. We are great at security. Security for other standards is just non existent in the simulation field. We're incredibly great through our qualities of service that come through with DDS, to create reliability, to create more real time fidelity, over these integrated systems that are distributed over vast different locations.
[9:47] How does Connext DDS compare to legacy solutions used in the industry today?
David Barnett: How does the approach taken by the Connext product in the DDS standard, compare to traditional or legacy solutions used in the training and simulation industry?
Rob Proctor: That's a really good question. We have to first tease out, what are these legacy systems within the industry today. There is an organization called SISO, which is Simulation Interoperability Standards Organization. They maintain about 13 different interoperability standards, specifically for simulations.
Rob Proctor: The most legacy related one is calls DIS, which stands for Distributed Interactive Simulation. The idea with DIS is that there's these PDU's or program data units. With these PDU's, they act as API's. They represent different entities. They run over a common wire protocol. It's publish subscribe. It runs on UDP. There's no quality of service though. There's no notion of discovery. There's no security there. The latest version of that standard, for the wire protocol at least, deals with 1990's technology.
Rob Proctor: Fast forwarding from there, SISO also maintains the high level architecture. This what we're gonna talk about, why we say Real-Time Innovations and not RTI. HOA, is the predecessor or what comes next after DIS. It has a better encyclopedia of different federation object models. They run over something called, a run time infrastructure, and RTI. HOA RTI is a common term you hear in our industry all the time. What a wonderful acronym, Real-Time Innovations creates. HOA RTI's are not well loved. It turns out, an RTI is not interoperable with another RTI. These things are proprietary. They're a couple partners out there, that Real-Time Innovations have, who are HLA RTI providers. The real issue here is that these things don't talk to each other.
[12:03] Rob tells a story of the importance of interoperability
Rob Proctor: Now is ... I have to tell you a story. ITSEC is the premiere event for training and simulation. It's in Orlando. It's in December. Come on down to Orlando, if you've never been there before. It's a wonderfully large simulation based event. For the last three years, they've had something called Operation Blended Warrior.
Rob Proctor: At ITSEC. Operation Blended Warrior is really precursors to things like STE and SCAR. This is, we're gonna prove how interoperable our stuff is, sponsored by NAVAIR. Three years ago now, NAVAIR got about 50 different folks in the room, said, "Hey. Do you wanna participate in this? We're willing to run events live, virtual constructive events. People are gonna come in and we're all gonna interoperate together. Let's find some ways that we can all talk to each other.
Rob Proctor: NAVAIR said, "Hey. We want to use our NAVAIR HLA RTI. How many folks in this room of about 50 people can run on RTI?" Less than half a dozen hands went up. They said, "Well fine. How many folks can run on DIS?" Everyone's hands went up. There is a real problem with interoperability with HLA, even though that's the premiere standard in that, one RTI doesn't talk to another RTI. What good is having a federated object model that we all agree to? Whether it be FOM, or some other FOM of some kind. You can't have those things talk to each other.
Rob Proctor: That works great, if you're in enclosed simulation. If I'm running my HLA RTI and someone else is running their HLA RTI and someone else is running a DIS, and now I have to have these three things interface back and forth with each. Oh, by the way, I want security. How am I gonna do that right now? What operation blended warrior proved at ITSEC is that the industry still has a long way to go for interoperability.
Rob Proctor: Now, of course, everything comes off without a hitch. When you're there, you see the interactions. You see the simulations and it's kind of neat and they added to it for the next couple of years. Now this year, they're not running it. What you didn't see, is that it took months and months of integration and work to get those things to work together, whereas, what the industry is really looking for, is the ability to quickly in an ad hoc network, setup these things, have them interoperate with each other, shut them, move on to something else.
Rob Proctor: I'd say that right now, with existing standards out there, HLA DIS, you just don't have that kind of interaction. Talk about other efforts in this industry. There's couple of things I wanna mention. Besides HLA and DIS, there's something called, TENA, that the Test Training Enabling Architecture. It's almost test and training. It's TATENA, is what it should be called, but it's just TENA, T-E-N-A. They have something in there called, JMETC, Joint Mission Environment Test Capability.
Rob Proctor: The whole point of TENA is yet a third standard within the simulation community for mainly, this one's mainly for testing. Then again, you now have, funny enough, for those out there, who are actually listening to this podcast, hello. It runs on something called, CORBA. CORBA is a common broker, where all of your messages have to go through two hops. You're clearly right there, losing some fidelity.
Rob Proctor: Every message going through to TETENA, or just TENA, has to go through two hops to get to where it wants to go. Now, there's HLA, there's DIS, there's TENA, there's no interoperability between the three of these things. They're all publish/subscribe. None of them have qualities of service. They don't do reliability. They don't do heartbeat, liveliness. They don't do any of the number of qualities of service that DDS does. That kind of begs the question. What's gonna happen next?
David Barnett: So Rob, what does come next?
Rob Proctor: I'm glad you asked. Funny you should ask. The SISO community, a couple of years ago, came up with an idea. The idea is this layered simulation architecture, which is basically to say, "I'm gonna take your chocolate, give it to my peanut butter, and come up with something better."
Rob Proctor: In this case, the chocolate would be the existing federate object models. There's a library of different things that are used out there by TENA, DIS, and HLA. SISO, layered simulation architecture, is something that was brought up by SISO a couple of years ago. The idea here is that we're going to take the chocolate that is the, federate object models, all these federation classes from TENA, from DIS and HLA, PDU's from DIS, federate object models from HLA. We're going to put them on top of DDS. We have the common wire protocol. We have the interoperability. We have the qualities of service. In the future, we also have the security that comes with things like Connext DDS.
Rob Proctor: The thing with the layered simulation architecture, the LSA, is that it's ... Right now we're at this chicken versus egg phenomenon problem, whereas we've identified that this is something we wanna try. We'd like to actually turn it into a fully functioning standard within SISO. We've had a working group, we've had working groups. They said, “Let's go off and do it now." Then it kind of stalled for a couple years, because do you write the full standard, and then create a product, or do you create a product and then write a standard based off of that?
Rob Proctor: I don't have the answer to that, although I will say this, with talking with our simulation partners, it seems like they wanna see some sort of, a prototype first, and then they'd be willing to come and work with us in the SISO community to fill out the LSA. I guess, this is kind of how it went with OMG and DDS. We had NDDS first. There were a couple other folks with different things. We went object management group with these products and then created a standard around that. That's what's happening right now for future, within the simulation market.
[18:56] RTI’s experience in the training and simulation market to date
David Barnett: Can you tell me a little bit then about Real Time-Innovations' experience in the simulation market so far?
Rob Proctor: Right. Real-Time Innovations within simulation is nothing new. When I started doing research within the company to see what we've done in the past, there was a literal treasure trove of white papers, data sheets, design-wins, that date back well over a decade at this point in time. I know Dr. Rajive Joshi has written some white papers with Dr. Pardo. These compare DDS to HLA for example. For us, this is really nothing new. We've been involved in the simulation and modeling market for quite some time. We have a good, fair bit of design-wins. What really happens now, is that this need for Real-Time interoperability over distributed locations with security, has kind of created a sweet spot for Connext DDS within simulation.
David Barnett: Okay. Well, thank you Rob. I certainly learned a lot today, including a lot of new acronyms, and some to try to avoid. Where can our listeners go to learn more about the training and simulation market and RTI solution?
Rob Proctor: Well, I'm so glad you've asked. There's a couple things to talk about. First, there will be a Florida simulation summit this year, that will be attended by our own Dr. Stan Schneider. He's going to be a panelist. That's the first week of September, right here at the Orange County Convention Center. The mayor of Orlando's gonna be there. That's kind of exciting for us. We're expecting about 400 folks in the room.
Rob Proctor: More exciting than that though, is that this year at ITSEC, we will be presenting a tutorial on data distribution service. That tutorial will be Monday at 8:30. Andre Odermatt and myself will be presenting. Be sure to be there. If you tell me you listened to this podcast and say the magic word, Rumpelstiltskin, I'll give you a free eval of DDS.
David Barnett: Is that the only way to take advantage of that offer?
Rob Proctor: It's the only way to get me to laugh at you. No. It certainly is not. Right there on our website, of course, you can grab a full 30 day download.
David Barnett: Great. Well, thank you very much Rob.
Rob Proctor: Hey. It's been a pleasure.
Steven Onzo: Thank you everybody for listening to another episode of the Connext Podcast. Stay tuned for our next episode, where we start part one of our interview series with principal software engineer, Ken Brophy on RTI Connext Tools for distributed system development. Please contact us with questions or suggestions on social media, or at firstname.lastname@example.org. Thanks and have a great day.