Joint Simulation & Training 2010: Graeme Duncan

Contributor:  Graeme Duncan
Posted:  06/22/2010  12:00:00 AM EDT
Rate this Article: (4.5 Stars | 2 Votes)

Defence IQ interviews Graeme Duncan, the CEO of Caspian Learning. Caspian Learning is the developer of the Thinking Worlds software platform that enables users to engage in a 3D virtual simulation designed to train and teach, particularly in the military realm.

Could you tell us which organisations and divisions you have worked with, and what sort of situations you have virtualised for these bodies, particularly in the military domain?

Sure, we have probably deployed the widest range of simulations and 3D immersive learning environments of any serious gaming or learning developer out there. And we have developed well over fifty individual simulations and in sectors as diverse as pharmaceuticals, motor and, which you pointed out, for defence. We have worked with major systems integrators such as IBM, Accenture, Price Water House Coopers. We have also worked in the education sphere with the Ministry of Education and regional bodies of education. But within the defence realm, or the security realm in particular, the ones that spring to mind are…we have worked in the UK with are the Defence College of Policing and Guarding, and for that particular group we have created 3D immersive simulations to enable them to put their guys in realistic scenarios where they have to go through entry search procedure processes, deciding how to enter the premises effectively, how to search for evidence effectively, and how to go through the process of seizing that evidence and making sure that it is usable if it is needed in a court of law.



In addition to that, we have worked with the UK Royal Navy last year in November. A simulation we developed for the Maritime Warfare School in Collingwood won the E-learning Age award in the UK. And that was used to take brand new navy recruits, and give them onboarding training in preparation for them going onboard a Type 23 war frigate. And then the key subject matter of that would then be to carry out health and safety and weapons rounds inspections. And we created a fully accurate, Type 23 warship, and created a simulation, but also a serious game in that environment, where there is a saboteur onboard and they’ve got to go and test their knowledge. We have also worked with the Defence Centre of Training Support here in the UK, with the Defence Academy in the UK. And in the US, through people licensing our technologies, we are working with the DoD, we are working with the Defence Security Arena. And a number of police forces throughout Europe as well.

So how realistic is the Thinking World’s platform–and what perhaps can be improved as technology continues to advance to make it as accurate to the real deal as possible?

Very good question. We attempt to make the simulations and the immersive learning environments that we can develop as real as they can possibly be. And that means three things. We tend to focus on the graphical fidelity when we talk about realism–how does it look on screen, how are the shading and the rendering package making it as realistic to our eyes as possible. And that is important, and we manage to add new shading capability, new rendering capability to enable the 3D environment to look as good as it possibly can. That part of realism is constantly evolving–we are constantly trying to keep up with it.

We have ray training software that is being launched. You have more powerful chips that are being launched, so from a 3D point of view, we try to go as far as we can. However I think that that is quite a myopic view of realism, to create a simulation that is highly realistic you also need to create scenarios and simulations of events in that learning environment that are highly contextual, and highly representative of what happens in the real world. And quite often we have seen simulations where you jump out to 2D overlays, or we jump out to other bits of learning, which is not how it occurs in the real world…which is not how [our] simulation occurs.

Through a vehicle checkpoint, you want to check someone’s virtual awareness as cars are arriving at a vehicle checkpoint. You need to place a contextual realistic environment which may look like a vehicle checkpoint in either Iraq or Afghanistan; you need to get the cars to arrive in exactly the same mode as they would. And then you need to get them to approach the vehicles, and role play those conversations exactly as they would in the real world. And then the third part of realism is the problem solving exercises, the decision making points that you give them need to again represent those key decision points, and those key problem solving areas that someone is faced with in the real world.

Another simulation that enabled us to do this that springs to mind, is that we have done a lot of work with the European Union customs and excise team, creating simulations where they search for drug precursors coming through borders, and in that scenario they had a variety of different investigative techniques that they can use as operatives on those land borders, such as x-ray machines, such as looking at the seals on the shipments, looking at the docking shipments of the lorries or the people that arrive, choosing to do full lorry x-rays using sniffer dogs. So, within the simulation we try to assure as much as we can that we give them all of the different investigative techniques that are available to them in the real world, in the simulated environment. Therefore they are not hampered by not having anything that they would not normally have in their day job, and we can therefore focus on using the right investigative technique, as the scenario unfolds and as the simulation unfolds.

So I think we try to make them as realistic, we try to make them as contextual as they can possibly be on the 3D graphics side of things. I think yes, we can continually improve that. You look at lip synching, you look at facial animations, you look at body animations, these are areas that we strive to drag into our simulation platform. But we should also look elsewhere in terms of realism and make sure that we are creating the scenario, and the simulation of events and decision points, and thinking like that enables us to do that in a highly realistic way.

It does sound like it is a very difficult thing to get right. Just how open is the coding for this platform or other platforms? I mean how much can your partners actually customise their experience throughout the process?

Another great question. How open is it? I think it is one of the most open platforms out there within simulation development. It is unique. It is not an entertainment games engine that is being used to create simulations or serious gaming, and it is also not an open source engine. There are other open source engines out there such as Delta 3D, where people are asked to go and code and programme. We have focused on a different level of author. We have enabled an authoring that enables structural designers, trainers, subject matter experts, people out in the field to be able to create these highly realistic 3D simulations for themselves without having to write a line of code or a line of script.

Now if I am someone who has a penchant for scripting, then the scripting interface is a very simple c derivative, and therefore I can go and create animations, interactions, events, that are not already in the authoring platform. So I am free to do what I want. I can take data in and out of the platform, using web services, using java, using PHP. So I can feed events into the simulation using other simulation engines. But I think where Thinking Worlds is fundamentally different, is–and we only launched it as a commercial license platform pretty much twelve months ago to the date–it enables a different type of person to be able to create these simulations, and that is something that we are critically focused on with Caspian Learning. We wanted to start to put people that are in theatre in the defence arena–instructional designers, subject matter experts, trainers, teachers–back at the centre of the design process. So rather than them creating story boards and handing it to physics programmers to then create algorithms, we wanted to be able to open up the number of people and the type of people that can create and design these experiences. 

Iif I’m an instructional designer, I can work with a scripter in Thinking Worlds to create brand new animation, brand new interactions and events within the games engine. But within Thinking Worlds, choosing where cameras sit, how long cameras look at particular objects or avatars, where the non player characters go, what animations I put on them, what events I trigger by them, what interactions I put at different points of the simulation, can all be done without having to right a line of code or a line of script.

So when you talk about how open is the coding platform, again, we probably tend to focus on “is it open sourced or is it a bit of proprietary platform?” Well we see our platform as extremely open because script and others feed data straight into the engine, but another view of openness is Thinking Worlds opens up the ability of a brand new developer community, and instructional designers have been chomping at the bit to be able to start to create 3D simulations for themselves without this long development process, constantly writing storyboards, going back to the physics programmers, seeing what they put on screen, saying “no, that’s not quite what I needed”, subject matter experts reviewing it, sending it back again into this box of c++ programmers that have to burn the midnight oil to try to get the algorithms right. We have opened up the platform to enable these people to be at the centre of the developmental process themselves.

And in that sense what sort of feedback have you received from your military partnerships when actually looking to tailor future simulations to accommodate the frontline operational experience?

Yes, I mean there is probably feedback linked to the simulations that we have deployed, and then there is feedback based on the engine itself. So let me deal with them in that order. Feedback simulations that we have developed, first. The two that spring to mind that have had the most robust feedback in terms of learners using it, learners feeding back on it, and some level of learner measurement versus either a control group or versus another mode of learning has been very encouraging. For example, the simulation that I talked about earlier that we did for the UK Navy, the Maritime Warfare School down at Collingwood, have done some evidence tracking where they took cohorts that used the simulation and then they took cohorts that went through a more traditional normal onboarding process before they got onboard a Type 23. And one of the measures that they looked at was a measure called BUTS, which stands for “Brought Up To Speed."

Now, there were two or three parts of the core curricular where they had to go through a significant level of BUTS training, now that means having gone through the class training or the traditional e-learning training, and then being put on a ship. In a number of particular subject areas they had to then go back to this, if you like, remedial “Brought Up To Speed” training. And they tracked a cohort of just under a hundred learners without the simulation, and then they created a cohort that had used the simulation, and looked at how the demand or the need of BUTS training had dropped. And they also compared it to two courses that were other problematic courses. 

What they saw with the simulation was a reduction by over 50 percent, in the need to carry out BUTS training in the cohort that went through the Maritime Warfare School simulation, linked to weapons rounds. Now the cohort was actually larger with the simulations, so whilst the reduction was over 50 percent, it was even better than that because the cohort of patients–or the cohort of learners should I say–were even higher. So you would expect there to be more people needing to go through that BUTS training purely as a result of the cohort being greater. So some of the hard and fast feedback, i.e. “quantitative measures,” is looking very positive.

Then you look at some of the qualitative measures, you look at the learner feedback–would they prefer this level of learning versus the traditional level learning? Do they feel more comfortable and ready to board the ship? And that has been an undoubted yes. The same is true of The European Union border control. Some quotes back from different European member states include “this heralds a new version of e-learning, a new version of immersive learning,” and therefore I think both quantitative and qualitative measures are starting to show that these types of simulations not only engage, motivate and appeal to the learners, they also have shown hard and fast demonstrable return on investment versus the other modes of teaching. So I think as far as the simulations themselves go, we have had a very positive feedback.

I would say, however, that the military in particular, and rightly so, are pushing us to increase and constantly improve the level of realism, constantly improve the ability o get these simulations out wherever they may be around the world. The days of bringing them into a structured IT suite are slipping away from us. We need to get the learning to wherever the guys are, and that is often in theatre, and we need to do it quickly. So I think the demands are not always necessarily linked to the simulation content, it is about how do we get it to our guys in a timely manner wherever they are in the world. So that is the type of feedback we have got linked to the simulations. Then linked to the technology itself, I would say this wouldn’t, but it has been nothing short of remarkable.

I think what is clear is that there is no one, and never should there be, simulation or games engine that solves all of our learning ills. There is a need for different simulation technology, both authoring platforms and delivery platforms to deliver different types of learning experience. And that could be the difference between single player scenarios and multiplayer scenarios. It could be the difference between the need for high-end physics, including ballistics, for battle scenarios versus the need for high fidelity animations when looking at cultural or body language training. So no one engine should ever solve all of these issues.

Where our engine comes in, is that it fills a gap. There are many engines that deal with what I would call the high-end simulators, creating new aircraft simulators, tank simulators, and there are more and more virtual worlds coming to the fore, gathering people in different locations together. But where our engine comes in is it deals with eighty percent of the training that you have to do with their guys, be it procedural, be it process, be it decision making or problem solving scenarios. Thinking Worlds enables us to do that, but it enables us to do it at far lower costs, and far lower timescales, and then send it through architectures in what is called a “thin client” way.

We don’t have to have things loaded onto machines, we don’t have to talk to servers externally, or even servers in a particular country, in a particular time. You can consume these 3D simulations as you would any other Web browser or Web page. And I think what the military are seeing here, is that this is needed. It isn’t going to replace all of these other simulation capabilities that they have, but it is another needed extension as it helps them get more training to their guys where they are, within theatre. And I think that our battle as an organisation is to continually understand those technical restraints that our military clients work in, and work tirelessly to make sure that we make the developmental processes as rapid as possible, and the ease of deployment as simple as possible.

Great points. Just on a final note, Graeme, what is your perspective on where you see military simulations operating this type, heading into the future, and perhaps how quickly we will see the face of this field changing and progressing?

There are a lot of in vogue phrases out there, so let me try to just give you two or three of the key movements that I expect to see, because we are looking them as an organisation and we are seeing them used in other proximate markets or some sectors. The first one would be undoubtedly cloud-based simulation delivery. So what does that mean? It means that the games industry is already moved to an arena where you have large server farms that do all of the processing, and therefore reduce the demand on the local PCs or the local handheld devices, and those handheld devices or PCs can just render the physics and the logic that is coming out of that large server farm. This means therefore that we should be able to deliver high quality fidelity 3D realism down our connectivity tubes and bandwidth to pretty standard kit. So I think cloud-based approaches are certainly going to come to the fore more and more.

As a precursor to that I think that we are going to see a lot more thin client delivery. We are probably leading the way there, but there are other organisations looking at it. And I think that thin client delivery again enables us to deliver what the military want, which is bite sized chunks, delivered rapidly, wherever the guys are. I think the third area that we will see developments in will be in the area of realism. I talked around ray tracing, and you can see some early deployments of that type of stuff around the world. And I think that ray tracing will enable us to create 3D visualisation with 3D scenarios, where the level of realism is so close to real life – the way that the shadows are thrown, the particles and the way that hamper troops line of sight–I think will continue to advance, and within the next year or two we will see significant improvements in those two areas.

Graeme Duncan Contributor:   Graeme Duncan


comments powered by Disqus


Advertise With Us

Join Defence IQ