Resilient Human-Machine Teams

Thanks for joining today's webinar, Resilient Human-Machine Teams, Building the Connected Force of the Future. My name is Vatche Arabian, Application Expert for Boston Dynamics. Andrew, you want to kick us off with some introductions? Yeah, thanks, Vatche, and thanks, everyone, for joining us today. My name is Andrew Wassef. I'm the Associate Director of Sales for our Government Division here at Boston Dynamics. Jake? Andrew, thanks for having us today. Jake McLeod, I'm the Business Development Director for Ground Robots at Persistent Systems. I'll hand it off over to Paul. Hey, thanks. It's awesome to be here. I'm Paul Hoffrichter. I'm a Project Manager in the Unmanned Systems with Persistent Systems. Well, hello, everyone. I'll start the presentation off. In the first portion, we wanted to highlight how Spot is transforming operational safety for our law enforcement and defense customers. And in the second portion of our presentation, our partners at Persistent System will talk about their network technology and how it enables Spot today and how it evolves this HMI feature for some of our operators. For those of you who are not familiar, Spot is our quadruped robotic platform at Boston Dynamics. It brings a high level of mobility, autonomous navigation, and sensor integration to field operations. So Spot's purpose is to walk anywhere in the world a human can walk. So it has the ability to navigate complex terrains, climb stairs, and carry payloads up to about 35 pounds. Because of this, Spot can go where many wheeled and tracked robots cannot go. So we're going to start with some of the Spot's most mission critical applications that we use today. Many of these applications you will see Spot used in a teleoperation manner, so meaning, there's a one-to-one control. An operator could use one of the tablets from Boston Dynamics or the RDC from Persistent Systems. So the first use case is around ISR, that's Intelligence, Surveillance, and Reconnaissance. Spot excels in perimeter reconnaissance or building clearing operations. It can give 360-degree visual feeds, thermal imaging, and then it also has a 30x optical pan tilt zoom camera. And it gives the operator situational awareness without exposing individuals. So in this use case, think of Spot as an amazing forward observer. The next use case is around EOD usage. Spot can safely approach, inspect, and manipulate suspicious devices using its integrated arm and high-resolution cameras. It reduces risk to bomb techs. It can carry and deploy EOD tools, like disruptors. And it's mobility allows it to get in position quickly, reducing the time to target. Lastly, we also have CBRNE use cases with the robot. It can be outfitted with radiation, chemical, and biological sensors. Spot can enter in a hazardous zone, stream data in real time, and create contamination maps, all while keeping personnel at a safe standoff. So beyond these frontline missions, Spot is also a powerful asset in support and infrastructure roles. And this is where you start to see some of the autonomous capabilities of Spot. In nuclear decommissioning and operation sites, Spot can navigate through contaminated GPS-denied environments. It can collect environmental readings, map the facility in 3D, have radiation overlay, and even move objects, especially in decommissioning efforts. This reduces the need for human entry. In the perimeter security use case, Spot can be scheduled to walk autonomous patrols. It can be along fence lines. It can check gates, quickly detect intrusions or some change in the environment. Unlike static cameras or wheeled robots, Spot can adjust routes in real time and handle rough terrain, stairs, or debris. And lastly, we use Spot also in mapping and exploration. So for base operations, train analysis, disaster response, Spot can generate a high fidelity map using LiDAR. And then these maps support mission rehearsal, site inspection, or it can create a digital twin of the environment. So those were some of the Spot's uses today. Now we can fast forward to what this functionality looks like with the human-machine integration. So in this example, you have a human-in-the-loop looking at multiple robotic platforms. It could be multiple Spots. You could have drones, other ground robots as well. And the operator is able to control multiple assets to perform different tasks or bring certain capabilities into any situation. This gives the operator two major benefits. One, it increases the situational awareness of the operator and simultaneously reduces the cognitive load on what this person needs to think about during these tasks. So when we talk about technology becoming a force multiplier, this is what we're talking about. So to accomplish this, we need powerful, secure networks that can connect all these devices in any environment. This is exactly why we partner with Persistent Systems. Their network architecture allows this critical data to be collected and reliably delivered in real time. So we'll play a little video that introduces Persistent Systems and who they are. And I will hand it over to Jake and Paul to talk about their technology in more detail and how it enables resilient human-machine teaming. [UPBEAT MUSIC] Hey, thanks, Andrew. So again, it's great to be here. And you can see a lot of these capabilities are very much in play in the now and both moving forward. And a lot of this stuff is empowered by a strong, robust network. And we're using these mobile ad hoc networks are managed for [? short. ?] That's where wave relay comes in. It's a dynamic, self-forming, and self-healing network that allows the network to move and be fluid with whatever operational environment that you might be in. So you can see you have a node or a user can leave the network and enter the network as much as is necessary for the operation but not have a negative impact on the operation of that network. It doesn't have a command node, so there isn't one specific appliance that is directing traffic. And it's not a hub-and-spoke architecture where all of the traffic is routed through a centralized location. It's truly dynamic. And you can see an example of-- it's simple, but an example of a MANET network in action on the slide on your right, where the gray dots represent those nodes. A really important thing to note, both in the now and moving forward with human-machine teaming, is scalability and security of the network. So wave relay is secured. It uses AES-256 encryption. It's FIPS 140-3 Level 2 certified through the National Institute of Standards and Technology, but it's also listed as a CSfC certified component with the NSA. So security is, across the board, especially in today's day and age, is very, very important. And that security applies to the network, whether it's a small network or it's a large network. So you can have networks four or five nodes, all the way up into the hundreds and it's going to operate just as efficiently as always. In fact, the more nodes you add to the network, the more powerful and robust the network actually becomes. So I'll pass it off to Jake to dig into some more of the details of how this can be used. Yeah, for sure. Thanks, Paul. So I'm just going to go over the why a little bit here. We look at traditionally used robotics and unmanned systems and we talk about human in the loop, which is that human in control. So a little example here on the slide is a warrant is being served or we're clearing a target building, something of that nature. And you have this singular operator who is actioning on that target with a Spot robot, and all its utility to bear, multiple missions it can use. We're using wave relay as the network radio to push the boundaries of that communications link, which is very important. But the why here is also the network. So what does that bring to bear as well? So we're going to enable that single operator to share across his network, his localized network there with his teammates, air assets, coordinations, and whatnot. So as a network builds this robust, this scalable data information, what we're really doing here is we're building information dominance. And that's really what makes these teams have a better edge with that decision dominance, by sharing across network. As that Spot goes in, the SWAT teams have situational awareness. The other patrol units, the air asset, they scale through and they see what the Spot's doing. Team sharing, team communication, information across the board. Go ahead and put through the animation. It'll give it a little display of it there, Vatche. So as you see, the network opens up locally there to all the assets that are on target. With our current capabilities, we also can expand globally. So as you add in cloud relay, that network now expands globally. Decentralized, disparate enclaves of local networks now become one global mesh network. And that network fabric can be balanced of how you pass information and data. So if you think about tactical decision-makers, strategic decision-makers, and how far back you want to push that envelope of data and decision-making helps these folks here on the edge make those decisions, those tactical decisions, those actions on even faster. So information is speed. Speed is effective. It creates efficient operators and more of a collaborative team on the edge. And there's another slide there that'll show how we patch through. Cloud relay provides that data backhaul to other connections. And then Paul's going to jump in and give us a quick slide on cloud relay. Thanks, Jake. So to really dig into cloud relay, at its very basic level, it is a way to connect to MANET networks geographically separated together into one network. It supports all of the capabilities that wave relay supports, so your data, your voice, your audio comms, tack, sensor data, anything that you can pass across. And it's transport agnostic. So that means you can use the local cellular infrastructure. You can use Starlink. You can use a hard line commercial line if you wanted to. So really, it's a way to extend that network globally, especially as global connectivity becomes more and more important as we operate more and more in different locations. A good example of this would be like a presidential trip. So say, he has an event in Houston, Texas. And for those of you who know, it's a huge man footprint, you know what I mean? So you've got the airport, you've got your motorcade route, then you have your event site. So you also have your operations center in Washington, DC that is wanting to monitor how these things are going, monitor comms, being able to talk to the team on the ground, look at video, record anything for the archives, whatever they need to do. So with wave relay being used in Houston on the event site, motorcade, route, and the airport, you'd be able to establish a secure connection through a commercial internet, and again, cellular, Starlink, whatever it may be, it doesn't matter, and that would connect that wave relay network to your operations center in Washington, DC as if it's on the same network. So you'd have seamless connectivity to be able to talk, monitor, watch video, whatever you needed to do. So at the end of the day, it is just extending that wave relay network out to wherever you need it to be in a secure manner. So I'll pass it back over to you, Jake, to talk more. Absolutely. Thanks, Paul. And we're talking about our current state and how really the end user is utilizing a network and these different platforms and these capabilities. The next step here is to look in the future. So you have a robust network that connects all these robots. What's the so what here? So where do we go from here really? And I think the user really needs to start thinking about collaborative behaviors. If you have multiple robots that do multi-mission functionality, which the Spot is totally capable of doing and has those current states, how do you have a human in the loop versus on loop? And when we dig into that, it really means the task saturation, the cognitive load on a single operator control and singular robots, singular actions. What we want to do in the future states are we want to put humans on the loop of that and let the collaborative behavior really problem solve. So it comes down to more efficient teams with actions on target that are more asymmetrical by that decision dominance, really. And what the network allows for you to do is develop that critical thinking autonomously through those robots on target and create those behaviors that allow you to manage the task and the outcomes versus driving a robot. So I really think the future state is going to come with multiple robots, not just ground robots, but air assets, unmanned, subsurface. You want to have a wide arrangement of hybrid robots that are able to accomplish a task given by those human-on-the-loop managers and decision-makers and let them do more things that humans are steered towards doing. Let the SWAT team run guns. Let the strategic commanders make their decisions, do risk management, things like that. So really, what we're trying to do is here, make the human team, through this human-machine team, we're trying to take that human operator, the human decision-makers, we're trying to make them more efficient, better at their job, actions quicker. So to do that, you really need a robust network that allows you to connect those robots. As they start doing collaborative behaviors, they need to be able to problem solve, that through the network, they're going to be able to talk across that fabric. And we want to be that connection across that fabric that allows them to problem solve. I need to see a CBRNE reading. I need to see ISR. I need to see different payload capabilities. They can collaborate. They can help make decisions, pass data back to have an overarching strategic decision made on the next actions on correct. And there's a little example I think that links them through on the slide. But it also reduces risk. You think about we can have more decision-making, more impact with operators, we can also reduce their risk. Bring them further back from the target. Let the robot be that first risk reduction there. Back over to you, Andrew. Thank you. Thanks, guys. So if you found the information useful today and you want to talk to us more, obviously, after this webinar, you can get in touch with either Persistent Systems or Boston Dynamics. We work together quite closely. If you'd like to come to one of these shows and maybe see this technology or talk about it live with our teams, both of our companies will be at these two upcoming shows. DSEI which is a defense show in London. Or if you're stateside, we are going to be at AUSA in Washington, DC in October. So come find us at one of those shows. We'll both be there. Otherwise, you can contact us after the webinar. It was a double homicide, suicide. [SOMBER MUSIC] This is one of those incidents that it was an apartment. And when the decision was basically made after negotiation, we weren't getting a response. At the time, nobody actually knew that the subject had already taken his own life. But basically, in this situation, once the decision was made to actually get eyes into the structure and see what was going on, this was one of those that the feedback that I got from the IMT personnel that were there, that were actually operating Spot is Spot cleared that structure in less than 6 minutes. Went into every room, was able to see everything, traversed all the obstacles that were present. It was able to identify the deceased in the structure. And basically utilizing that information, the decision was made that let's send in the SWAT team, finish up, clear the structure once we knew that the subject was no longer a threat? One of the problems with a lot of the ground robotics that exist on the market today is, especially the ones that have [? arms, ?] is they have difficulty trying to climb stairs. Some of them cannot climb stairs. And the ones that do climb stairs, a lot of them don't have an arm to manipulate opening doors. That's one of the things that just having Spot had basically been a piece of mind and then that ease for us knowing that I can actually overcome this. There's a great reduction in fatigue utilizing a robot like Spot simply because of how fast Spot can operate and clear obstruction. So you can imagine, historically, if you're waiting for the robot to clear a house and it takes upwards of an hour, that's an hour for all the SWAT personnel that are on the perimeter. You could be baking in the sun. It just adds to the amount of fatigue wearing your vest, holding point. Spot makes things a lot faster. [DRAMATIC MUSIC] Maybe something to mention there, Vatche, that that testimonial came from Lieutenant Rego from Miami-Dade. And Miami-Dade just got their second Spot robot last week. So we were really pleased and happy that the first one proved useful enough that they expanded their usage of Spot. Awesome. Thanks to our presenters. So we've got quite a few questions. If you've been using the Q&A button, thank you. If you've got questions, please use that and we'll get to as many as we can. To kick things off, we also had some presubmitted questions, so I'll start with one quick one for Andrew. What are the best ways to address the concern of human-to-robot replacement? Yeah, that comes up quite a bit, Vatche. We've always heard that we don't want robots replacing personnel. I think Jake articulated it really well with his example, is you're looking to reduce risk for our operators, not reduce personnel. So they're making better decision-making and reducing risk. That's what the technology is used for. We haven't seen any reduction in personnel from drones or other technologies that get introduced. This is about increasing capabilities. Yeah. There was a question that came in through chat. If I have a Spartan robot and a PackBot robot that work with MPU5's, can Spot be used in that same network? Yeah. do you guys want to answer that, PS team? Yeah, absolutely. Once you're connected, if it has a wave relay system, if it's part of that network, it takes a little bit of administration coordination between those teams. But yes, they can be on the same network. Absolutely. Cool. Speaking specifically about wave relay technology, is that only sold to local, state, federal or military agencies or is it available to private companies that use Spot for other applications, like perimeter security or plant inspection? Oh, absolutely it is offered. And we do have private companies that do use our radios quite often. This came in at one point. It's just a clarification. When you say globally, does this work all over the world and not just in the US? I think I know the answer but wanted to ask you. I'll let Paul take that one. Oh, absolutely. All is required is a connection to the internet, effectively. So connection to the internet and the right configuration, which is simple. And yes, it works globally. Yeah, mentioned secure connection as well. So examples of that is you can plug into Starlink, satellite connection, landline, even hotspots, 5G. Anywhere you can connect to the internet, you can plug-in through cloud relay. Cool. This might be a little in the weeds, but I thought it was an interesting question. Does Spot come with a set IP from the factory? How easy can you change in order to get multiple Spots on the same network? Who wants to take that one? I'll start. And maybe the PS guys can piggyback off of it. Yes, so Spot comes with configurable IP. It does come with defaults from the factory. But you just set the IP onto the radio's network and then you can match the controls to be on that same network as well. So I don't know if they need more details than that, but it's all configurable. And same on our side. Our radios are configurable. They're completely tailorable to whatever your network is. So like I said, we were talking earlier, different hybrid robots being able to be on the same network, as long as those IPs are matched and administratively coordinated, those teams using those robots or the robot on the same team that are different, yeah, they'll talk, absolutely. And if you've got follow-up, too, I think we'd be happy to talk in more detail about the technical side of how to configure robots. Why are networks critical in human-in-the-loop operations? Maybe one for Paul? Well, so networks are critical for human in the loop because that's how you get your situational awareness. That's how you control your unmanned systems. That's how you share that data across the team. So having a robust and high-speed network that connects not only your users, but also your unmanned systems is crucial, especially if your organization wants to maintain that awareness of their environment or where they're moving into. It also allows your unmanned systems to talk to each other as well for even more coordination. Yeah. The critical and contested environments, that's what it all comes down to, how well can you stay connected. We know what happens when that robot isn't connected anymore, if the human is in the loop, you're basically going to abort that mission. You're going to try and get connection. Either you're going to use lost linkage logic to bring that robot back or the operator is in a scenario where they have to go and get close to the robot, get through electronic warfare, get through range complications, reconnect to that robot. The more resilient you are and the robust that network and that fabric, that radio connection is, that's your success rate. It really depends on how you stretch that. I guess to follow-up on fabric, we had a presubmitted question and then another question that I think I can tie these two together. So are fabric networks strong enough to provide live video for multiple sources back to command? And how far can the tactical command or command post be from assets in the hot zone? And then to tag on to that, how well does the system work in something like an underground tunnel? Because this individual says that they're planning on using Spot inside metro tunnels that are coming online in their city and that's a consideration for them. Yeah, absolutely. We have great success. We've mapped through the English Channel. We've done multiple mission sets with DOD entities and teams partnerships doing sub-t terrain, not just training, but in real life, too, has been used. So we perform really well in subterrain. And that question is very open-ended and loaded when you ask how big that network be and grow. What we do well as a networking company is that we scale. That is our bread and butter, we scale well. So I would say that the more nodes you have, the larger network, the better we are at passing that information and creating those connections. Cool. There's a question about rad mapping. Just what sensors can folks use for that, Andrew, and what kind of level of detection? Yeah, there's a bunch. If it's like a RadEye, or a Pendar, or there's probably 30 or 40 sensor companies out there. Spot has an open API. So usually, there's hooks to connect that data coming from the sensor and pass it through where it can show up on the tablet screen or get the information out to another area through those communication channels. So I would say there hasn't been a sensor that we haven't been able to integrate. But it's very specific because there's a lot of them out there. So if the person asking the question has a specific one that they have in mind, they can talk to us about it and we can give them some specific documentation. Yeah, and just to add on, I think it's very common, like customers will already own certain sensors and want to integrate their sensors. And we've seen a lot of folks do that. Do you support tactical data links or do you envision supporting them in the future? That is a great question. As far as-- I guess, [CHUCKLES] to ask a question in return, what type of tactical data links are we talking about, like DOD, Are we talking like Link 16, Link 11 types of things? That I don't have a good answer to. Any kind of interactions with tactical data links are going to require a lot of coordination with the Department of Defense and the owners of that system when it comes to connecting to that network. I think it's safe to say that if it's data passed in a traditional sense, we can pass that data, we can provide the throughput, the pipe. It just comes down to coordination. And with our government certifications, NIA, CSfC, that makes it even more so compliant to a lot of those government and DOD data links. So it really just comes down to coordination, this data. Cool. Has the PT 5 been used on Spot for cloud relay? We have tested out PT 5 with Spot. It's relatively a new product. But I think we've had good results from both teams working together. The PT 5 is fantastic. It really resonates well with your domestic market space, your SWAT teams, your bomb, your fire, your rescue because you have infrastructure already in place with these cellular networks. And then what the PT 5 provides is that throughput, that data pipe right through your common network. So you're not really trying to do your line-of-sight radio-to-radio anymore. You can incorporate the PT 5 in that and you can start transporting data out using available infrastructure. So it's going to be a phenomenal tool for those domestic markets to markets use. Andrew, are there easier options for financial aid other than grants or can grant options be discussed? Yeah, especially in this world where a lot of the procurement is around procurement cycles or in a tender process. And if you're not in that cycle, people ask, how do we get a Spot? We work with a company called [? Gov Funds ?] that help us locate sourcing state and federal ways to get grants to help collect donations. Some departments have done private funding. But, of course, that's always tricky and difficult to have that scenario set up to get some of this equipment. So if you talk to us and talk about what we're looking for, we use those companies like [? Gov ?] [? Fund ?] to help try to locate some sourcing of funding. Follow-up on the PT 5, does it work with cell providers outside the US? It does with certain cell providers. I don't have the list available of all of the overseas providers that have been tested with. But that work is ongoing. Cool. And obviously, if follow-up questions like that, we're happy to follow-up with you offline. I think we'll take a couple more. So what specific lessons have emerged from real world deployments involving human teams and robots in degraded common environments or low connectivity environments? And how have those insights shaped the current design of mobile networks for tactical operations? Yeah, I think the largest lesson out front is that the future is going to be collaborative autonomous behaviors. So those far-reaching networks that stretch the boundaries of connections, it's important to have a robust network, but it's also important to have these collaborative machines be able to connect to each other and work together on the edge. And not only that, but you can't have one or two robots fall offline and then this collaborative behavior, these models go offline. So for example, if I have five Spots, they're doing a collaborative behavior model, an agent type model around a target. And you have two of those robots fall offline or they come disconnected. The whole system can't fall apart. They can't be impinged on one or two passing data back and forth from each other. And that's why I think we're positioned really well as a product to support that. One, we have a robust network. Our waveforms pass data very well. We hold the bandwidth. But also on the edge, we're not dependent on singular nodes to hold together that glue, that network together on the edge when you look at those multiple platforms. So I think that's important to consider not only just-- we get wraparound range and distance bandwidth. You have to also look at the future of collaboration and how these robots are going to interact with each other. If you chop one leg off, is it going to continue to move on as a group or is it going to fall apart? So that's really important to think about. Yeah, maybe I'll add to that, Jake, a little bit from the robotics side, we're also looking to increase the capability, like the individual capabilities of the robots. So some of the stuff that we do, we have to do teleoperation today and help guide the robot to do, those things in the future, we like to get that collaboration essence that Jake is talking about, we want to do some of those things autonomously. Like, be better at opening doors, being able to go across environments, get into a building without direct input from people on the ground level. So I think from the robotic perspective, we're getting better at that all the time. And then to the point of the collaboration element across all those different systems, that's where PS is going to help us. Awesome. Yeah, great answer. I think we're going to take two more. There are multiple sensor partners that Boston Dynamics work with. Does each sensor partner develop their own core? Or do you use or only able to use the Boston Dynamics core? Andrew, do you want to take that one? Yeah, we've had people use different type of edge computers on Boston Dynamics, different types of sensors. We're very much concentrated on the robot platform working very well over time. So we can only test a limited number of computers, and sensors, and things. We don't have infinite engineers and infinite time to test everything. But we try to go where the customers have the most need. So if they say, we want to test this particular configuration, we spend a lot of time making sure that it works reliably. Because the idea is that it has to work at the moment of need. You can't have stuff work only 70% or 80% of the time. It's got to work 100% of the time. So for us, yes, you can use your own custom bespoke equipment. But, of course, as a company, we want this stuff to be up and running like 100% of the time. So we can only test a limited number of those things. Yeah. This is a perfect question to close us out. Who should we contact for procurement, defense procurement? Is it Boston Dynamics? Is it partners? Andrew, you want to take that one? Contact us at Boston Dynamics. We're constantly in collaboration with our partner ecosystem, companies like PS, because we're always trying to marry the technologies to be more effective. And if you contact PS, the team is great about funneling stuff back to us as well. So if it's comms related, you can start with PS. And if it's robot related, you can start to Boston Dynamics and we'll find the right path. Awesome! Right, and like Andrew said earlier, we're going to be at DSEI in September and AUSA in October. So if you're interested in seeing us in-person, definitely catch us at that show. If you're watching this on-demand and that time has passed, you can always reach out direct using our contact sales form. We're happy to answer any questions, any follow-up questions you have about the robot, about the radios, and any other questions you have. So thank you, Jake, Paul, and Andrew for your time today and your presentation. Thank you to our audience for your questions. And we'll see you next time. Thanks, Vatche. Thanks, Andrew. Thanks, Vatche. Thanks, Andrew. Thanks all.