Chapter 11

Defense Advanced Research Projects Agency Development Center, Nevada
June 17, 2005, 1400 Pacific Standard Time

Charles used the time waiting for people to arrive to review the plan he had developed with Alan's help. He had tried without success to persuade Alan to take the job as his deputy, right up until the day he left for England.

It was often the case, you lost the people who were or could be most valuable to a project, and those you would happily lose, stayed no matter what. Getting and keeping the right people, and the right combination of people, was an important part of Charles' job. He was far from sure he had the right people to make MAADS work, or that he could keep them, never mind getting them to work productively together.

Charles began the weekly MAADS team meeting in his usual way by summarizing the topics they would cover. “This afternoon I'm going to outline the capabilities we need to make MAADS a reality. I will then cover responsibilities and the schedule.

“What I have done is assign those capabilities into three different categories. Category-one are improvements we need to make to things we have already. Category-two are things we can buy or obtain from elsewhere. Category-three are things we have to design and perhaps build ourselves.

“Category-one improvements to what we have already, are primarily upgrades to the tactical mobile robots and the aerial vehicles.

“We are planning several improvements to the tactical robots. Their current power source is three twelve-volt car batteries. We will add an auxiliary power system, another three batteries, both for redundancy, and to give them a longer duration between recharges.”

Alan asked, “With the extra batteries, how long will the tactical robots go before they need a recharge?”

“Sammy, can you answer that question?”

“It depends on how much moving around they do. The robot’s electric engine drains the batteries. The cameras and computers, in comparison, consume relatively little power. I haven't run tests, but a rough calculation is about a hundred and fifty hours between recharges, if the tactical robot doesn't use its engine.”

“A week without a recharge is impressive.”

Charles continued, “We need to add a weapon to the robots, and then integrate the weapon with the targeting software, but I'll cover that as a category-three activity. Otherwise, apart from adding an adapter for our data network, which is straightforward, we have a program of incremental upgrades to improve the tactical robot's software. Once we have the weapon integrated, we should be able to freeze development at any time.

“We need to add a number of capabilities to the aerial vehicles. The most important is to use a fuel cell as the power source, which I will cover as a category-two activity. Other things we need to do are, install the same adapter for our data network, and integrate the autonomous search software we developed for Porcupine. I think this will require a second processor, which unfortunately adds more weight. I'll cover autonomous launch and refuel as category-three tasks.

“We have three category-two activities, things we can purchase or obtain elsewhere.”
Charles referred to his notes in order to see where he was on his list of topics. “The first of these is the control unit's power source. We need to incorporate a power source into the MAADS control unit to drive its systems, and to recharge the tactical robots.

“The second is the aerial vehicle's power source. The good news is, small fuel cells are a hot research area. A number of groups, here in the US and overseas, are working on fuel cells that are a similar size and weight to current batteries, but provide substantially more power.

“I am hoping it will just be a matter of finding the one that best suits our needs, then ensuring an adequate supply. The Lawrence Livermore Labs are producing prototypes of a small direct methanol fuel cell with replaceable fuel canisters. They are small enough to use in our aerial vehicle. You replace the fuel canister by unscrewing it and screwing in a new one, which makes replacement by an autonomous robot, a viable option. The fuel is a nonflammable methanol water mix, which should satisfy the military's safety issues.

“Samsung in Korea and Fujitsu in Japan have fuel cells designed to power a laptop, which is the size we need for the aerial vehicle. I’m trying to get evaluation units, so far without success.”

Fiona asked, “Why a mixture of methanol and water as the fuel?”

Charles replied, “It’s specifically to make the fuel nonflammable. The water is inert, and doesn’t produce power.

“Originally, I’d intended we'd use the same power source in the control unit, so as we minimize the number of technologies we have to deal with, but it appears a propane-powered fuel cell is a better option for the control unit's bigger power requirement. They are available from a couple of US companies as evaluation prototypes. Apparently, the main market is as a power source for Recreational Vehicles. They are bulky about a cubic meter including the tank, but Sammy assures me we will have the space. Unfortunately, propane is flammable, which is an issue for the reasons I raised earlier. If it comes down to a propane system, I will have to take it back to DARPA for them to decide.

“The other category-two item is an encrypted wireless data network. This is an off-the-shelf military specification tactical Internet with a four-megabit per second capacity, that's about a half a megabyte bandwidth. It supports real-time video, and is more than adequate for our purposes. The manufacturer's specification says it works up to forty kilometers, but the range is lower in urban environments and mountains.

“The network controller goes in the control unit, and a network adapter card and aerial go in each of the tactical robots and aerial vehicles. The adapter and aerial are small, together weighing less than a hundred grams, although that's still significant to the aerial vehicle.

“Does anyone have any questions on these activities before I go onto the things we need to design and build ourselves?”

There were no questions. Everyone was clearly anxious to hear who gets to do what. Charles was well aware that category-three activities were the 'interesting stuff' where people got to do original design and development. Most of the decisions on 'who worked on what', had been straightforward, but one he had discussed with Alan, and debated with himself for most of the previous week.

“I have identified five major components, we need to design, and in some cases build, ourselves in order to achieve the MAADS package. I have designated each of them as a sub-project, and given one person responsibility for delivering the sub-project.

“If any one of these sub-projects fails to deliver then MAADS as a whole fails. If any one of these sub-projects is late then MAADS as a whole is late.“

Charles was well aware that should one of the sub-projects fail, the blame would attach to him, irrespective of who was actually at fault.

“The first sub-project is a command and control capability. This is a software system we can give general commands to, and it will work out the operational details. For the computer gamers here, it is a lot like a computer gaming engine that has a number of scenarios and a set of rules built into it.

“The problem is not as difficult as it might sound. The command and control software directs the tactical robots and aerial vehicles to certain places in response to more general objectives we provide. As an example, we give MAADS an objective of defending a particular geographic area against infiltrators coming from a particular direction. The C2 software then determines where to send the individual tactical robots and aerial vehicles using its preprogrammed scenarios and internal maps.

“The C2 software will also know what to do in response to various events, like when a robot is lost or damaged. Richard and Jennie are confident we can get the software up and running by the delivery date.

“The second sub-project is a refueling facility for the tactical robots and the aerial vehicles. Obviously, we have to decide what fuel cell we use in the aerial vehicle before we can determine how to refuel it.

“I like replaceable fuel cartridges because refueling will take a few minutes at most, whereas recharging batteries will take hours. Cartridge replacement would make a single aerial vehicle MAADS system, providing continuous aerial surveillance, a viable proposition. A single aerial vehicle system will solve some problems that Sammy and Fiona have identified in building a two-vehicle system.

“Until we resolve these issues, Sammy will concentrate on recharging the tactical robots, although for a short deployment, we could get away with not recharging the tactical robots.”

Arjay asked, “Why don’t we use a fuel cell in the tactical robots?”

“Why fix something that isn’t broken? Conventional batteries work just fine in the tactical robots.”

After his abrupt dismissal of Arjay's idea, Charles returned to his agenda. “The third sub-project is a launch mechanism for the aerial vehicles. I think we all agree, this is the hardest problem we face, and the one I am least confident of being able to deliver. Fiona has several ideas how we could achieve autonomous launch, and that in itself concerns me because it means the solution is not clear to us. I am thinking of asking DARPA to provide us with an outside expert who will help guide us to the best solution.”

Fiona asked, “What happens if we can't deliver a launch facility for the aerial vehicle?”
Charles hesitated before answering. “Without an autonomous launch facility we have no aerial surveillance capability. Without an aerial surveillance capability, we can't position the tactical robots in the path of infiltrators. The robots would have to rely on chance to be in the right place, and troops would find it easier to maneuver around them.

“Without an autonomous aerial launch facility, MAADS is a much less useful system. It doesn't provide much of value over the same number of individual tactical robots.”

Charles paused to let them consider the implications of the statement before continuing. “The fourth sub-project is to provide a deployment capability. This combines all the MAADS elements into a single package that can be transported to wherever it's needed. The package must be capable of autonomously deploying the aerial vehicles and the tactical robots. The package could be transported to its deployment location in several ways, including lifted in by helicopter, paradropped by a GPS-guided parachute, accurate enough to hit a landing site within a few meters, or even driven to wherever it is needed on a flatbed truck.

“This sub-project includes cabling and fitting everything into the control unit's container such that it's in the best position for deployment, not glamorous, but necessary work.”

Arjay asked, “Does MAADS need to self-deploy?”

“What do you mean by self-deploy?”

“With all the MAADS components packaged into a container, we could put it on a vehicle chassis, and get it to drive itself to wherever it needs to go.”

“An interesting idea, but not currently a requirement. I think we should file it under potential new capabilities, and bring it out the next time DARPA asks us what else we could do with MAADS.”

Charles knew one of the biggest problems in managing smart technical people was getting them to stick to the documented requirements. Left to their own devices, they would spend most of their time on things they thought were cool or interesting. At the same time, they often had fragile egos, and it was necessary to tell them their idea was good, and make them think it would be included, just not quite yet.

“The final sub-project is to develop a weapon and camera assembly we can install onto the tactical robots and the control unit. The weapon and the cameras both need to be on rotating platforms, and capable of changing their vertical elevation. This presents very similar engineering problems, and by combining them into a single assembly, I anticipate we need only solve the problems once.”

Charles waited to see if there were any questions, but everyone was waiting for him to tell them who got to do what. “I am going to put one team member in charge of each of these sub-projects. That person will own the sub-project, and be responsible to me for on-schedule delivery.

“I am putting Richard in charge of the command and control capability, because it's primarily a software project.”

Charles looked across at Jennie, who as usual had said nothing in the meeting. “Of course, while Richard is nominally in charge, we all know who is really in charge. Don’t we, Jen?

“Arjay will be responsible for the refueling facility. I put him in charge of this sub-project because of his experience with industrial robots, and because of his knowledge of the tactical robots. I've included refueling the aerial vehicle as part of this sub-project, although it's closely related to the relaunch capability.

“Fiona will be responsible for the aerial vehicle launch mechanism because it is heavily dependent on the vehicle's own capabilities. How we launch, how we refuel, and how many aerial vehicles we need, are linked in a complex dependency, but I'm confident, if anyone can resolve the dependencies and reach a solution, Fiona can.

“Sammy will be responsible for the deployment capability, which means he is responsible for designing the layout of the control unit because the control unit is our deployment container. It's primarily an engineering design problem, and Sammy is our senior, as well as our best, engineer.”

Charles thought the deployment capability and design of the control unit was the most thankless of the sub-projects. He was glad he had a good conscientious, detail-oriented person like Sammy to do it.

“The weapon and camera assembly requires we bring together weapons expertise, software, mechanical, and electrical engineering aspects. For that reason, I will run the sub-project myself, although I will need a lot of help from Sergeant Jackson, Richard and Sammy.”

Charles couldn't say the real reason he was taking charge was to ensure Jackson and Richard worked effectively together. He had seen worrying signs of friction between them.
He had thought about giving Sergeant Jackson control of the sub-project, but decided against it.

Charles didn't know if Jackson was capable of running a project. It didn't sound like the kind of skill a combat soldier would acquire. Even if Jackson knew how to run a project, he wasn't confident the soldier could manage the relationships with the people he needed to work with, to achieve his objectives. Charles had seen projects fail because people let egos and personalities get in the way of delivering.

Sergeant Jackson wondered whether he should feel offended that Charles trusted Richard, Sammy, Arjay and Fiona to run their own sub-project, but not him. He resolved to speak to Charles immediately after the meeting, and ask him face-to-face why he didn't get the chance.
Charles displayed a chart comprising a series of horizontal bars. “This is a Gantt chart showing our schedule over the next ten months. Each horizontal bar represents one of the activities I have described this morning.

“But first, I need to talk about what the schedule means. This is a rapid development project. The schedule is fixed. The project will be terminated if we cannot deliver within the planned dates.

“Ladies and gentlemen, this is a one shot deal.”

Charles paused to let the statement sink in. He had already told them individually the same thing, but he would use the opportunity to reinforce the message. He had also told some of them, if the project was terminated, DARPA could decide to reconstitute it, but only some of the people on the current project would be invited to rejoin. It was the sporting world's 'team cut' applied to high-tech development projects.

“Our first key milestone is a working prototype of the command and control software at this point here, in just over three months from now.” Charles pointed to the Gantt chart. “Five weeks later, we should take delivery of our first control unit. Two weeks after that, we complete autonomous refueling. At the same time we finish the weapon and camera assembly. Two and a half weeks later, we demonstrate a prototype of our autonomous launch capability. That takes us through to early December. By then, we should have a clear idea of whether or not MAADS will work within our project schedule.

“I realize this is an aggressive schedule, but we need to know as soon as possible if there are any problems, we can't in principle solve, and I'm confident the working prototypes will tell us.

“The thing that concerns me most about this schedule, is that the highest risk of the five sub-projects, launching the aerial vehicle, we complete last. Good risk mitigation practices require you to do the highest risk things first, but, as is often the case, the riskiest activities require the most investigation and design upfront, so we can't complete the prototype any sooner. However, it means Fiona's sub-project gets priority.”

Arjay asked, “What's our drop dead date, Charles?”

Charles didn't like to encourage people to think in terms of drop dead dates. If the schedule said you delivered on a certain date then that was the date, not some later, unstated date. At the same time, everyone knew there was some slack in the schedule. They could tolerate a degree of slippage, and still meet the final delivery date.

“I have a review meeting with DARPA immediately after the holidays. If I can't go into that meeting saying our big five development activities are all at the working prototype stage, and we are into the integration phase, then I will be telling DARPA we can't meet our original schedule, and we should be considering whether and how to proceed with MAADS.

“If we leave for the holidays, and any of the five sub-projects are unfinished, then there is a good chance we will not be coming back.”

“Can you define exactly what you mean by a prototype?”

“Sorry Fiona, I should have done that already. 'Prototype' is one of those words that people bandy around, and use to mean, often subtly different things. What I mean by a prototype is something that has all the important requirements in some form, and there is no deficiency, we do not have a clear idea of how we will solve.

“As an example, we need a way to launch the aerial vehicles, then recover, refuel, and relaunch them. The working prototype needs to demonstrate these capabilities, not necessarily integrated into the control unit, and not necessarily integrated together. The prototype doesn't need to be reliable or robust. Just as long as we think, we know how we will integrate, and how we will make it sufficiently reliable and robust.”

Fiona said, “Thanks Charles, I think that answers my question. But what happens if we get to one of the prototypes, and an important requirement is missing or doesn't work? Will DARPA terminate the project?”

“First, we decide if we can fix the deficiency, and if so how long it will take, and if we need additional resources. We then go back to DARPA, tell them about the deficiency, and present our recommendations on how we propose to continue. If we can find a way to fix the deficiency within the schedule, then that's what we propose. If we need more time and resources than the current schedule allows, then we propose terminating this project, and present a proposal for a new project with a new schedule.

“Either our sponsors agree to our proposal, or they drop the requirement, or they amend the requirement, so that we can deliver it within the schedule. It depends on what we have a problem delivering. Let's say, it is the ability to recharge the tactical robots, then we might change the requirement to something we can do, like fit them with a solar panel.”

Charles knew the project was high risk. It was meant to be. The biggest advantage of a short duration project was that you found out quickly it had failed and why. You could then go back to the drawing board, and start again with a much better chance of success, in the time a longer project would take to just fail.

He needed to encourage people to take risks and court failure, without projecting the perception, failure was the likeliest outcome.

“I don't want you to dwell on the possibility of cancellation. High-tech development projects get canceled for many reasons, most of them beyond the development team's control. Let's focus on solving the problems we have to solve, solving them one at time, until we don't have any more problems left to solve.

“That’s all for today.”

Charles noticed Jackson was not getting ready to leave the meeting like the others. When they were the last two left in the room, Charles looked across at him, inviting him to initiate the conversation.

“I want to know why I don't get to run the weapon and camera assembly sub-project? Is it because you think I am not capable or is there another reason?”

Charles sat back down, and pressed his fingertips together that signaled, for those who knew him, he was about to tell you something you probably didn't want to hear.

“When I was a child, my mother used to tell me 'everything is easy when you know how'. It wasn't until years later that I really understood what she meant. I don't know whether you are capable of running your own project, but you are an intelligent and resourceful man, and I must assume you are. The issue is not whether you are capable. The issue is, 'do you know how'.
Richard, Fiona, Arjay, and Sammy have all run projects before, and more importantly, they have seen others do it. Consequently, they have a good idea of how to run a project, even though they may not be as good at running an actual project as you might be.

“I don't have to explain schedules, resources, milestones, critical paths, and drop dead dates to them. They know about these things. Whereas, I don't know if you do, and I can't afford to find out, or more correctly, I can't afford to find out you don't know. That's the reason why I am running the weapon and camera sub-project, and not you.

“In case you think you are missing something, running a project is a mundane activity. Doing the project work is much more interesting. Perhaps, one day you will get to run your own project, and you will realize, this time I am doing you a favor by running it for you.”

Jackson wasn't sure he really understood Charles' point, but he did recognize the man was genuine in what he said.

“Thanks for the explanation.”

Jackson had an appointment the next day with the SOCOM Assignments Manager who'd arranged his secondment to the MAADS project. He had decided to request a return to active duty as soon as possible. His knee was healing nicely, and the doctor said he was almost ready to start heavy training.

Chapter 12