Lightship

by Stephan Besik

  “Next time come to me as soon as you know there is a problem.” Klaus turned back to look at Kevin again and Kevin got that head-drilling feeling. “The sooner I know we have a contingency the sooner I can review budgets to see if we can handle the problem internally. Mr. Smith is spending a great deal of money on this project. He will do what is necessary to keep it moving, but I would rather not impose on his largesse. We should try to act responsibly when something comes up. Next time, let’s try to get the design right the first time.” With that Klaus turned and walked away.

  Kevin watched Klaus go. He felt as if he was a seventh grader getting called on the carpet for not doing his homework. ‘Not impose on his largesse?’ Really? Did he think they were out buying sports cars for the engineers? They’d done their best; sometimes that was how engineering worked. Get the design right the first time? The best you ever got was pretty good the first time. It wasn’t unusual for a first try to be dead on arrival. That was just the way things went.

  Kevin was angry, but he did his best to get past his feelings. Klaus might be an asshole, but he was an asshole who got things done. He knew that Klaus would go scrounge money in the budget and then pick the robotics guys’ pockets to get Artie his refab. In the process Artie would have a chance to watch a real money guy handle a vendor. At the same time Kevin would be able to continue working with the engineers instead of going to meetings in which vendors treated him like a prime sucker.

  Kevin pulled out his cell and called Artie.

  The Fix is In

  “Hey Art, this is great!” Lucy yelled across the drone room at Arturo. “Smooth as silk. Set the sight and the old teapot takes off and heads straight to the target. Much better than trying to line up the drone first. She stops on a dime, too. Real good stuff! Next time let’s go for this first!”

  Lucy wasn’t looking his way, so Arturo grimaced and shook his head. Lucy had caught an important glitch but she was still a pain in the ass. Kevin must have an iron stomach to have her around. She was a good programmer and all, but the yelling was definitely hard to take. He wished she would go light on the teapot label, too. There were superficial similarities, but a teapot was definitely not a highly complex semi-robotic device. He guessed he’d better get used to it, though, at least as much as Kevin. Especially after her catch on the drones, it wasn’t likely she was going anywhere.

  Arturo looked again. Lucy had some sort of visor on her head. What the heck was that about? He walked over to her workstation.

  “Hey, Luce, what’s that on your head? Afraid you’re going to catch a bug?”

  Lucy smiled, but kept her eyes on the screens in front of her.

  “It’s a little gadget I thought up. I call it EDD. With two D’s. Stands for ‘Eyes Down Display’. There are a bunch of parameters we need access to, but they’re hidden at the bottoms of the screens and they’re hard to read. I was thinking about trying a HUD, like the Heads-Up Display that the Air Force uses, but that goes right over your field of vision and things can get obscured. The problem with our setup is that the numbers are in sort of the right place, out of the way, but they’re too far away and you kind of have to look for them on the screens. So I thought, what if there was a way to put them closer to us, but not over the view screens somehow.

  “Then I ran into my old grandma, who’s still wearing a visor she got someplace after the last flu bug. And I thought, what if we could put our data on the bottom of the visor somehow? Then you could just glance down and pick up what you needed without losing track of stuff on the big screens. We might even be able to see the data without having to look down at all.

  “So I played around a little and came up with EDD. I tried putting my cell phone at the bottom of the visor and programmed a cutout to bring the data over, but that was pretty clunky. It made the visor heavy and funny feeling. Then I found this flexible stuff that you can display data on, and Whamo! I glued a piece to the bottom of a visor that I shortened, plugged it into one of the accessory ports, and figured out how to duplicate the viewscreen data out to the port. Now I’ve got data and a view, all at once. How’s that for slick?”

  Arturo looked at the rig, and looked at Lucy. Not bad at all. Not just a cute idea, but some damn good programming, too. Maybe they could gin up some of these for the other drone drivers. Hm.

  Maybe Lucy wasn’t going to be that much of a pain in the ass after all.

  Alvin’s Debut

  Lucy, Suzette and Robbie Kay were grinning like Cheshire cats as Kevin, Arturo, and Klaus entered the room. Suzette was the team mathematician, orbital mechanics and navigation specialist, and Robbie was the human pilot and computer management specialist.

  The three men knew only that their programming team wanted a meeting. The e-note from Suzette included no additional information. The object of the meeting was still a mystery. Apparently there were some things that couldn’t be handled in the normal weekly roundups.

  Klaus, who had little time for games and mysteries, opened the conversation. “We are all here. What is the purpose of the meeting?”

  A voice from the computer interrupted. “Hello, Klaus. How are you?”

  Startled, Klaus was confused for a moment. “Who is that? Do we have someone phoning in to our meeting?”

  Suzette smiled impishly. “Well, yes, in a manner of speaking. Gentlemen, say hello to Alvin.”

  Klaus, Kevin, and Arturo all looked mystified. “Do we have a new man on the team?” asked Klaus.

  The three programmers still had on their silly grins. Lucy shook her head. “No, Alvin isn’t a man. He’s an artificial intelligence. I guess you could say he lives in the computer that’s going to go into the first ship.”

  A light came on and Kevin spoke up. “Wait a minute. Are you guys saying that you developed an artificial intelligence to pilot the ship?”

  Robbie spoke up. “To be fair, only Suzette and Lucy created Alvin. I just talked to him a lot. It was a little tough at first, but now he’s a pretty good conversationalist.”

  “Piloting the ship turned out to be the easy part of the software,” said Suzette.

  “We knew Kevin was planning for artificial intelligence to operate the ship. We started to kick ideas around at the beginning of the software development.

  “In addition to the need for a smart navigator, Alvin came about because we needed a program that could deal with maintenance and system failures with as little human intervention as possible. The software had to be heuristic, able to make decisions based on incomplete data in unstructured situations. The software then had to be able to make the appropriate fixes or control the maintenance drones to make them. We also wanted a system that would forecast the need for preventive maintenance in advance of failures to the extent possible.”

  “If this was going to be a once-in-a-lifetime flight, Alvin wouldn’t be necessary,” added Lucy. “We could wing it with human monitoring and intervention, doing navigation in a couple of computers linked in a ground-to-space network and continuing to control the maintenance drones with operators on Earth. We realized, though, that the Moon ship wasn’t going to make just one run unless it was a failure. If things worked as planned, we wanted the ship to take care of itself to the extent possible. When we go interplanetary, we’ll really need a ship that can take care of itself. Based on communication turnaround alone, we aren’t going to be able to handle any genuine interplanetary emergencies. Even at the distance to the Moon, things may get sticky. If the ship’s on the other side of the Moon and something bad happens, it had better be able to take care of itself. Thus,” said Lucy, “we have Alvin.”

  “I believe my parents have taken the correct approach to their needs.” Alvin’s contribution startled the entire team. “Indeed, Suzette is quite correct. Piloting the ship is relatively easy given its ability to change speed and maneuver efficiently. It’s really just a matter of taking frequent measurements. Once I know the relative positions of the current destination and the ship, along with the ship�
�s current velocity, I simply make adjustments as needed. As I will begin with a rough estimate of the required trajectory, the rest should be easy.

  “The more difficult part of my job will be the prediction and execution of maintenance needs as well as decision making in case of unanticipated failures. Even though maintenance and repair is likely to take a small fraction of the time I will spend on navigation, that fraction is at least as important. The problems will include a far larger number of variables than are needed to keep on track to the Moon, and there may be situations in which the information available to me will include elements of uncertainty for which I have to compensate. With the maneuverability of this ship, navigation errors will be correctable on the fly and should not be costly. Any maintenance and repair issues are likely to be extremely costly, even though they may be few and far between. The sooner and the better they are contained, the less expensive and the less dangerous they will be.

  “Then there is the issue of time lags. Even on a short trip like one to the Moon, the lag in communications is a second or three if our ship is between the Earth and the Moon. If on the far side, the lag is many minutes. A failure on the far side that can’t be quickly addressed may well be fatal to our mission. In an extreme situation, even the short lag on the Earth side could be catastrophic. Therefore, having an intelligence local to the ship is a necessity- not a luxury.”

  Klaus acted somewhat annoyed after Alvin’s comments. “Was it really necessary to develop a speech module for the ship?”

  Suzette responded to the question. “We didn’t develop a special module, strictly speaking. We bought a package that we integrated into Alvin’s intellect via his neural net. That was a lot less expensive than trying to build up Alvin’s communication system any other way. That was really the easy part. The hard part was teaching him what he needed to know about the ship and its workings. That took weeks.

  “Frankly, his speech capabilities are a very minor part of his personality. If he needs speed of communication he will use his high-speed network to display messages to the human crew back on Earth, his drones and other devices on the ship. His ability to vocalize is more about his social interactions than about the job of piloting and maintaining the ship.”

  “Do you think Alvin can pilot and maintain the Moon ship on his, er, its own?” asked Kevin.

  Robbie nodded. “We used the most recent ship simulations to train him. He’s done well with the scale models. He’s at least as good as I would be as a pilot and far superior at monitoring and maintaining the ship. Don’t get me wrong, I’d really like to pilot ships to the Moon, but Alvin can do the job without me unless there’s a real catastrophe of some sort and a lot of improvisation is required. Even then he has engineering capabilities at Arturo’s level, expertise that I don’t have. In case of a genuine emergency I think I’d rather have him as pilot than me. Humans will only be necessary in case of a complete loss of power. Under those conditions I’m not sure even a person would be much use.”

  Suzette leaned forward. “We want Alvin to train on the Moon ship so that he can be cloned into the interplanetary ships once they are available. There should probably be a human crew on board any interplanetary ship, but we’d like to be able save the humans for emergencies. On the multi-month voyages, crew and passengers should be hibernating to save stress on the humans and reduce food and fuel requirements.

  “Alvin’s our test bed. If he can handle months of cargo runs he should be able to handle a months-long run to Mars or Venus.”

  Klaus shook his head. “I’m not sure that having an AI as pilot is safe. It may not sell well to potential passengers to the Moon, either. What if it made a mistake in a real emergency? We might end up completely shut down.”

  “That’s why we want to do this now,” pressed Suzette. “We can afford mistakes during unmanned cargo runs. He can safely learn now. Once we have passengers he’s either ready or he’s not and there won’t be any opportunities for “live” exercises any longer. And we’re not going to leave him completely on his own even now. Robbie will be watching and paralleling his decisions on these cargo runs. We’ll have drone operators in the drone room to take over if there is a really thorny maintenance problem that Alvin can’t handle. We certainly don’t want to lose the ship, but if Alvin is going to fail let’s let him fail now.

  “If it turns out that he can’t handle things by himself, we’ll have to have a human crew alert and ready, which will raise our costs substantially. And it will probably be a huge bore to the crew. The last thing we need is excitement that requires the intervention of humans all the time.”

  Kevin looked at Klaus, who seemed to be the only one having trouble with a ship piloted and maintained by artificial intelligence. “They’re right, you know. An artificial intelligence can be tested now. Once we have passengers, we’d better have enough data from the cargo runs to make a decision to commit human lives to an AI pilot or sacrifice cargo capacity to a human pilot. We might want to leave the maiden voyage in Robbie’s hands, but we should test out the AI on the subsequent unmanned cargo runs.”

  Klaus looked unhappy but he nodded slowly. “Now is the time if we think this is a viable option. I don’t think Alvin should be left alone on the maiden voyage, but the AI pilot should be tested as soon as possible after that.”

  After letting the humans thrash out his future, Alvin chimed in again. “That sounded like a positive decision to me. I’d like to be on board for the Moon ship’s maiden voyage but I’ll take the next trip if that would make everyone more comfortable. I’m ready when you want me to take command.”

  The Final Design

  By the time the design was finalized, the ship wasn’t very sleek and it wasn’t very pretty. It wasn’t a beauty contest, though. Even if it was, the other man-made space objects looked almost as ugly as the new ship would be.

  While most of Kevin’s design had survived, sizes and shapes changed as the project progressed. The biggest changes had occurred in the dimensions of the cargo space, the length of the plasma engines, and the size of the gimbals. Naturally, those changes made changes in other parts of the ship necessary as well.

  The frame for cargo got wider and longer to enable the ship to carry a larger, although not necessarily heavier payload. The change had come about as the design team began to consider the requirements for a passenger or crew pod. Even though humans took up space and required room for supplies, much of a passenger pod was empty breathing- and moving-around room. That wouldn’t be true for many cargo loads, in which nearly every cubic inch of space was occupied by something. The result was that a crew pod generally had to be physically larger than a cargo pod while not weighing as much.

  The change in the cargo frame made a change in the design of the of the central spars of the solar panels necessary, as well as requiring more strength in the gimbals that made the turning of the ship possible. The wider cargo cylinder meant that more room was needed between the sails in order for the ship to turn and reverse the direction of the engines. Otherwise the ship couldn’t slow down as it reached its destination. The spars that connected the sails to the maneuvering gimbals had to be longer. That ensured that things like turnover and reasonably narrow turns were still possible.

  As the sails got farther apart, the gimbals had to be bigger and stronger to manage the extra torque. The changes in spar length and gimbals required more weight, which made everyone unhappy. The heavier the ship itself was, the less bill-paying cargo the ship could carry. That problem resulted in some complicated tradeoff analysis to figure out at what point the cargo pod would be too big to justify the additional cost of a larger, heavier ship.

  Then there was the problem of the plasma motors. The linear accelerators that pushed the fuel to plasma speeds ended up longer than Kevin had planned. Again, this required more weight. The complexity didn’t increase but the radiators ended up longer and needed more support from the conduits and the frame that made up most of the length of the acc
elerators. Since the conduits and the frame were longer, they ended up heavier too.

  The control pods ended up larger because the design was changed to enable humans to move about inside them. They weren’t places to live in, but in a pinch a human could conduct onboard diagnostics in case of a failure that the drones couldn’t handle. In extreme emergencies, a couple of humans could stay in one for a limited time, if for example a crew pod was damaged in a way that made survival there impossible.

  The bottom line was that the ship that Kevin had designed had gotten bigger and heavier even though it still looked a lot like the original design.

  Bigger it was, heavier it was, and perhaps a little less cost effective. Beautiful, however, the ship was not.

  The ship was built around two key structural components- a keel and the gimbals that enabled the ship to swing around its sails.

  The keel was a long titanium alloy cylinder that ran from the tail of the ship up past the sail ring. Once the rear section of the keel was completed, the forward section of the keel would be added. Some distance past the ring, the forward keel was split in two, forming a sort of closed U into which large cargo pods would fit. The cargo section of the frame would also be where a human passenger compartment would fit.

  In total the section that would go in front of the sail ring looked like a pill capsule with a stubby tail at each end. The rear stub would fit into the cylindrical keel just ahead of the ship’s battery back, which was connected just ahead of the sail ring. The stub at the front of the cargo frame would start back up like the rear keel, but its sole function was support of the control module that made up the nose of the lightship. Cargo pods or a passenger module would be fastened to the upper and lower sections of the keel section that made up the cargo frame.

  The lightship didn’t have a real keel, like the ones in ocean ships. It was, however, one of two primary structural supports of the lightship. Running down the length of the ship’s hull, everything would be connected to it in one way or another. In that sense, it was at least as important, or more so, than the keel of an oceangoing ship.