The DARPA Challenge
“How is the conceptual design for the unmanned aerial vehicle coming along?”
“We’ve concluded that a medium-altitude UAV with a maximum takeoff weight of 5,700kg, capable of 24-hour operations, is the most suitable. We can handle the airframe design in-house, but the issue lies in the weaponry to be mounted on a medium-altitude platform.”
President Jo Yoon-je decided, due to technical limitations, to first develop a turboprop-powered, propeller-driven UAV similar to the MQ-1 Predator, with plans to develop a high-altitude UAV in the future.
“Is it a structural issue related to weight?”
“Yes. Because we had to extend the wingspan to increase operational time, there’s no room for anything other than lightweight Hellfire missiles. And there’s a possibility that the U.S. won’t grant an export license for those.”
“Phew! The lack of foundational technology for the aerospace industry in South Korea is a real problem. We’ve only just started to catch up to the U.S. in aviation systems, but we always hit a wall with missiles and engines.”
“Representative, the fact that we’ve developed the T-50 after the F-5 Jeogongho and F-16 assembly production is a significant achievement in itself. If we invest a little more budget and time, I believe we can develop our own aircraft.”
“I will instruct President Park of Starlink Defense to develop a Hellfire-class missile and an air-to-ground guided bomb that can be mounted on the TA-50 tactical lead-in fighter trainer. Also, are you sending the pilot data to Deepmind?”
“Yes, we are sending the accumulated monthly data to President Cha Young-jin.”
After inspecting the TA-50 aircraft equipped with the AESA Radar and giving the development orders to President Park In-tak of Starlink Defense, Taesik headed to Seoul. Upon entering his office, he found President Cha Young-jin of Deepmind and Researcher Lee Sang-hoon waiting for him. Seeing the two men, who rarely made appearances, waiting for him, Taesik wore a puzzled expression.
“What brings the two of you from Deepmind to see me? Is there a problem?”
At Taesik’s question, his junior, Lee Sang-hoon, smiled and replied.
“Representative, the autonomous vehicle is finally complete. We plan to enter the DARPA Challenge, and because of its symbolic importance, we wanted to get your approval before proceeding.”
“So you’ve finally completed the autonomous driving technology. Where is this year’s DARPA Challenge being held?”
“They say it’s being held at a military airfield in California. Unlike last year, it’s not just about finishing the course; we have to follow traffic rules, avoid obstacles and other vehicles, and complete a 100km distance within 6 hours.”
“Is there prize money for winning?”
“Unlike last year, there is prize money this time. With Google and automotive companies sponsoring it, the prize is $2 million.”
“If we win the prize money, we can use it as a bonus for the researchers.”
Researcher Lee Sang-hoon looked burdened by Taesik’s words.
“Representative, this is our first time participating—are you putting that much pressure on us already?”
Taesik signed the application form handed to him by Researcher Lee, handed back the approval board, and asked President Cha Young-jin.
“What’s the status of the pilot data received from Starlink Aviation?”
“We’re using software similar to the autonomous vehicle’s to train the AI, but there are a few problems. We have enough data for simple takeoffs and landings, but we’re limited by the lack of data on weapon deployment.”
“That’s data that even Starlink Aviation can’t collect. It looks like we’ll need the Ministry of National Defense’s help after all.”
“The autonomous flight program for aviation was actually simpler and easier to develop than the AI for cars. It’s simpler because we only need to reduce the probability of collision using the aircraft’s radar information. However, the issue is performing formation flight and tactical missions; it seems we’ll need to use EO/IR (Electro-Optical/Infrared) cameras instead of radar.”
“For fighter formation flight, we’ll need to utilize IRST (Infrared Search and Track) equipment.”
“Yes. Fortunately, if we utilize the target identification and tracking algorithm you provided, Representative, limited formation flight should be possible.”
“It all comes down to securing the training data. I’ll see if I can get help from the Ministry of National Defense.”
Once the conversation with President Cha ended, Researcher Lee Sang-hoon asked as a junior.
“Senior, since this is our first time participating in the competition, you’ll come with us, right? I’m too nervous to go alone.”
Taesik decided to accompany them, as he needed to deliver the mobile processors produced by Misung Electronics to Aple. Despite Taesik’s repeated warnings, Misung Electronics had decided to ignore the smartphones Aple was developing and focus only on the mobile phones they were currently producing, which Taesik found regrettable. Ultimately, just like in his previous life, Misung Electronics would launch their smartphone later than Aple and lose many opportunities.
Two months after the North Korean ICBM removal operation ended, Section Chief Choi Jae-ho appeared at Starlink with a smile.
“Representative, thank you for your help in the last operation.”
“It’s been ages since the operation ended… More importantly, thank you for ensuring our employees received their performance bonuses.”
“You helped for the sake of the nation, so it was only natural. More importantly, I’ve recently secured the technical data you mentioned.”
“…….”
“I’ve finally secured the solid-fuel production technology. It wasn’t on the NIS server, but I barely managed to find it in hard-copy form at the Defense Research Institute. They didn’t want to give it up, so I forcibly took it.”
Taesik received the USB from Section Chief Choi.
“Is this technology provided by Russia?”
“Yes. However, even though I’m giving you this data, it will be difficult for Starlink to develop a solid-fuel rocket. As you know, due to the ROK-U.S. Missile Guidelines, there are restrictions on range and warhead weight, which makes utilization problematic.”
Taesik knew that launching a satellite rocket would be difficult, but he thanked Section Chief Choi because the technology could be utilized for similar missiles like the Hellfire needed by Starlink Aviation.
“Thank you, Section Chief Choi. Developing a solid-fuel rocket might be difficult, but it seems it can be utilized for developing aviation missiles.”
“Hahaha. I also reported to the command that it was for missile development, not solid-fuel rocket development.”
“There hasn’t been any follow-up information on the operation conducted last time—how is North Korea reacting currently?”
“They still think the U.S. removed their ICBM through a secret operation. Instead, the U.S. has requested that we notify them in advance if there are any similar operations in the future. It’s convenient that the U.S. is taking the heat for North Korea’s anger instead of us.”
“Please ensure that information regarding Starlink’s involvement is controlled. It wouldn’t be good for us to get involved in unnecessary disputes.”
“The U.S. probably already knows that Starlink’s satellites were used. Instead, I will be careful so that no related information leaks to the North Korean side.”
Taesik sent the technical data received from Section Chief Choi to Starlink Defense via courier. After confirming the solid-fuel production technology, President Park In-tak was astonished and called him immediately. Taesik reassured President Park by saying it was technical data obtained from the government, then gave instructions to develop a ground-based guided weapon similar to the Hellfire before heading to the United States.
Upon arriving in California with Researcher Lee Sang-hoon’s team, Taesik was stunned by the military airfield where the DARPA Challenge was being held. For the competition, the U.S. military had installed various obstacles and traffic lights within the airfield to create a realistic environment, and in particular, they had placed military vehicles in several locations to simulate real road conditions.
While Researcher Lee Sang-hoon inspected the vehicle they had sent in advance, Taesik toured the cars prepared by each representative and began comparing them with the ones prepared by Starlink Deepmind.
The cars released by universities and research institutes were equipped with LiDAR and cameras on top, and the screens they were operating displayed the information obtained through LiDAR and cameras in 3D. After touring the competition venue with President Cha Young-jin, Taesik found Researcher Lee Sang-hoon and delivered the information on the cars produced by each country to confirm Starlink Deepmind’s state of preparation.
“It seems they aren’t using electronic maps this time—how did you decide to identify the roads?”
“We decided to obtain road and object information using radar and LiDAR. The cameras are set to acquire lane information, and the radar and LiDAR are for object information. It would be helpful if we could use electronic maps, but the organizers told us not to.”
“So, you’re getting lane information from the cameras and performing autonomous driving with the information coming in through radar and LiDAR.”
“That’s right. And since radar and LiDAR only provide current information, we have to calculate predictive information through neural networks. If it hadn’t been for the AI neural network program you gave us, Representative, predicting vehicle information would have been difficult. Ultimately, the hardware is all at a similar level; the victory will be decided by the autonomous driving program.”
Starlink Deepmind received the 50th starting position out of 98 participating teams. When the competition began, the steering committee decided to conduct the preliminary round first and advance only 10 teams to the finals. Compared to the finals, the preliminary course was simple, involving a round trip on a road made using the airport runway. Cars that couldn’t drive on the road arranged without obstacles were eliminated, and the final 10 teams were selected.
The autonomous vehicle showcased by Starlink finished the road driving without any issues, and when their advancement to the finals was decided, President Cha and Researcher Lee Sang-hoon rejoiced as if they had already won.
By the afternoon, the 10 teams that passed the preliminaries started sequentially, and Starlink started in the final position. Countless people watched with interest as the Stanford University vehicle, which started first, departed. When the first obstacle appeared, the crowd tensed up, watching the vehicle’s movements.
The first obstacle was an intersection where they had to stop and then proceed; the judges watched to see if they complied with the "STOP" road markings on the ground. The Stanford vehicle stopped without issue and approached the second obstacle. The attendees clapped and cheered as the vehicle passed the first obstacle, and they observed the movements of the remaining vehicles.
The remaining vehicles stopped simultaneously when the lead vehicle stopped for the stop sign, and they began to move together when the front vehicle moved. Since they were autonomous vehicles at the level of participating in the preliminaries, everyone passed the stop sign safely, but a bottleneck occurred at the second obstacle, the traffic light.
The Stanford University vehicle crossed the intersection slowly in accordance with the green light, and the remaining vehicles experienced a bottleneck due to the vehicle in front and the traffic light. A moment later, when the light turned green, the vehicles began to drive slowly, and the vehicles that entered the straight road began to increase their speed little by little.
When the Starlink car, which started in the last position, stopped at the second obstacle, Taesik looked at Researcher Lee Sang-hoon’s face.
“It seems it stopped driving because other vehicles are blocking both sides of the two-lane road. The signal is normal, so let’s wait a moment.”
As the congested vehicles cleared the second obstacle, the Starlink car gradually began to pick up speed. The attendees’ eyes were focused on the Stanford and Carnegie Mellon vehicles leading the pack, and at the 30km mark, an absurd scene was witnessed.
As a flock of birds passed in front of the autonomous vehicles, the Stanford and Carnegie Mellon cars stopped, and they only began to move again once the entire flock had passed.