Rebirth: Starting from Lighting Up the Tech Tree

Chapter 172 Rocket Chapter

Two weeks later, Li Guodong led three members of his technical research team into the conference room.

Zuo Cheng was already inside. Yu Ying sat on the left, and Shen Yiming sat opposite them, their laptops on.

Li Guodong placed the documents on the table, the top one bearing the words: Comparative Analysis of Rocket Recovery Technology Solutions.

"Get straight to the conclusion," Zuo Cheng said.

"Two main approaches," Li Guodong said, walking to the whiteboard and picking up a black marker. "The first is vertical recovery. After the first stage separates, the rocket turns around, uses its main engines to decelerate, adjusts its attitude, and finally lands vertically on a sea platform or a land landing site. SpaceX's Falcon 9 follows this route."

He drew an arc on the whiteboard from the top left to the bottom right, and then drew a simplified rocket at the bottom.

"The second method is parachute recovery. After primary separation, the parachute deploys to slow the descent, eventually landing in the sea for retrieval by ships. This method has lower technical barriers, but post-recovery maintenance costs are high, and the weight of the parachute also occupies payload capacity."

Yu Ying twirled a pen: "What are the advantages and disadvantages?"

"Vertical recovery has obvious advantages. The post-recovery maintenance cycle is short; SpaceX's experience shows that the first stage of a rocket can be reused more than ten times. Landing accuracy is high, allowing it to return to the vicinity of the launch site." Li Guodong paused, "The disadvantage is the extreme technical difficulty. Attitude control, engine thrust adjustment, landing cushioning—every aspect is a world-class challenge. SpaceX spent ten years, destroying more than a dozen rockets, to achieve its current recovery success rate of over 90%."

"What about parachuting?"

"Parachute landings don't require complex attitude control systems, and the engine doesn't need to be ignited multiple times," said Li Guodong. "But parachutes have weight, which reduces carrying capacity. After landing in the sea, seawater corrosion and immersion cause significant repair costs. Salvage is uncertain; large waves might prevent it from being found."

Zuo Cheng placed the documents on the table.

What is our goal?

"Lowest cost, highest reuse rate, and maximum carrying capacity," said Li Guodong. "Working backwards from this goal, vertical recovery is the only viable option. Parachute landing has a lower barrier to entry, but its long-term economic viability is not as good as vertical recovery."

Zuo Cheng nodded.

"I agree. Vertical recycling."

No one in the meeting room objected. This conclusion had been largely formed during the past two weeks of research, and today's meeting was simply to formally finalize it.

"Now let's get into the specifics," Li Guodong said, turning to the next page of the document. "Vertical recovery presents three core technical challenges. First, attitude control. The rocket's reentry into the atmosphere is at extremely high speeds, with complex aerodynamic forces. Precisely controlling the attitude to ensure the engine nozzles are aligned correctly is incredibly difficult. Second, engine thrust adjustment. Multiple ignitions are required during recovery, and the thrust magnitude and duration must be precisely controlled; an error exceeding five percent can lead to landing failure. Third, landing cushioning. The impact force during vertical landing is enormous, and the landing legs must absorb the impact while ensuring the rocket body remains undamaged."

Shen Yiming began, "Our AI team has an advantage in attitude control. The real-time inference latency of neuromorphic chips is less than one millisecond, making them an order of magnitude faster than traditional PID controllers for rocket attitude feedback control."

"Theoretically, yes," Li Guodong said. "But the actual flight environment of a rocket is far harsher than the simulation model. High temperature, high pressure, and strong vibration—whether the wafers can work stably under such conditions requires specialized testing and verification."

"It can be done," Shen Yiming said. "Let's start with hardware hardening, and then do environmental simulation tests."

Zuo Cheng stared at the diagram on the whiteboard, the "Satellite Attitude Control" blade on the system panel flashing through his mind. When that blade was activated, he obtained an attitude control algorithm framework, the underlying logic of which was a generation ahead of current mainstream theories. But he couldn't say that.

"Let's start with simulation," Zuo Cheng said. "We'll use AI to build a full-process simulation model of rocket recovery, with three modules working in parallel. I want to see the first simulation result within a month."

Li Guodong nodded and wrote it down.

"There's another issue," Yu Ying said. "Regarding the overall design of the rocket, should we develop it independently or find partners? Designing a rocket from scratch takes at least three years and requires numerous test launches, which is extremely costly."

"We're taking a two-pronged approach," Zuo Cheng said. "On the one hand, we're negotiating cooperation with the China Aerospace Science and Technology Corporation to modify their existing rocket platforms and shorten the development cycle. On the other hand, our own R&D team is simultaneously launching preliminary research to accumulate independent design capabilities. We rely on cooperation in the short term, and on independent research in the long term."

Han Lu added, "I'll handle the liaison with the China Aerospace Science and Technology Corporation. They have a prior cooperative relationship with us on space photovoltaics, so discussing rocket modification should be easier than starting from scratch."

"Okay," Zuo Cheng said. "We'll proceed along three lines. Li Guodong will be in charge of the technical line, completing simulation verification within three months. Han Lu will be in charge of the cooperation line, obtaining feedback from the China Aerospace Science and Technology Corporation within one month. Yu Ying will coordinate the pre-research line, drawing key personnel to build an independent design team."

After the meeting, Zuo Cheng kept Li Guodong behind.

"I have some ideas about the attitude control module," Zuo Cheng said. "I'll organize my thoughts later and give you a technical memo."

Li Guodong nodded: "Okay, Mr. Zuo."

Zuo Cheng didn't explain. He couldn't explain the origin of that algorithmic framework, but he could use the name of "technical intuition" to gradually release the knowledge.

That night, Zuo Cheng stayed in his office until very late.

He left only a table lamp on the desk. Outside the window, cars were moving on the distant overpass.

He opened the system panel: 517 points, six branches. The passive effect of a 1.3x technical boost was active, making his thinking clearer and his approach more direct than usual.

The detailed information about the satellite attitude control blade was still in his memory. The core of that algorithm framework was an adaptive control strategy based on deep reinforcement learning, capable of adjusting control parameters in real time under unknown disturbances. The original version of this strategy was designed for deep space probe landings, and its application in rocket recovery follows the same principle.

He sketched several diagrams in his notebook, breaking down the key steps of the algorithm into modules that the engineering team could understand.

After finishing writing, he closed the notebook. Rocket recovery—these four words represent one of the most difficult hurdles in the global aerospace industry. SpaceX took ten years to overcome it, and no Chinese company has yet truly achieved it.

However, if the algorithm within that blade can be successfully transferred to engineering implementation, the development cycle can be shortened by more than half. This is the certainty provided by the system.

He put the notebook in the drawer, turned off the desk lamp, and walked out of the office. The corridor was quiet, with only the emergency exit indicator light flashing green. In the distance, Yu Ying's office light was still on. She peeked through the crack in the door and saw Zuo Cheng walk by, saying, "Still not gone?"

"I just finished writing something," Zuo Cheng said.

"Several sets of thermal vacuum test data from the second-generation satellite are incorrect; I'm checking them."

"Get some rest."

"Know."

Zuo Cheng walked towards the elevator. Attitude control, thrust adjustment, and landing cushioning—he overcame each of these three challenges one by one.