giving spacecraft the ability to see where they were going, understand the terrain beneath them, and adjust in real time
In the high desert of West Texas, a rocket carrying no passengers prepared to carry something more consequential than tourists: a set of sensors and algorithms designed to teach spacecraft how to land themselves on worlds where no human hand can reach the controls. Blue Origin's New Shepard, funded in part by a $3 million NASA contract, lifted SPLICE — a precision landing system of cameras, lidar, and autonomous software — to the edge of space and back, in a 12-minute test that asked a quiet but enormous question: can we trust machines to find safe ground on the moon and Mars? The answer, if it comes, would open lunar and Martian terrain that human ambition has long imagined but never safely reached.
- The moon and Mars are littered with boulder fields and shadowed craters that have kept spacecraft from landing in their most scientifically valuable regions — SPLICE exists to change that calculus.
- Two previous launch attempts in late September were scrubbed over a power-supply fault, a reminder that even a 12-minute test flight carries the full weight of spaceflight's unforgiving complexity.
- The system works by fusing terrain-mapping cameras with Doppler lidar lasers that measure altitude and velocity, all processed by a new onboard computer that runs the entire descent without waiting for commands from Earth.
- NASA's Tipping Point program distributed $44 million across six private companies to accelerate exactly this kind of technology, betting that commercial partnerships can compress the timeline to operational readiness.
- A successful October test would feed real flight data into systems already scheduled to fly on commercial lunar landers in 2021 — including the mission carrying NASA's first moon rover since Apollo.
Blue Origin's New Shepard rocket lifted off from West Texas not to carry tourists, but to test one of space exploration's most stubborn problems: how to land a spacecraft safely on the moon or Mars when no human pilot can intervene.
The payload was SPLICE — Safe and Precise Landing–Integrated Capabilities Evolution — a NASA-designed suite of sensors, software, and a specialized computer built to guide spacecraft autonomously to the surface of other worlds. NASA paid Blue Origin $3 million through its Tipping Point program to test these systems in actual flight, sending them 62 miles up and back down in a 12-minute suborbital arc. The test had already been delayed twice in late September due to a power-supply issue, underscoring the complexity lurking beneath even a brief flight.
The problem SPLICE addresses is deceptively simple to name: the moon and Mars are dangerous, uneven, and full of terrain that previous landing systems could not safely approach. SPLICE answers with two sensor systems working in concert — cameras that map the landscape in real time and compare it to pre-loaded terrain data, and a Doppler lidar that fires lasers at the ground to calculate altitude and descent speed. A new onboard computer ties it all together, running the descent without waiting for instructions from Earth.
What gave the system particular promise was its modularity. NASA designed SPLICE so its components could operate together or independently, letting mission planners choose only what each landing requires. The Doppler lidar alone was already scheduled to fly on two commercial lunar missions in 2021, including the lander carrying NASA's first moon rover since the Apollo era ended in 1972.
For NASA's broader ambitions — a permanent lunar base, and eventually human boots on Mars — the New Shepard's 12-minute journey represented a crucial waypoint. A successful test would validate years of engineering and push autonomous landing technology closer to the missions that will define the next chapter of human exploration.
Blue Origin was preparing to send Jeff Bezos's New Shepard rocket on a mission that had nothing to do with tourists or joy rides. Instead, the company had agreed to carry NASA's latest attempt at solving one of space exploration's hardest problems: how to land safely on the moon or Mars without a human pilot at the controls.
The payload was called SPLICE—Safe and Precise Landing–Integrated Capabilities Evolution—a collection of sensors, software, and a specialized computer that NASA had designed to guide spacecraft down to the surface of other worlds with precision and autonomy. The agency had paid Blue Origin $3 million to test these systems in actual flight, lifting them 62 miles above Earth to the edge of space before bringing the rocket back down. The entire test was scheduled to take 12 minutes. If everything worked, the New Shepard would touch down safely in West Texas on a Tuesday morning in October 2020, and NASA would have real data about whether its new landing technology could survive the rigors of spaceflight.
The challenge SPLICE was designed to solve was deceptively simple to state and enormously difficult to execute. The moon's surface is treacherous—boulder fields, shadowy craters, uneven terrain that previous landing systems had deemed too risky to approach. Mars presented similar hazards. NASA wanted to change that equation by giving spacecraft the ability to see where they were going, understand the terrain beneath them, and adjust their descent in real time, all without waiting for commands from Earth. Two sensor systems would do the heavy lifting. One used cameras to map the landscape in real time, comparing what the spacecraft saw to pre-loaded maps to determine its exact position. The other, a Doppler lidar, would fire lasers at the ground and measure the returning reflections to calculate altitude and landing speed. A new computer would process all this data and coordinate the different systems, running the entire descent autonomously.
John Carson, who worked on precision landing technology at NASA's Johnson Space Center, explained the significance in a September news release: testing SPLICE on an actual suborbital rocket represented a major step beyond what the agency had accomplished in laboratories and helicopter field tests. This was the first of two flight tests that Blue Origin would conduct through NASA's Tipping Point program, which had distributed $44 million across six private companies to help push next-generation space technologies toward operational readiness. Blue Origin had also received an additional $10 million to test a propulsion system using ultra-cold liquid fuels for lunar landings.
The path to this October test had not been smooth. Blue Origin had originally scheduled the launch for September 24, but scrubbed it due to a power-supply issue. A second attempt the following day was also canceled while the company worked to verify a fix. The delays were routine in spaceflight, but they underscored the complexity of what was being attempted.
What made SPLICE particularly elegant was its flexibility. NASA had designed the system so that individual components could work together or independently, allowing mission planners to select only the elements needed for a particular landing. The Doppler lidar, for instance, was already slated to fly on two commercial lunar robotic missions in 2021, including the lander that would carry NASA's new rover to the moon—the first such landing since Apollo ended in 1972. A third sensor system, designed to scan the surface and identify hazards before selecting the safest landing spot, would be tested in future flights.
The implications extended far beyond a single test flight. If SPLICE worked as designed, it would unlock vast regions of the lunar surface that had previously been considered too dangerous for landing. On Mars, it could enable human missions to touch down in terrain that current technology could not safely approach. For NASA's ambitions—establishing a permanent base on the moon and eventually landing astronauts on Mars—this test represented a crucial waypoint. The New Shepard's 12-minute journey would either validate years of engineering work or reveal problems that needed solving before the next phase of human space exploration could begin.
Citações Notáveis
Testing SPLICE technologies on a suborbital rocket expands the envelope beyond previous lab tests, helicopter field tests, and lower-altitude suborbital rocket tests.— John Carson, NASA Johnson Space Center precision landing technology specialist
A Conversa do Hearth Outra perspectiva sobre a história
Why does landing on the moon need to be so automated? Can't mission control just guide the spacecraft down?
The signal delay. Commands from Earth take seconds to reach the moon. By the time a pilot sees a problem and sends a correction, the spacecraft has already moved. Autonomous systems have to see, think, and act in real time.
So SPLICE is basically giving the spacecraft eyes and a brain.
Exactly. Two different kinds of eyes—one that recognizes landmarks, one that measures distance using lasers. And a computer that processes both feeds simultaneously to figure out where to land safely.
What happens if the sensors disagree with each other?
That's what the software is designed to handle. The systems are built to work together, cross-checking each other. If one sensor gives bad data, the others can compensate. That's why testing them together in actual flight matters so much.
Why test on a suborbital rocket instead of just sending it to the moon?
Cost, mainly. And risk. You get the spacecraft into space, expose it to the real environment, see if it survives and functions—then bring it back safely to study what happened. It's a proof of concept before you commit to an actual lunar mission.
What happens if the test fails?
Blue Origin and NASA learn what broke and why. That's still valuable. But the real win is if it works—then those technologies move from theory to proven hardware that can actually fly to the moon.