Researchers at the Massachusetts Institute of Technology (MIT) have developed a new ultra-efficient chip that enables tiny autonomous robots to generate detailed 3D maps of their surroundings in real time while consuming only a fraction of the power required by existing systems.

The new MIT robot navigation chip, called Gleanmer, could help small drones and robots safely navigate complex environments, from industrial heating and ventilation systems to confined inspection spaces where battery life and computing resources are limited.

According to the researchers, the chip consumes just 6 milliwatts of power—roughly the same amount needed to run a single LED—while constructing detailed 3D maps for navigation.

The findings were recently presented at the IEEE Very Large-Scale Integrated Circuits Symposium.

Designed for battery-powered robots

Autonomous robots rely on 3D maps to understand their surroundings and avoid obstacles. However, generating these maps typically requires significant computing power and memory, making the process difficult for small, battery-powered devices.

The MIT team tackled this challenge by combining a highly efficient mapping algorithm with custom-designed hardware that minimizes memory usage and energy consumption.

“This paper showcases a key example of how you can leverage co-design of the algorithm and hardware to really push energy efficiency,” Vivienne Sze, professor in MIT’s Department of Electrical Engineering and Computer Science and senior author of the study, said in a media statement.

“While there has been a lot of work looking into compact 3D maps, what stands out about this work is that it also ensures that the process to generate those maps is as efficient as possible. Our chip allows you to store very large maps in a very small space, and do it in a very energy efficient manner,” she added.

Replacing cubes with ‘Gaussian blobs’

Traditional mapping systems represent environments using millions of cube-shaped units known as voxels. These structures require substantial memory and processing power.

Instead, the MIT researchers employed a technique that represents objects using flexible ellipsoid-shaped structures known as Gaussians.

Because these Gaussian representations can adapt to the shape of real-world objects more efficiently, the system requires far less memory than conventional approaches while still preserving detailed information about obstacles and free space.

The chip uses a mapping algorithm developed by the researchers called GMMap, which can generate accurate 3D maps from depth images in a single pass, eliminating the need to repeatedly process and store large image datasets.

“At any point in time, we only need to store a few pixels in memory, which significantly reduces the memory footprint our algorithm requires,” co-lead author Peter Zhi Xuan Li said.

Improving efficiency through hardware-software co-design

As robots move through an environment, they often observe the same object from multiple viewpoints, creating overlapping representations that can increase map size.

To address this, the researchers developed a technique that merges overlapping Gaussian representations directly, without revisiting the original image data. This further reduces memory requirements and power consumption.

The chip also keeps frequently used map data in small on-chip memory units located close to the processing hardware, reducing the need to access more energy-intensive external storage.

“By having a dedicated memory that just stores the objects you’ve seen in the previous few frames, you can access the data much more efficiently,” co-lead author Zih-Sing Fu said.

Potential uses beyond robotics

The researchers tested the chip using a range of existing 3D environments and live data streams from an iPhone camera. In these experiments, Gleanmer generated detailed maps in real time while consuming only about 2.5% of the power required by the best existing map-construction chips.

The team believes the technology could be useful not only for autonomous robots and drones but also for lightweight augmented reality headsets, particularly in applications such as medical training, repair work, and industrial assembly.

“We reduce the memory consumption by making sure the algorithm is efficient. Then we accelerate the workload that is performed by that efficient algorithm, so in the end, our chip is as efficient as possible,” Li said.

Researchers now plan to further improve the technology by bringing processing components closer to sensors and exploring additional applications, including AI systems that need to analyse complex engineering schematics.

At 20:30 hours IST yesterday, NASA’s Johnson Space Center in Houston, Texas held a live event their engineers, scientists, the astronaut corps and the media attended. The space agency officially announced the crew of Artemis III, the agency’s next human space exploration mission, paving the way for humanity’s planned return to the moon in 2028, over fifty years after the Apollo program.

Half-way through the hour-long presentation, Jared Isacson, the NASA administrator, walked to the dais to announce the all-men crew of Artemis III: NASA mission commander Randy Bresnik, mission specialists Andre Douglas and Frank Rubio, and European Space Agency pilot Luca Parmitano, an Italian national. 

Three of the astronauts excluding Douglas, a US Coast Guard reserve, are both spaceflight and military veterans. Bresnik, a US marine colonel and test pilot clocking 7,000 hours, commanded the International Space Station. So did Parmitano, the first Italian commander of the station, and who survived a 2013 spacewalk when water abruptly filled his helmet and had an asteroid named after him. Rubio, a US army helicopter pilot, holds the record for the longest time spent in space. 

Screengrab from the YouTube livestream of the event at NASA Johnson Space Center, Houston, Texas. Credit: NASA

Mission timeline

The mission could take off in the second-half of 2027. Originally,  NASA planned Artemis III to be the first soft-landing lunar mission since 1972’s Apollo 17, with a slated launch date in 2028. However, in March, the agency updated mission timelines, with the mission relegated for testing its mission critical docking mechanism, ahead of Artemis IV’s planned soft-landing that year.

The crew will fly aboard a Space X Orion capsule into low-earth orbit. Unlike its predecessor, Artemis III won’t leave earth orbit and conduct a flyby past the moon. Instead, it will test life support systems and docking with Artemis’ era lunar landers, built by private space companies Space X and Blue Origin, the Starship Human Landing System (HLS) and the Blue Moon respectively. In addition, Artemis III will carry on science experiments, including using instrumentation to test effects of atmospheric drag upon the spacecraft, amidst hostile space weather.

The Apollo and Artemis-era lunar landers drawn to scale. Credit: NASA

Lunar landers 

There has been skepticism whether the Blue Moon lunar lander’s launch schedule would be affected, in the aftermath of last week’s mishap involving New Glenn, the flagship rocket of Jeff Bezos-owned Blue Origin, exploding during a hot-static test ahead of its slated launch of Amazon’s satellites. The explosion destroyed the company’s custom-developed launchpad at Cape Canaveral Space Force Station in Florida. However, the company CEO, David Limp, posted on X, they’ll return to full-swing operations latest before the end of this year.

Whereas Starship HLS, the other lunar lander design, will feature a variant of the Starship rocket, with the latter design being still tested over repeated space flights in the past year. 

Either lunar landers designed to ferry astronauts from lunar orbit to the surface, and back. In a future Artemis mission, the astronauts, who will ride aboard Space X’s Orion crew module from earth, will dock with the lander in lunar orbit, before transferring to the lander module. 

It’s unclear which lander design’s slated to make the soft-landing attempt in Artemis IV. 

Dual-mode propulsion system technology developed by MIT engineers could give small satellites the ability to perform both powerful manoeuvres and fuel-efficient long-distance travel using a single propellant source.

Small satellites have transformed space research by making missions cheaper and more accessible. Yet they continue to face a fundamental limitation: propulsion.

Traditional chemical thrusters provide powerful bursts of speed but consume large amounts of fuel. Electric propulsion systems, on the other hand, are highly efficient but generate only gentle thrust over long periods. Spacecraft designers have typically had to choose between the two.

Engineers at the Massachusetts Institute of Technology (MIT) now believe they have found a way to combine both approaches in a single compact system, potentially giving small satellites the agility of much larger spacecraft.

The breakthrough centres on a special propellant capable of powering both chemical and electric thrusters from the same fuel tank.

“If you can have chemical and electrical propulsion in one small package, it’s the best of both worlds,” said Amelia Bruno, lead author of the study and a former postdoctoral researcher in MIT’s Department of Aeronautics and Astronautics, in a media statement.

“This opens the door for small satellites to do even more science, more observations, and more interesting missions, all on a smaller and cheaper platform.”

The findings have been published in the Journal of Propulsion and Power.

Dual-Mode Propulsion System Combines Two Technologies

The MIT team tested a propellant known as Advanced SpaceCraft Energetic Non-Toxic propellant, or ASCENT. Originally developed by the U.S. Air Force as a safer alternative to hydrazine, ASCENT was designed for chemical propulsion systems.

Researchers discovered that the same propellant can also power miniature electric propulsion devices known as electrospray thrusters.

These tiny thrusters use electric fields to charge particles within a liquid propellant and eject them into space, creating precise and fuel-efficient thrust. While chemical thrusters are ideal for rapid manoeuvres, electrospray systems are better suited for gradual course corrections and long-duration journeys.

By enabling both systems to share a single fuel source, the technology could significantly reduce the size and complexity of propulsion systems aboard CubeSats and other small spacecraft.

Dual-Mode Propulsion System Could Expand Deep-Space Missions

Dual-mode propulsion system can expand deep-space missions. The implications extend beyond Earth orbit.

CubeSats have become popular for scientific research and technology demonstrations, but their limited propulsion capabilities have restricted their use in deep-space missions.

According to Paulo Lozano, the Miguel Alemán Velasco Professor of Aeronautics and Astronautics at MIT, the new system could change that.

“We could send CubeSats to Mars, or the asteroid belt, where they could make the journey slowly, using electrospray thrusters,” he said.

“You could then use your chemical thrusters to quickly move to look at interesting features. You could have a lot more flexibility to do a lot more things.”

Testing the Technology

To evaluate the propellant’s performance, the researchers filled small CubeSat reservoirs with ASCENT and tested them in a vacuum chamber designed to simulate conditions in space.

During the experiments, electrospray thrusters powered by ASCENT successfully generated thrust for extended periods, in some cases operating continuously for up to 100 hours.

NASA Mission Will Put the Technology to the Test

The next major test will come later this year.

MIT researchers are working with NASA on the Green Propulsion Dual Mode mission, a CubeSat that will carry both chemical and electrospray thrusters powered by a single propellant tank. Scheduled for launch in November, the mission will be the first demonstration of such a system in a small spacecraft.

If successful, the mission could help pave the way for a new generation of versatile satellites capable of switching between rapid manoeuvres and highly efficient long-distance travel.