Chinese scientists at the Beijing Institute of Technology have run satellite proximity simulations to extend potentially hostile observation of U.S. space assets in Geostationary Earth Orbit (GEO), seemingly inching both countries closer to a potential war in space.

Although the new simulations focus on orbital adjustments to extend observation times and increase proximity to U.S. satellites, including critical missile warning and other space defense-oriented systems, Chinese scientists also plan to integrate reinforcement learning methods into future space assets to give satellites even more independence from human operators.

Study Includes Accusations of Cold War in Space Activities

In the 20th century, the concept of war in space remained mostly in the realm of science fiction, as depicted in popular films like Star Wars. Today, advances in space access, orbital maneuvering technologies, and weapons systems have made the possibility of a future conflict in space a very real security concern, and several nations engage in espionage using reconnaissance spacecraft to obtain information on valuable defense targets in Earth’s orbit.

According to the authors of the new study, “acquiring multi-aspect payload information of non-cooperative targets to infer their functions and mission statuses is of great significance for space security assessment.”

The study authors highlight a series of relative motion configurations that modern spy satellites employ, including fly-around, hovering, and drifting flight. They also point to real-world “adversarial” U.S missions they suggest were designed to gather intelligence on their GEO defense assets.

“Since 2014, the US GSSAP (United States Geosynchronous Space Situational Awareness Program) satellite has carried out hundreds of maneuvers and has conducted close observation operations on dozens of normally working satellites near GEO orbit,” the Beijing University of Technology scientists write.

The study authors also include a specific July 2021 accusation, where they charge that “the U.S. satellite ‘USA-271’ deliberately approached the operational ‘Practice 20’ satellite in geostationary orbit and flew in close proximity.”

In an email to The Debrief, a U.S. Space Command Spokesperson did not address the specific incident or the accusations of “dozens” of similar operations, but offered some insight into the agency’s thinking.

“USSPACECOM, working with allies and partners, plans, executes, and integrates military spacepower into multi-domain global operations in order to deter aggression, defend national interests, and when necessary, defeat threats,” the spokesperson told The Debrief, adding that “we have nothing additional to add in response to the article.”

Getting Closer to Adversarial Satellites and Extending Observation Time

In a step toward potential dogfights between satellites, the Chinese scientists looked at several challenges faced by current spy satellites.

For example, they explain, since most spacecraft observation cameras are mounted in a fixed position, “frequent large-angle attitude maneuvers are time-consuming and consume substantial fuel.” The researcher also noted that several different angles and approaches could yield different categories of intelligence data.

However, changing angles and orbital planes requires fuel, an extremely limited resource in GEO. Therefore, the researcher accounting for all the variables required to gather intelligence on adversarial spacecraft as efficiently as possible “has become a key bottleneck in enhancing space situational awareness capabilities.”

To find an optimized solution, the team designed what they called a “waypoint trajectory planning scheme” and a “multi-impulse maneuver strategy” using a common game-theory tool called “minimum integral squared control” theory. Unlike previous research aimed at extending spy satellite observation time and fuel efficiency, the Chinese scientists said their use of sequential coalition game theory “is introduced for the first time into the multi-model combined observation problem.”

Next, the researchers designed a “game tree.” A press release said this approach lets the spacecraft select the “optimal combination of initial motion points and close-range operation models at each stage under minimum fuel constraints, thereby generating an optimal observation strategy sequence.”

Finally, the researchers ran simulations to see whether their approach improved sensor observation times and field of view while reducing fuel use.

Successful Simulations and Future Reinforcement Learning Plans

As hoped, the Beijing University of Technology team said their simulations supported the theories. Specifically, the study authors said the simulations showed that their spacecraft efficiency algorithm “can successfully generate a multi-model combined observation strategy sequence under minimum fuel constraints.”

“The effective observation time increases with larger observation distances and field-of-view angles, and reducing the model configuration size also contributes to improved observation performance,” they explained. The study authors also said that the simulation results “demonstrate that the algorithm can successfully generate the optimal combined observation sequence under minimum fuel constraints.”

When discussing next steps, the Chinese scientists said they will explore optimal initial points to further improve maneuver efficiency. They also plan to explore a “reinforcement learning method,” suggesting future spacecraft could learn and adapt to evolving conditions to further enhance their cold war in space information-gathering capabilities.

Study’s War in Space Implications “Extends Beyond any Single Algorithm”

To understand the potential implications of the Chinese study, including its impacts on a potential future war in space, The Debrief reached out to former U.S. Air Force officer Shelli Brunswick, CEO and Founder, SB Global LLC.

According to the author and futurist, the Chinese team’s research matters not just because it proves China has deployed a “new operational surveillance capability.” Instead, Brunswick said that the work is significant because it “illustrates how sophisticated coordination and optimization could make spacecraft observation more efficient, persistent, and strategically useful.”

The Blue Marble Space Visiting Scholar also told The Debrief that the work “should not be interpreted as demonstrating a complete pursuit-and-evasion capability against a spacecraft that is actively responding,” since the work assumes a stationary target without maneuvering capabilities. Still, Brunswick noted, the Chinese framework “shows how game-theoretic planning could help an observing satellite gather information from multiple angles while managing limited resources.”

When discussing the potential implications for the broader U.S.-China space-security competition, including how such enhanced spacecraft could lead to more activities that could be construed as hostile, the space strategist said this area of research “extends beyond any single algorithm.”

“As rendezvous and proximity operations become more sophisticated, the line between inspection, space situational awareness, intelligence gathering, and potentially threatening behavior becomes increasingly difficult to interpret,” she explained.

When suggesting what potential strategic responses the U.S. and its allies might employ in the emerging cold war in space, Brunswick told The Debrief that they should include “more resilient architectures, improved space-domain awareness, clearer expectations for close approaches, stronger communication mechanisms, and deeper coordination with commercial and international partners.”

“The challenge is not simply detecting another spacecraft—it is understanding capability and intent early enough to respond responsibly and effectively,” Brunswick explained.

The study “Spacecraft Proximity Operation Model-Based Sequential Coalitional Observation Game Strategy Design” was published in Space: Science & Technology. 

Christopher Plain has spent the last six years as Associate News Editor and Head Science Reporter at The Debrief. Follow and connect with him on X, learn about his novels at plainfiction.com, or email him at christopher@thedebrief.org.

Scientists from the Beijing Institute of Technology have revealed a colloidal quantum dot (CQD)-based infrared-to-visible upconverter that transforms the infrared spectrum into full color, allowing people to ‘see’ the infrared in the visible light spectrum.

The research team behind the novel invention said their upconverter can be used in eyeglasses to improve human vision and for augmented reality (AR) applications. They also reported a successful implantation of an upconverter in mice, opening the possibility of an implantable photoreceptor to repair or improve vision in human eyes.

Attempts to Convert Infrared to Full Color Face Limitations

In a published study detailing the novel, full-color IR glasses, the human eye is “inherently blind” to infrared radiation. Although IR radiation interacts with the retina, infrared wavelengths lack the energy to trigger the photoisomerization of the retinal chromophore that lets people distinguish colors. As a result, humans don’t have visual access to over half of the solar spectrum.

Still, the researchers write, “harnessing this imperceptible radiation is critical, as infrared perception unlocks a vast reservoir of inaccessible information.” For example, this added visual information enables critical functions in other animals, such as navigation in darkness or detecting predators and prey from their heat. Still, they observe that our natural visual-light spectrum vision is “utterly inadequate” to access this information that so many animals rely upon.

In their study, the researchers highlight recent advances in accessing the IR spectrum, including “ocular injectable photoreceptor-binding upconversion nanoparticles” that convert infrared light into visible emissions. Other studies have also evaluated wearable lenses with nonlinear upconversion projecting infrared-transformed visible light onto the cornea and photovoltaic nanowire networks.

Although these approaches achieved some IR sensitivity, the study authors said they remain “fundamentally constrained” to narrow near-infrared (NIR) spectral response. These prior efforts also relied on high-intensity laser excitation that the team notes is “incompatible with natural infrared radiation.” The result is a “mostly monochromatic output” lacking infrared spectral discrimination in full color.

Subtle IR Wavelength Variations Facilitate Upconversion

For their new approach, the research team investigated a mercury telluride (HgTe) CQD-based infrared-to-visible upconverter. They note that this material combination leverages the material’s “quantized energy level states” in synergy with a separate, dual-emissive-layer organic light-emitting diode (OLED).

Specifically, the engineered layers captured IR photons in strategically engineered hole-trapping barriers. The team explains that this “allows for subtle differentiation of infrared input by tuning the emitted visible color and luminance” based on infrared wavelength and intensity.

“This upconversion yields a discrimination sensitivity to subtle infrared variations exceeding two orders of magnitude higher than conventional single-color modes, capitalizing on the human eye’s intrinsic superiority in chromatic differentiation over simple luminance contrast,” they explain.

To test their material’s IR upconversion ability, the researchers designed a pair of lightweight (23 grams) semi-transparent wearable eyeglasses. The team said these tests were a success, achieving “high-resolution infrared imaging” derived from the infrared spectral information. However, instead of monochrome, the research team said that data is displayed as full-color visible patterns directly onto the retina, “enabling intuitive infrared vision without obstructing natural sight.”

Surpassing the Evolutionary Boundaries of Biological Photoreception

Although the current version of the full-color, infrared-capable glasses is still in the development stages, the team said the technology could lead to advancements in visual prosthetics, augmented reality (AR), and “navigation under environmental conditions where normal sight is degraded.”

Following the successful validation in mice, the team said their technology could also serve as the technological basis for an implantable new-generation retinal photoreceptor. In that design, the team said an implanted device could transform infrared light into visible light emissions “that stimulate light-sensitive proteins on retinal neurons to bypass damaged photoreceptor cells and potentially restore visual function across both the visible and infrared spectra.”

“By surpassing the evolutionary boundaries of biological photoreception, this technology paves the way for next-generation visual prosthetics, high-fidelity augmented reality, molecularly sensitive substance identification, and robust navigation in degraded visual environments,” they conclude.

The study “Multispectral infrared-to-full-color upconversion expanding human vision” was published in Science Advances.

Christopher Plain is a Science Fiction and Fantasy novelist and has spent the last six years as Associate News Editor and Head Science Writer at The Debrief. Follow and connect with him on X, learn about his books at plainfiction.com, or email him at christopher@thedebrief.org