Product1 publisher2 min readPublished
SoftBank tested its laser ground station against a passive mirror 20km up over Kochi
The August demonstration off Cape Muroto put a corner cube reflector on a Sceye airship and left the demanding part on the ground. There, a portable ranging system had to hold a narrow pulsed beam on a moving target and catch the return.
The Product Desk · Product desk
What happened
- SoftBank, Hitotsubashi University and Japan's National Institute of Polar Research tracked a high-altitude communications platform from the ground using laser pulses, a step toward optical links between ground, stratosphere and satellites.
- The runs took place on August 23 and 24, 2026, following an initial observation and software adjustment on August 22, with the platform flying over waters off Cape Muroto in Kochi Prefecture.
- The platform was an LTA-type HAPS operated by the US company Sceye, and the optical payload on board was a corner cube reflector.
- Instead of passing communications data between an airborne terminal and a ground station, the test created a round-trip laser path from the ground to the HAPS and back.
- The experiment also produced measurements of how much the laser signal was attenuated on the way between the ground and the stratosphere.
Compiled by The Product DeskSomething wrong?How this is made
Why it matters
- constraint A reflector can only send light back, so nothing gets communicated until transceivers fly on the platform itself; a HAPS still has to carry the whole list of hardware.
- capability The attenuation measurements give engineers install and operating requirements for terminals that have not been built. A bench test could not have produced that evidence.
- decision Teams building a two-ended optical link now have a worked example of buying evidence about the ground segment before committing a terminal to a flight article.
- exposure The beam shares airspace with crewed traffic, and the demo's shutdown-on-detection rule is the sort of operating condition a production ground station will inherit.
A corner cube reflector does one thing: it sends incoming light back toward its source, whatever angle the light arrives at [7]. The ground side had the rest. A compact, portable ranging system had to find a platform operating around 20 km up whose position and orientation change constantly. Then it had to hold a narrow pulsed beam on that platform while picking up a return that comes back inside a very narrow angular range [8][13][3]. SoftBank groups those functions as pointing, acquisition and tracking [9].
SoftBank says the pointing and receiving-tracking accuracy required for satellite laser ranging is roughly an order of magnitude higher than that used in conventional optical communication links [14]. The ranging job is the tighter one, so a ground station that passes it has been tested harder than the comms link it is meant to support. And the company says the tracking system it is developing for its optical communications terminal uses the same software architecture as the laser-ranging system [20].
Optical links out of the stratosphere are older than this test. The German Aerospace Center's STROPEX experiment ran a 1.25 Gbit/s downlink from a balloon at about 22 km in 2005 [16], 21 years before the Kochi flights [1]. The US Navy's Naval Information Warfare Center Pacific has developed the Stratospheric Optical Link Demonstration as a high-altitude balloon experiment to reduce risk for optical communications technologies [17]. SoftBank's claim is narrower than a first for optical work at that altitude. It calls this the world's first demonstration of this specific form of ground-based laser ranging of a stratospheric HAPS, based on publicly available information the three organisations reviewed through September 8, 2026 [4].
Calling the reflector a cheaper payload than a terminal is an inference. SoftBank did not disclose cost or mass figures for the reflector or for a flight terminal [2]. The stated objective is a three-dimensional communications network in which HAPS platforms connect ground users [21].
For a team sequencing a link that has two ends, the useful question is which end holds the uncertainty you are paying to remove. Then whether a passive stand-in on the far end produces the same evidence as the real thing. In this case the stand-in answered the ground-side questions and left the airborne ones open. What came out of Cape Muroto is a ground station that acquired and continuously tracked a moving stratospheric target and read its return [3].
What to watch
- Whether SoftBank publishes the measured atmospheric attenuation figures and the link budget they imply for a HAPS terminal.
- Whether the next flight carries an actual optical transceiver on a Sceye platform and the ranging software transfers without rework.
- Whether the crewed-aircraft transmission cutoff used in the test becomes a condition on operational ground stations.