High Altitude Platform Stations, or HAPS, are often described rather simply as "flying base stations". The idea is to place a communications platform in the stratosphere, typically around 20 km above the Earth, where it can provide mobile coverage across a much larger area than a conventional terrestrial base station. SoftBank's latest HAPS trials in Japan suggest that this description may increasingly undersell what these platforms could eventually do.
In August 2026, SoftBank and US-based Sceye conducted a series of HAPS communications trials off Cape Muroto in Kochi Prefecture. The tests included conventional smartphone connectivity, communications with drones and, perhaps most interestingly, edge computing performed directly aboard the HAPS. SoftBank sees these capabilities eventually becoming part of a three-dimensional communications network integrating terrestrial networks, HAPS and satellites.
The platform itself was Sceye's lighter-than-air, or LTA, HAPS. Unlike fixed-wing HAPS aircraft that rely on aerodynamic lift, an LTA platform uses buoyancy to remain aloft. The Sceye HAPS departed New Mexico on 9 August and travelled more than 15,000 km across the Pacific in the stratosphere, reaching Japanese airspace around 13 days later. Once over the test area, Sceye was able to keep the platform within a radius as small as 5 km for an extended period despite changing stratospheric conditions.
This part of the demonstration is important in its own right. Communications payloads are only useful if the platform carrying them can remain reliably where coverage is required. Endurance, station keeping, energy availability, payload capacity and the ability to operate in changing stratospheric winds are therefore just as important to HAPS deployment as the radio technology itself.
For the mobile connectivity tests, a 4G LTE-equivalent base station installed on the HAPS communicated directly with smartphones on the ground. The airborne base station was connected through a ground gateway to SoftBank's core network. The companies demonstrated voice calls, text messaging, social media applications, video calls and video streaming. They also tested earthquake and tsunami emergency alerts and calls to Japan's 118 maritime emergency number. SoftBank says the trial achieved communications performance equivalent to terrestrial networks while reducing interference with terrestrial base stations.
Disaster recovery is one of the obvious applications for HAPS in Japan. An earthquake, tsunami or other major event can damage terrestrial base stations, transmission links and power infrastructure at exactly the time when communications become most important. A HAPS positioned above an affected region could potentially restore a wide layer of coverage without waiting for terrestrial infrastructure to be repaired.
But SoftBank's trial went considerably further than simply providing connectivity to people below. The companies also established communications between the HAPS and a drone. Using HAPS connectivity, they demonstrated remote automated flight control, transmission of the drone's location and video transmission. SoftBank sees potential applications including disaster assessment and supply delivery, monitoring remote islands and offshore areas, logistics in mountainous regions, infrastructure inspection and forest monitoring.
3D Communication Networks and the Future of Connectivity - https://t.co/oqeDcXLxRi via Connectivity Technology Blog#Free5Gtraining #3G4G5G #ConnectivityTechnology #3DNetwork #3DCommunications #Terrestrial #Satellite #NTN #HAPS #UAV #UAS #FSO #Fiber pic.twitter.com/AfruXCZp9q
— Free 5G Training (@5Gtraining) March 9, 2026
This demonstrates why HAPS is interesting in the context of future three-dimensional networks. Conventional mobile networks have primarily been designed to serve users and devices relatively close to ground level. Drones, autonomous aircraft and other aerial systems create a different coverage problem. A platform looking down from the stratosphere potentially has line-of-sight connectivity to both devices on the ground and aircraft operating below it.
The most technically interesting part of the experiment, however, may have been what happened inside the HAPS itself. SoftBank installed a mobile core network and a web server aboard the platform so that smartphone traffic could be processed directly in the stratosphere rather than being sent through the terrestrial network to an Internet cloud. SoftBank describes this, based on publicly available information available to it at the time, as the world's first successful test in which a mobile core and web server aboard a HAPS processed smartphone communications entirely within the airborne platform and returned the results to the smartphone.
In SoftBank's test, the average round-trip response time when processing was performed by the server aboard the HAPS was 68 milliseconds, more than 40% lower than when processing was performed in the cloud via the Internet. This should not be interpreted as a general HAPS latency figure, since it reflects this particular trial and network configuration, but it demonstrates an interesting architectural possibility. Instead of acting purely as a radio access point that forwards everything towards infrastructure somewhere else, a HAPS could potentially host parts of the mobile network and computing infrastructure itself.
That could be particularly useful where terrestrial connectivity is damaged, unavailable or geographically distant. An application could communicate with the HAPS, have some of its data processed locally in the stratosphere and receive a response without every transaction making a round trip through a remote terrestrial data centre. SoftBank sees possible applications including real-time video analytics and Physical AI, where AI systems interact with cameras, sensors, robots and other physical systems.
There are, of course, practical limits. Computing equipment consumes power and adds weight, both of which are scarce resources aboard long-endurance airborne platforms. Processing capacity, cooling, reliability and maintenance all become more challenging when the edge data centre is around 20 km above the ground. The amount of computing that ultimately makes sense aboard a HAPS will therefore depend heavily on the application and platform design. Nevertheless, the trial demonstrates that the HAPS architecture does not necessarily have to stop at the radio access network.
Another experiment associated with the same HAPS flight points towards how these platforms might eventually connect into a much larger network. On 17 September, SoftBank, Hitotsubashi University and Japan's National Institute of Polar Research announced results from a laser-ranging experiment conducted while the Sceye HAPS was operating off Cape Muroto in August. A corner cube reflector was installed on the HAPS and a portable laser-ranging system on the ground continuously tracked the moving platform while transmitting laser pulses towards it and receiving the reflected signals.
This was laser ranging and tracking, not an optical communications link. That distinction is important. The experiment was intended to gather information relevant to future optical wireless communications, including tracking performance and attenuation of light as it travels through the atmosphere between the ground and stratosphere. The longer-term architecture is potentially much more interesting. SoftBank is studying optical wireless communications between terrestrial networks, HAPS and satellites. It says it plans to demonstrate bidirectional optical wireless communications between a LEO satellite and HAPS in 2027.
Put these pieces together and the eventual architecture becomes quite interesting. Smartphones and other terrestrial devices could connect to HAPS over conventional radio interfaces. Drones and other aerial systems could connect to the same airborne layer. Some traffic could potentially be processed aboard the HAPS, while other traffic could be carried back to terrestrial infrastructure through feeder links.
In the future, high-capacity optical links could add another dimension, potentially connecting ground stations to HAPS and HAPS to satellites. This would make the HAPS much more than simply an unusually high base station. It could become an intermediate network layer connecting terrestrial, stratospheric and space infrastructure.
That is particularly relevant as the industry moves towards 6G and increasingly incorporates Non-Terrestrial Networks into the wider communications architecture. Rather than considering terrestrial mobile networks, HAPS and satellites as three independent connectivity solutions, future networks could potentially select and combine different layers depending on coverage, capacity, resilience and application requirements.
HAPS could also occupy an interesting position between terrestrial networks and satellites. At around 20 km altitude it is vastly closer to users than a satellite, while still having line-of-sight over a large geographical area. Unlike a satellite it can potentially remain over a particular service area, and unlike conventional terrestrial infrastructure it is less dependent on local ground conditions. These characteristics make HAPS particularly interesting for remote coverage, emergency restoration, temporary capacity and aerial connectivity.
There are still substantial challenges before HAPS becomes a routine component of commercial mobile networks. Platforms need to operate reliably for long periods, maintain their position, generate sufficient energy, carry useful communications and computing payloads and coexist with terrestrial networks. Operators also need viable economics, appropriate spectrum and aviation arrangements, operational processes and a sufficiently strong advantage over expanding terrestrial coverage or using rapidly evolving satellite connectivity.
SoftBank and Sceye are nevertheless moving beyond demonstrating that a base station can simply be placed in the stratosphere. The latest tests combined a long-endurance airborne platform, ordinary smartphone communications, disaster services, drone connectivity and computing aboard the HAPS itself. SoftBank says it is now working towards commercialising HAPS services in Japan from 2027 onward.
The familiar description of HAPS as a "flying base station" therefore remains useful, but it may soon become incomplete. If onboard computing, aerial connectivity and eventually high-capacity links between terrestrial networks, the stratosphere and satellites develop as planned, HAPS could become an entire network node in the sky rather than simply another place to mount a radio.
Related Posts:
- Free 6G Training: SoftBank Advances Nullforming Towards Practical NTN Spectrum Sharing
- Connectivity Technology Blog: Softbank's Cylindrical Antenna for HAPS to Reduce Handovers
- Free 6G Training: Softbank's 12 Challenges for Beyond 5G / 6G
- Connectivity Technology Blog: SoftBank to Promote Non-Terrestrial Network (NTN) Solutions
- Telecoms Infrastructure Blog: Softbank Demoed Drone Wireless Relay System
- Connectivity Technology Blog: HAPSMobile and Loon Partner for Stratosphere-Based LTE Communications Solution
- Connectivity Technology Blog: HAPSMobile - Bringing Connectivity from the Sky for Unserved Areas & Emergencies
Comments
Post a Comment