A phased-array antenna points its beam without moving. Instead of a dish swinging on a motor, it uses a flat panel of many small radiating elements and steers the beam by adjusting the relative phase of the signal fed to each one — shifting where the combined wavefront constructively reinforces. Change the phases, and the beam swings electronically, in microseconds, with no moving parts. That capability is the enabling technology behind low-Earth-orbit (LEO) broadband: a LEO satellite crosses the sky in minutes, far too fast for a mechanically steered dish to track cleanly across a constellation handover, so the user terminal needs a beam it can re-aim electronically. In the patent record, this technology lives in a specific, findable corner — and reading it correctly means separating the general capability from what any single patent actually claims.

Classification is the entry point. The Cooperative Patent Classification routes satellite-communications inventions into group H04B7, "Radio transmission systems, i.e. using radiation field," with the H04B 7/185 series specifically covering systems using a satellite or space-borne relay station — and finer codes such as H04B 7/18513 and 7/18515 for the space-segment and ground-segment ends of the link. Antenna structure itself is classified in H01Q (antennas), with codes like H01Q 21 for arrays and H01Q 3 for beam-steering arrangements. A phased-array satellite-terminal patent therefore typically carries a cluster of codes at once — H04B7 for the satellite-link function and several H01Q codes for the array hardware — and that pairing is itself a signature. When you see H04B 7/185 sitting alongside H01Q 21 on a grant, you are almost certainly looking at a satellite phased-array filing.

To see what such a patent covers, read a real one. U.S. Patent 10,958,336 B2, titled "Phased array antenna for use with low earth orbit satellite constellations," is directly on point, classified in H04B 7/1858 and a stack of H01Q codes. Its claim 1 reads:

"A phased array antenna system for use with a Low Earth Orbit ('LEO') satellite constellation, comprising: a plurality of antenna panels positioned in a dome, wherein each of the plurality of antenna panels comprises an array of meta-structure cells; and an antenna controller to control the plurality of antenna panels, the controller directing a first antenna panel to transmit a first signal and a second antenna panel to transmit a second signal to a LEO satellite, the first signal having a first phase and the second signal having a second phase different from the first phase."— US10958336B2, claim 1, patent record

What the claim actually protects

The claim makes the engineering explicit. The system is a set of antenna panels arranged in a dome — a geometry that gives the terminal coverage across a wide swath of sky as satellites rise and set. Each panel is built from "an array of meta-structure cells," meaning engineered metamaterial elements whose electromagnetic response is tuned cell by cell. A controller drives different panels to transmit at different phases — "a first phase" and "a second phase different from the first phase" — which is the phase-steering mechanism stated in claim language. The dependent claims in the same patent fill in the implementation: a feed network, a radio-frequency integrated circuit (RFIC) generating the phase shifts, varactors whose applied voltage sets each cell's phase, and subarray groupings. Those are the knobs that turn the general idea into a buildable terminal.

The decisive point for any competitive read is the limitation "meta-structure cells." The abstract speaks broadly of a phased array for LEO constellations, but claim 1 requires meta-structure (metamaterial) cells specifically. A LEO phased array implemented with conventional patch elements and discrete phase shifters — rather than tuned meta-structure cells — would not read on this claim, however similar its function. This is the recurring lesson of antenna IP: the field is crowded with phased-array patents, and they do not all cover the same thing. Each one stakes out a particular implementation — a particular cell technology, feed architecture, packaging geometry or beamforming method — and the boundary between them is in the claim limitations, not the shared abstract language about steering beams to satellites.

Why this is the comms battleground

It helps to understand why the dome geometry and electronic steering recur across these filings. A LEO satellite is not parked over one spot the way a geostationary satellite is; it sweeps across the sky and is replaced by the next satellite in the constellation minutes later. A user terminal must therefore acquire a fast-moving target, track it across the pass, and hand off to a rising satellite without dropping the link — potentially while also forming a second beam toward the incoming satellite. A mechanically steered dish cannot slew fast enough or form two beams at once; an electronically steered phased array can re-aim in microseconds and, with enough elements and the right control logic, support multiple simultaneous beams. The domed arrangement in the patent above addresses the geometry problem directly: panels facing different parts of the sky keep at least one panel pointed near any satellite as it rises and sets. These are the physical constraints that the claim limitations encode, which is why so many LEO-terminal patents converge on similar structural language even as their specific cell technologies and beamforming methods diverge.

Beamforming is where LEO-communications competition concentrates, because the antenna is the architecture: spectrum efficiency, the number of simultaneous beams, handover smoothness between satellites, and terminal cost all trace back to how the array is designed and controlled. That is why the H04B7/H01Q corpus is dense and why assignees from satellite operators to terminal specialists file heavily in it. For an analyst, the right altitude is to use the classification to gather the set — pull H04B 7/185 paired with H01Q array codes — and then read each candidate's independent claim to see which slice of the design space it fences. The classification tells you a patent is about satellite phased arrays; only the claim tells you whether it covers meta-structure domes, a particular RFIC beamforming scheme, or something else entirely. Treat the two as one and you will overstate overlap; read the claim and you will see exactly where each filing's wall actually stands.