020 | Hedy Lamarr, George Antheil and the Frequency-Hopping Patent

“The Signal That Would Not Sit Still”

On 11 August 1942, while cinema audiences knew Hedy Lamarr as one of MGM’s most glamorous stars, the United States Patent Office knew her rather differently. Patent 2,292,387 names Hedy Kiesler Markey (her legal name at the time) and George Antheil as the inventors of a ‘Secret Communication System’. There is no publicity photograph on the cover. There are two sheets of circuit diagrams, two synchronised paper records, a radio-controlled torpedo and six claims.^[2]

The problem was brutally practical. A ship could steer a torpedo by radio, but the enemy could discover the control frequency and transmit on it too. Enough noise might drown the instruction; false commands might corrupt it. A weapon that stopped obeying its launcher once the target objected was not an especially satisfactory weapon.

Lamarr and Antheil’s answer was to move the control signal repeatedly from one radio frequency to another. The transmitter would hop. The receiver inside the torpedo would hop through the same sequence at the same moments. An enemy monitoring one frequency would hear only fragments. An enemy jamming one frequency would block only the brief periods in which the signal happened to visit it.

This is the mechanism now widely known as ‘frequency hopping’. It is also where the popular version of the story begins to overheat. Lamarr and Antheil did not conjure the broad idea of changing frequencies from nothing; earlier patents existed, and the Patent Office rejected their broadest claim. Their issued patent protected a particular, ingenious synchronisation apparatus based on paired record strips—the mechanical logic of a player piano (pianola) applied to contested radio.^[2]

That qualification does not demote Lamarr. It rescues her from two bad histories at once: the old one in which a beautiful actress could not possibly have invented anything, and the new one in which admiration must be inflated until her patent personally becomes Wi-Fi, Bluetooth, GPS and every mobile telephone. A patent can be important without being a creation myth. A woman can be restored to engineering history without being turned into a saint, nun or miracle.

One frequency, one easy target

A radio transmitter places information onto a carrier wave. The receiver is tuned to that carrier frequency and extracts the information: in this case, momentary instructions to move a torpedo’s rudder left or right. If transmitter and receiver remain on one known channel, an opponent’s task is relatively simple. Find the channel. Put enough competing energy onto it. Continue until the legitimate receiver can no longer distinguish command from interference.

The interference need not contain an intelligent message. It may be deliberate jamming, another transmitter, electrical noise or an accidental collision with unrelated radio traffic. From the receiver’s point of view the result is similar: the signal-to-interference ratio has become too poor for reliable detection. The message may still exist in a philosophical sense. The torpedo, inconveniently, requires the message in an engineering one.

Now suppose the transmitter occupies frequency f3 for a short interval, then f17, then f6, then f12, following a schedule known to the receiver. A jammer sitting on f3 wins briefly and then finds itself shouting into an empty room. To keep winning, the jammer must predict the schedule, detect and chase the hops quickly enough, or transmit across the whole possible band.

That changes the economics of attack. Guessing is unreliable. Chasing introduces detection and retuning delay. A barrage jammer can cover many channels at once, but its available power is then spread across a wider bandwidth; unless it has considerably more power, the interference delivered to each channel falls. The jammer now has a choice: guess, chase or shout across the whole band. All three cost more.

The clever part was moving together

Changing frequency is easy to describe. Changing it usefully is a problem of shared state and time.

If the transmitter moves to f17 while the receiver is still listening on f3, nobody has jammed the link. The system has defeated itself. Both ends need the same channel list, the same order, the same starting point and sufficiently accurate clocks. They also need a plan for drift, interruption and recovery. A receiver waiting on the wrong frequency is not secure. It is merely lonely.

Lamarr and Antheil proposed two matching records, one at the transmitter and one inside the torpedo. Perforations in the moving strips operated switches that selected different tuning components. Constant-speed spring motors, like those used in clocks and chronometers, pulled the records past their readers. When the torpedo was released, an electrical connection holding both strips at a start hole was broken, releasing them together. The patent also allowed for synchronising impulses to correct the receiving mechanism’s phase.

The 88 frequencies repeated in almost every account did not come from a sacred radio calculation. A conventional player-piano roll could provide 88 longitudinal rows because a piano has 88 keys. Each row could select a carrier frequency. The published drawings use a smaller illustrative system (seven transmitting channels and four receiving channels) but the principle could scale.

Antheil’s musical machinery was not a whimsical decoration added to make the story marketable. He had written Ballet Mécanique for an ensemble that originally involved sixteen player pianos and had learned, painfully, how difficult it was to make mechanical systems perform together. He also held an earlier French patent concerning a device and paper for recording music. Synchronisation was part of his technical vocabulary.

The trick was not moving. It was moving together.

A hopping sequence, simplified (illustrative frequencies). The receiver follows the same schedule; a jammer covering one channel disrupts only the slot in which the legitimate signal happens to visit it.

Time slot	1	2	3	4	5	6
Transmitter	f3	f17	f6	f12	f2	f14
Receiver	f3	f17	f6	f12	f2	f14
Jammer fixed on f3	BLOCK	miss	miss	miss	miss	miss

Inside the patent

The patent rewards close reading because it is more interesting than its slogan. Lamarr and Antheil were not merely proposing a radio that changed channel every so often. They designed a remote-control system around very short, infrequent command pulses.

At the transmitting station, keys marked L and R selected left or right rudder. The carrier was modulated with a 100-cycle tone for one direction and a 500-cycle tone for the other. At the receiver, filters separated the tones and electromagnets moved a ratchet mechanism, nudging the rudder by one increment for each command. The patent emphasised that a key need be closed only momentarily. A short transmission was harder to discover, and a single received pulse caused a bounded mechanical action rather than an indefinite turn.

The drawings also include deception. Seven channels are available at the transmitter but only four at the receiver. Signals sent on the other three are deliberately ineffective: the torpedo does not listen there. A lamp tells the operator when the two legitimate ends are on a shared channel. To an enemy, however, the false transmissions complicate the question of which bursts matter. The patent even suggests arranging the records so the receiver is never twice tuned to the same frequency.^[2]

So the design combined frequency agility, synchronisation, brief exposure, decoy traffic and constrained commands. It was not modern digital spread spectrum. It was a mechanical-electrical security system thinking in layers.

What hopping buys

Frequency hopping can improve availability because a disturbance affecting part of the spectrum need not destroy the whole link. If one channel is noisy, only the transmissions occurring during visits to that channel are lost. A later digital system may detect the loss, add error-correcting redundancy or retransmit. Lamarr and Antheil’s torpedo scheme was simpler, but its short incremental commands reduced the consequence of any one missed pulse.

Against deliberate jamming, the benefit depends upon the attacker. A narrowband jammer that does not know the hopping sequence covers only a fraction of the visits. A reactive jammer must first notice a burst, identify its frequency and emit useful interference before the dwell time ends. A wideband jammer avoids the prediction problem but pays in power: covering ten times the spectrum with the same total energy delivers roughly one tenth as much interference per unit bandwidth, before allowing for implementation losses.

None of this makes jamming impossible. A sufficiently powerful barrage transmitter may overwhelm the entire receiver band. A fast follower may chase slow or predictable hops. An attacker may jam the synchronisation mechanism, saturate the receiver front end, compromise the hopping schedule or attack the command protocol instead. Radio security remains physics with an adversary, which is less forgiving than a diagram.

The important change is asymmetry. The defender is no longer offering one stationary point at which a modest attacker can apply all available effort. The signal spreads its exposure across possibilities; the attacker must spend more power, more bandwidth, more sensors, more speed or more knowledge to achieve the same denial.

Interception is not encryption

Hopping also makes casual interception harder. A receiver tuned to one channel hears disconnected pieces. If the hopping sequence is secret and false transmissions are mixed in, the listener must discover where and when to follow. In 1941, when monitoring equipment was narrower and more mechanical, that was a substantial obstacle.

It is not the same as encrypting the command. A wideband recorder can capture many frequencies at once. An array of receivers can watch several channels. A known or predictable schedule can be followed. If the underlying left and right signals remain intelligible, anyone who collects the complete hopped transmission can still interpret them—and perhaps forge them unless the system authenticates commands.

Frequency hopping is therefore primarily an availability and interference-resistance technique, with interception-resistance as a useful side effect. Confidentiality requires cryptography. Integrity and origin require authentication. Replay protection requires freshness. Secure synchronisation requires its own design. One mechanism may help several properties, but security collapses quickly when ‘harder to hear’ is promoted to ‘impossible to understand’.

Frequency hopping is not encryption wearing dancing shoes.

The wrong sort of expert

Hedy Lamarr was born Hedwig Eva Maria Kiesler in Vienna in 1914. She became internationally famous as an actress, arrived in Hollywood under contract to Louis B. Mayer and appeared in films including Algiers, Ziegfeld Girl and Samson and Delilah. Her face became public property in the way the film industry has always preferred women’s faces to be: endlessly described, evaluated and reproduced.

She had no engineering degree. Neither did she need permission to notice an engineering problem. Her first marriage, to Austrian munitions manufacturer Friedrich Mandl, brought her close to the weapons business. Later retellings embellish precisely what she learned and at which dinner table; the documentary record is safer on the essentials. She understood that radio-controlled weapons were vulnerable to interference, initiated the torpedo-control idea and explained its basis to Antheil.

They met in 1940 at the home of actress Janet Gaynor and costume designer Gilbert Adrian. Antheil was an American avant-garde composer, pianist, film-score writer, inventor and enthusiastic wanderer across disciplinary boundaries. Their collaboration joined Lamarr’s problem and central hopping concept to Antheil’s experience with recording mechanisms and synchronised machines. The surviving Smithsonian collection contains their invention notebook, sketches and correspondence with patent attorneys and the National Inventors Council. This was work, not a cocktail-party anecdote that somehow acquired six patent claims.^[1][3]

Other labour sat around the named inventors. The lawyers at Lyon & Lyon shaped the application. Patent-prosecution research also identifies Samuel Stuart Mackeown, a Caltech physicist and radio engineer, as helping to reduce the concept to workable radio engineering. The surviving evidence does not permit a neat percentage allocation of inspiration. It does permit us to avoid the habit of making technical assistance disappear merely because only two names fit comfortably in the headline.^[6]

Correcting Lamarr’s erasure does not require erasing Antheil, Mackeown, the patent attorneys or earlier inventors. It requires putting Lamarr where the records put her: at the beginning of the project and on the patent as co-inventor.

What, exactly, did they invent?

The broad idea of changing transmission frequencies for secrecy pre-dated their application. Willem Broertjes filed a US patent in 1929 describing messages transmitted over a group of secret working frequencies, alternated during the message, with the receiver retuned in synchronism. His design also proposed open-circuit or spacing frequencies to confuse an unauthorised listener.^[4] Still earlier systems divided communications among multiple frequencies.^[5]

This matters because patents are composed of claims; they are not medals for having a generally splendid idea. Lamarr and Antheil’s original application included a broad seventh claim covering synchronised frequency changes according to an arbitrary, non-repeating pattern. The examiner rejected it in light of prior art including Broertjes. Their attorney cancelled that claim. The six claims that issued are tied to the particular apparatus: elongated record strips, record-responsive switching, synchronous movement and the remote-control implementation.^[6]

The historically defensible description is therefore precise. Lamarr and Antheil were co-inventors of an important early frequency-hopping communication system and of the patented player-piano-style mechanism for synchronising it. They were not the sole originators of every abstract form of frequency hopping, and the patent did not claim all future spread-spectrum radio.

That is not a pedant’s demotion. Their design attacked the operational difficulty that makes the principle useful: how two separated radios agree on a rapid, irregular sequence while an enemy listens. It added short signals, false channels and a physical method that could be manufactured with the technology they knew. Ideas become engineering when they survive contact with timing, switches, motors, noise and the requirement to fit inside something travelling through seawater towards an armed opponent.

A patent can be important without being a creation myth.

The Navy’s shrug

Lamarr and Antheil offered the invention for the American war effort. It was not adopted as described during the Second World War. A paper-strip mechanism inside a torpedo raised serious questions: vibration, moisture, space, manufacturing tolerance, clock drift and reliable starting were not details that salt water would politely overlook. The Navy also had a long institutional talent for distrusting ideas that arrived from outsiders.

Antheil later believed that the player-piano analogy had encouraged officials to imagine an absurdly bulky instrument rather than a compact control mechanism. That recollection may capture the encounter; it should not be upgraded into a complete engineering evaluation. Nor should the popular claim that the patent was classified and locked in a vault be repeated as fact. The patent was published in 1942, diagrams and all. No wartime secrecy order prevented publication.^[1]

The exact mechanical system appears never to have been fielded. Related electronic frequency-changing systems emerged after the war, when stable electronics made rapid tuning and synchronisation more practical. A mid-1950s sonobuoy communication project cited the Lamarr-Antheil patent while developing a mechanical hopper, then encountered exactly the maintenance difficulties one might expect. The concept travelled; the piano rolls did not become standard naval equipment.^[6]

Neither inventor profited from the patent, whose term ended in 1959. Antheil died that year. Lamarr lived until 2000, long enough to see belated public recognition but not a royalty stream from every radio later placed into the same extended technical family.

Written out in plain sight

Lamarr was not literally absent from the formal record. ‘Hedy Kiesler Markey’ is printed on the patent’s first page. The Smithsonian archive preserves her sketches and correspondence. The erasure happened in the story told around the record.

For decades, she was introduced as an actress with an amusing technical hobby, if the invention was mentioned at all. Her beauty was treated as evidence against intelligence. Her lack of institutional engineering credentials was treated as evidence against technical agency. Antheil’s eccentric musical reputation did not help either, but the disbelief directed at Lamarr had a specifically gendered shape.

SPECIFICALLY GENDERED DISBELIEF A woman whose appearance had been industrially marketed was presumed incapable of thought that did not serve the marketing.

Hollywood put her face on posters. Engineering history removed her name from the diagram.

The Electronic Frontier Foundation recognised Lamarr and Antheil with a Pioneer Award in 1997. The National Inventors Hall of Fame inducted both in 2014. Those honours corrected public memory, but the correction still needs care. Lamarr should not have to become the solitary mother of wireless communications before we allow that she was an inventor. Equal historical citizenship is not conditional upon superhumanity.^[7][8][9]

The scandal is not that a film star could invent. The scandal is that people found the two roles mutually exclusive.

From torpedoes to Bluetooth—carefully

Frequency hopping became one branch of spread-spectrum communications: techniques that distribute a signal over more bandwidth than the minimum immediately required. Modern implementations replace punched paper and clockwork with oscillators, digital logic, shared keys and algorithms. They may hop faster, adapt the channel set, encode data across losses and authenticate the devices taking part.

The family resemblance is real. The original IEEE 802.11 wireless-LAN standard, ratified in 1997, included a frequency-hopping spread-spectrum option at 2.4 GHz alongside direct-sequence spread spectrum. Later Wi-Fi generations moved to other modulation schemes, so the sentence ‘Hedy Lamarr invented Wi-Fi’ is an affectionate overstatement rather than a technical history.^[10]

Bluetooth provides the cleaner modern illustration. Connected Bluetooth devices switch among channels according to a shared schedule. Adaptive frequency hopping also records which channels are performing badly and excludes them from the usable map; both devices share that updated map so they continue to move together. The purpose in an office or kitchen is usually coexistence rather than wartime anti-jamming, but the governing idea is familiar: do not let one noisy patch of spectrum own the entire conversation.^[11]

GPS is often added to the celebratory list, but GPS principally uses direct-sequence spread-spectrum and code-division techniques, not Lamarr and Antheil’s frequency-hopping mechanism. Cellular and military radio contain many related spread-spectrum ideas with complex, multi-inventor histories. Influence is not a straight wire from a 1942 drawing to every wireless chip.

The correct legacy is better

Lamarr and Antheil produced a striking early anti-jam implementation, centred on synchronised agility, that sits within the history from contested wartime radio to resilient digital communications. It was cited, remembered, reinterpreted and eventually honoured. The resemblance to Bluetooth is conceptual and technical; the claim that their exact patent ‘became’ all modern wireless is neither necessary nor true.

Frequency hopping in a modern security team

The patent’s most useful modern lesson is not ‘change things rapidly’, despite the evidence to the contrary on the ground. It is instead that agility must be coordinated, bounded, and attached to a threat model.

Moving-target defences rotate addresses, credentials, infrastructure or exposed services so an attacker cannot rely indefinitely on one observation. Cloud systems shift workloads. Authentication systems expire tokens. Resilient networks reroute traffic. The comparison should not be taken literally, but the security economics are similar: force the adversary to rediscover, follow or cover a larger space.

The cost arrives immediately. Every moving part creates state that legitimate participants must share. Clocks drift. Updates are missed. Recovery paths are attacked. Old schedules remain in logs. A compromised hopping pattern turns mystery into timetable. The control plane that coordinates agility may become more valuable than the channel being protected.

The layered lesson is sharper:

• Use hopping or diversity to preserve availability.
• Use encryption to protect meaning.
• Use authentication to reject false commands.
• Use anti-replay measures so yesterday’s valid instruction does not steer today’s torpedo.
• Monitor whether the adversary has adapted.
• Design a safe failure mode for loss of synchronisation.
• Do not award one clever mechanism responsibility for every security property because the architecture diagram has run out of space.

And notice the false channels. Lamarr and Antheil’s decoy transmissions imposed uncertainty on the listener without confusing the intended receiver. Modern deception systems, canary credentials and honeypots use the same broad defensive instinct: make hostile observation more expensive, then ensure your own operators can still distinguish signal from theatre.

Security agility without synchronisation is outage. Synchronisation without protection is a map for the attacker. Resilience requires both movement and trust.

The artefact

The artefact is US Patent 2,292,387 itself.^[2] On its cover, beneath the neutral title ‘Secret Communication System’, are the names Hedy Kiesler Markey and George Antheil. Inside, a strip of perforated paper passes over a pneumatic reader. Condensers select radio frequencies. Matching motors turn. L and R keys steer a torpedo. Three channels lie.

It is an object poised between eras. The threat is electronic warfare. The answer is partly wireless, partly clockwork and partly music technology. The security property depends upon two distant machines maintaining a shared sequence while an adversary tries to learn or disrupt it. There is nothing quaint about that problem.

The patent also preserves the correction in unusually durable ink. Lamarr’s contribution was not discovered by generous historians and bestowed upon her afterwards. Her name was filed with the idea in 1941 and printed when the patent issued in 1942. Society spent the following decades looking at the same page and choosing to see the actress first, the inventor second, or not at all.

Read the names. Then read the mechanism.

A signal survived by refusing to remain where the enemy expected it. Hedy Lamarr’s place in the history of security should no longer depend upon the same trick.

Footnotes and Source notes

The technical and patent-history claims have been treated conservatively. In particular, the draft distinguishes the issued record-strip apparatus from the broader idea rejected during prosecution; it does not repeat the unsupported ‘classified and locked away’ story; and it describes modern wireless systems as relatives, not products of a single patent.

[1] Richard Rhodes, Hedy’s Folly (Doubleday, 2011). Detailed narrative history of Lamarr, Antheil and their collaboration; also the source cited by the Smithsonian finding aid for key recollections.
https://books.google.com/books/about/Hedy_s_Folly.html?id=8o0NnwEACAAJ

[2] US Patent 2,292,387, Hedy Kiesler Markey and George Antheil, ‘Secret Communication System’; filing and issue dates, paired records, 88 possible frequencies, simultaneous release, short rudder commands and false channels. https://patents.google.com/patent/US2292387A/en

[3] Smithsonian Institution, Guide to the Hedy Lamarr and George Antheil Invention Papers, NMAH.AC.1590; invention notebook, correspondence, biographies and the documented division of the collaboration.
https://sova.si.edu/record/nmah.ac.1590

[4] Willem Broertjes, US Patent 1,869,659, ‘Method of maintaining secrecy in the transmission of wireless telegraphic messages’; 1929 filing for synchronised alternation among secret frequencies and decoy frequencies.
https://patents.google.com/patent/US1869659A/en

[5] Martin, Blackwell and Vernam, US Patent 1,598,673, ‘Secrecy Communication System’; earlier multi-frequency communication prior art.
https://patents.google.com/patent/US1598673A/en

[6] David Rand Irvin, ‘Analysis of Hedy Lamarr’s Contribution to Spread-Spectrum Communication’ (2024); analysis of the USPTO prosecution record, cancelled claim 7, Samuel Stuart Mackeown’s assistance, publication status and later sonobuoy work.
https://researchers.one/articles/24.01.00001.pdf

[7] National Inventors Hall of Fame, ‘Hedy Lamarr’; named inventorship, biographical summary, 1997 recognition and 2014 induction.
https://www.invent.org/inductees/hedy-lamarr

[8] National Inventors Hall of Fame, ‘George Antheil’; musical and technical background, collaboration and 2014 induction.
https://www.invent.org/inductees/george-antheil

[9] Electronic Frontier Foundation, 1997 Pioneer Awards; contemporary record of the belated Pioneer Award to Lamarr and Antheil.
https://w2.eff.org/awards/pioneer/1997.php

[10] IEEE Standards Association, ‘Actress/Inventor Hedy Lamarr, and How Far Wireless Communication Has Come’; the original 802.11 frequency-hopping option and later Wi-Fi development.
https://standards.ieee.org/beyond-standards/hedy-lamarr/

[11] Bluetooth SIG, ‘How Bluetooth technology uses adaptive frequency hopping to overcome packet interference’; channel hopping, shared channel maps and avoidance of noisy channels in modern Bluetooth.
https://www.bluetooth.com/blog/how-bluetooth-technology-uses-adaptive-frequency-hopping-to-overcome-packet-interference/