Skip to content
problem genome

One of 592 unsolved problems in the Problem Genome, a collection written up for student design and engineering teams.

infrastructure · health

getting the fire alarm to the pillow

In Japan in 2006–2009, about 12,000 light-alarm units were shipped against 67.7 million home fire alarms, and only about 1 alarm in 10 could connect to a light or shaker; a free 2011–2013 scheme for deaf people on public assistance reached 3,613 homes, and in three Tokyo wards today's equipment grants mostly cover only severe certified hearing loss, while sleep tests abroad found shakers and a low-pitched sound wake most hard-of-hearing sleepers

Source: FDMA (Fire and Disaster Management Agency) study group Open ↗source: primaryLast updated: 2026-11-17

Problem statement

Many fire deaths at home happen at night, and a smoke alarm is only useful if it wakes the sleeper. Deaf and hard-of-hearing people need the alarm to reach them another way: a flashing light or a shaker under the pillow or mattress. In Japan, every home has had to have a smoke alarm since June 2011. But the law sets technical rules only for the sound. For anything else it says only that the alarm must be able to tell the people inside "effectively" (our translation). A 2011 study group of Japan's Fire and Disaster Management Agency (FDMA) found that of about 67.7 million home alarms shipped from 2006 to 2009, about 12,000 were light alarms (about 0.017%). Only about 6.7 million (about 10%) had the terminal needed to connect a light or shaker. Connection between makers was not assured, because no common standard existed (though the report adds that a check of products on sale found that devices from different makers could be connected). The same report found that alarms with a terminal cost much more than ones without. A free scheme for deaf people on public assistance then installed alarms with lights or shakers in 3,613 homes, and closed in 2013. Linked systems do exist: the scheme supplied alarms joined by wire or radio to a light or vibration unit, with a pillow shaker in the bedrooms of deafblind people. Units that listen are sold in Japan too: one maker's sensor learns the sound of the home's own alarms and sends text and vibration to a wristwatch, at a reference price of ¥87,400 before tax on a Tokyo welfare-equipment listing. And today some city welfare offices pay toward a fire alarm or a light or vibration signalling device through the disability "daily living equipment" grant. But in the three Tokyo wards whose pages were read (Meguro, Edogawa, Minato), the fire alarm grant is limited to grade 1–2 (Edogawa: grade 2 hearing) certificates and to households judged to find it very hard (Edogawa: hard) to notice a fire and escape. The indoor signalling device in Meguro and Edogawa needs a grade 2 hearing certificate, age 18 or over, and a household of hearing-impaired people only (or similar). Only Minato lists a vibration-and-light item open to any certified hearing disability, with a 10% co-payment (free for people on public assistance or exempt from resident tax). Japan has about 450,000 hearing certificate holders but about 6 million people are said to have hearing loss of 40 dB or more, so most people with hearing loss hold no certificate and fall outside grants like these. Renters need the landlord's consent before an alarm is fitted (a rule of the 2011 scheme). Sleep studies abroad show that strobe lights wake few hard-of-hearing people. Shakers wake most; in the Australian study a low-pitched sound woke slightly more people with mild to moderate hearing loss, but at the standard level neither woke everyone there. The unsolved problem is getting a waking signal from whatever alarm a home already has to the bed of a deaf or hard-of-hearing person, reliably and cheaply, without replacing the alarms or needing a certificate or a landlord's permission.

Why this matters

The FDMA report cites about 450,000 people in Japan holding a disability certificate for hearing or balance in 2008, and says about 6 million are said to have hearing loss of 40 dB or more, counting older people. A 2018 Japanese engineering paper notes, citing FDMA data, that people aged 65 or over made up 69.9% of home-fire deaths, that being "too late to escape" accounted for 49.7%, and that many deaths happen between 2 and 6 a.m. (our translation). When 277 deaf and hard-of-hearing people were asked in 2010 what alarms would help in a bedroom (more than one answer allowed), 63.5% chose a shaker and 49.8% a flashing light. Asked what would most help them install non-sound alarms at home, the largest share named a local-government subsidy (34.7%), then publicity (24.2%) and products that reflect users' needs (17.3%).

What’s been tried and why it hasn’t worked

(1) Lights. Strobes are the usual non-sound alarm in the US; where one is installed in a sleeping area, the US code (NFPA 72) sets its brightness at 110 or 177 candela. In a sleep study in Australia (38 adults with hearing loss of 25–70 dBA and 32 young adults given alcohol, signals sounded in deep sleep), "three quarters of both groups slept through the strobe light" even at above the brightness required by the US standard (the journal version reports 27% of the hard-of-hearing group woke to it). A US sleep-laboratory study presented in 2005 (111 people: 32 profoundly deaf, 45 hard of hearing, 34 hearing) found the strobe woke sleepers on 57% of presentations for deaf, 34% for hard-of-hearing and 32% for hearing people. (2) A low-pitched sound. In the Australian study, for people with some hearing, the strongest positive result was a sound, not a shaker. In the Australian study a 520 Hz square-wave sound at 75 dBA woke 92% of the hard-of-hearing group, against 56% for the usual high-pitched (about 3,100 Hz) alarm. The authors conclude it "should be adopted across the whole population as the normal smoke alarm signal" and is "an appropriate emergency alerting device" for hearing loss in the range tested (25–70 dB). It does not help deaf people: in the US study a low-frequency (450 Hz) alarm woke deaf sleepers on only 11% of presentations (92% for hard of hearing). (3) Shakers. In the US study, an intermittent bed shaker woke sleepers on 100% of presentations in every hearing group and a continuous one on 93% for deaf sleepers. In the Australian study the shakers were "much more effective than the strobe": only 3% (pillow) and 11% (bed) slept through every intensity level, and 80–84% woke at the intensity as purchased. But at that intensity they did not wake 17–20% of the hard-of-hearing participants, those aged 60 or over were less likely to wake to the bed shaker, and both shakers were less effective than the 520 Hz sound (the journal version found this gap not statistically significant). The Australian authors call for study with people with greater hearing loss, including deaf people. (4) Linking the shaker to the alarm. In Japan the standard home alarm and the light or shaker are separate devices joined by wire or radio, and in 2006–2009 only about 10% of alarms had the terminal. The FDMA group proposed fitting a terminal to every home alarm, possibly as a manufacturing requirement, and a common connection specification. It also listed unsettled technical requirements for shakers: "vibration strength, frequency, rhythm, reliability, measures against falling off" (our translation). A US bedside unit sold for people with hearing loss listens for the alarm and drives a bed shaker. Its manual says it detects only alarms that emit the UL 217 or ULC S531 "T3" signal, and it was recalled in Canada in December 2018 (Health Canada notice) because the display and optional bed shaker could stop working; its manual also says it warns when the bed shaker is unplugged, when AC power is lost and when its batteries are low or missing (the shaker then twitches once a minute). A second US listener (SafeAwake) also responds only to US T3 smoke alarms; it sounds a low-frequency tone and drives a pulsed bed shaker, has a weak-battery check, and is listed by a US retailer at $399.95. Learning the alarm is not new either: a US patent with 2004 priority (US7522035B2, now listed as expired) describes a bedside unit whose alarm sound "can be 'learned' by activating the alarm for setup purposes", with several sound templates at once and a bed-shaking device among its waking outputs, and in the UK, Geofire's own page says its Agrippa pillow fire alarm, a battery unit with a pad under the pillow, uses "listen and learn" technology that "listens for the unique sound of a specific fire alarm". In Japan, the Ronk "知らせる君" sensor learns "あらゆる警報音" (any alarm sound) and sends text and vibration to a waterproof wristwatch or a light receiver; its maker says that because it detects only learned sounds it "誤報しません" (makes no false alarms), a claim we found no test of. Its wristwatch runs about 120 hours on standby between charges, and its page names no pillow or bed shaker. A Tokyo Shinyu wristwatch system with a microphone sensor (reference price ¥110,000) likewise reports a sounding fire alarm by text and strong vibration. The same maker sells an add-on radio transmitter (AT-0101) that is screwed to a home fire alarm's relay terminal and passes the alarm to its wristwatch and light receivers; its page says it works only with alarms that have that terminal (「移報接点付のみ」, in our translation "relay-terminal type only") and names no bed or pillow shaker. In 2018, engineers at Okayama Prefectural University built a microphone-and-filter circuit that drives a pillow vibrator when it hears a Japanese home alarm (buzzer peak at about 3,293 Hz). But they tested it with one alarm model and ten young men with normal hearing, awake, wearing earplugs, so it says nothing yet about waking. (5) Paying for it. In 2011 the FDMA funded a free supply of alarms with a light or shaker for households on public assistance that include a person with a hearing disability certificate, or one whose certificate application is pending (budget ¥320 million; about 14,000 people assumed eligible). The notice gave the reason: ordinary alarms only sound, and light alarms sell at high prices. In rented homes the alarm could be fitted only after the landlord agreed. A notice in July 2012 says the budget was planned for about 4,000 households and applications had reached about 2,000. It asked municipalities to strengthen outreach, for example by switching from direct mail to home visits. The scheme closed in March 2013 with 4,317 households applying and 3,613 installed. The 704 households that applied but got nothing were cancelled for reasons such as long hospital stays.

What would unlock progress

A link from the alarm to the bed that does not depend on the alarm having a terminal, and that is proven to wake people. The listening approach avoids replacing the alarms already on Japanese ceilings, which is why the Okayama team chose it. But it must recognise that home's own alarm and not other sounds. The US bedside listeners accept only the North American T3 pattern. The Japanese learning sensor accepts any alarm it is taught, but it reports to a wristwatch that must be worn and recharged, has a reference price equal to the Meguro and Edogawa grant ceiling for an indoor signalling device (¥87,400), and we found no published test of how often it misses or false-triggers. The vibration must stay strong through the pillow or mattress. In the Okayama bench test, a low-rebound urethane pillow cut the peak vibration to 1.25 m/s², against 6.53 and 8.27 m/s² for feather and polyethylene pillows. And the device must warn when it is unplugged or flat, since the recalled unit's shaker could stop working. The FDMA group's own suggestion, made to help such devices spread, was to combine the alerting device with things used every day, such as a doorbell or an alarm clock; a unit in daily use is also more likely to stay plugged in and be trusted. Because the Australian study found a 520 Hz square-wave sound woke at least as many hard-of-hearing sleepers as the shakers did, the bedside unit could also sound that low tone for people with some hearing, and keep the shaker for deaf people. To reach the people the grants leave out, it must cost little, need no certificate, and fit without drilling or rewiring, so that renters do not need a landlord's consent.

How we’d know it’s being solved

(a) In sleep-stage-monitored tests, the device wakes deaf and hard-of-hearing sleepers, including those aged 60 or over, at least as well as an intermittent bed shaker did in the US study (100% of presentations), and, for hard-of-hearing sleepers, at least as well as the 520 Hz sound in the Australian study (92% at 75 dBA). That means doing better than the 17–20% miss rate for shakers as purchased in the Australian study. (b) It triggers on the household's existing alarm in every room tested and rarely on other household sounds (a false-alarm count per night). (c) Vibration measured through common pillow and mattress types stays above a set level. (d) A missing power supply or an unplugged shaker shows itself without anyone checking. (e) In Japan, many more homes of deaf and hard-of-hearing people with a working light or shaker than the 3,613 reached by the 2011–2013 scheme.

Where to start

Check what is sold first. Much of this exists. Japan has wired and radio-linked pillow shakers and light units, but they need an alarm with a link terminal (about 1 in 10 alarms shipped in 2006–2009), and some city grants pay for them only for people with severe certified hearing loss. Listening units that hear the alarm already on the ceiling and drive a shaker are also sold. There are two US bedside units with a bed shaker (both accept only the North American T3 pattern; one was recalled in 2018 because its shaker could stop working), and there are at least two UK bedside units with an under-pillow pad. One of the UK units, by its maker's manual, responds to the smoke alarm supplied with it and lists two other alarm models it has been tested with. The other is said to trigger on an alarm sound above a set loudness that lasts a set time. Japan already sells a sensor that learns the home's own alarm sounds, but it alerts through a wristwatch that must be worn and recharged, its page names no bed or pillow shaker, and its no-false-alarm claim is the maker's own. Learning the alarm by sounding it once at setup is also described in a 2004-priority US patent now listed as expired. The 2018 Okayama prototype is one more. So a team should buy or read up on these before building, and aim only at what none of them shows.

Facility-free door (make a thing): A plug-in or battery bedside unit that needs no wiring, no new alarm and no landlord's consent. Like the Japanese learning sensor, it learns the sound of the alarms already in that home, whatever their pattern (a Japanese home alarm's buzzer peaks at about 3,300 Hz; the US listeners accept only the T3 pattern), and ignores other sounds. Parts of this are already sold: in the UK, Geofire's pillow fire alarm learns the sound of one specific fire alarm and vibrates a pad under the pillow, and the US Lifetone unit warns when its shaker is unplugged, its power is lost or its batteries are low, but hears only T3 alarms. What is still open, on the evidence here, is one unit that does all of this for the alarms sold in Japan: it drives a pillow or bed shaker strong enough to be felt through the bedding, so nothing has to be worn or charged at night, and it reports a measured false-trigger and missed-alarm count rather than a claim. For people with some hearing it can also sound the 520 Hz low tone. And it warns when it is unplugged, flat or its shaker is disconnected. Build it from a microcontroller, microphone, vibration motor and battery, in a housing that people can set up by themselves. Bench-test three things. First, recognition: sound each alarm with its own test button (never with real smoke or fire), using several models bought or recorded, and play a library of household sounds (TV, kettle, phone alarms, the listening units' own trigger rules as rival designs) to get a false-trigger count. Second, vibration through different pillows and mattress toppers, measured with a phone accelerometer, repeating the Okayama method. Third, the unplugged, low-battery and loose-shaker warnings. No partner or restricted data is needed.

User door (awake volunteers only): Deaf and hard-of-hearing adults, including older people with hearing loss, can be reached through a campus disability office, a local deaf club or society, or a senior centre. Run supervised daytime sessions, awake and lying down, on how noticeable different vibration rhythms and strengths are through a pillow, and on whether people can set up, test and trust the unit themselves. Students never test whether the unit wakes anyone, and no session involves sleep. Sessions with people need the campus's ethics review, and deaf participants need a sign-language interpreter or written materials agreed with them.

Design-the-trial door: Whether a signal wakes a sleeping deaf person is a sleep laboratory's job, under its own ethics review. The team writes the protocol only: sleep-stage monitoring and sounding in deep sleep, as in the Australian and US studies, with profoundly deaf participants, whom the Australian study did not cover, and older participants, and the pass mark in "How We'd Know".

What carries over to Japan: The listening, false-trigger and vibration-through-bedding tests carry over directly. The alarm sounds may not: the US listeners accept only the North American T3 pattern, and the sources here do not describe the pattern Japanese home alarms sound, so a unit tuned on US or European alarms must be re-tested on Japanese home alarms (bought or recorded) before it means anything there. The renter and no-certificate goals are the reason for the no-wiring design in any country. Relevant skills: embedded electronics, audio signal processing, industrial design, human factors.

Genes

Related problems

Stuck the same way, in other fields.

field: health & medicinethe sleeve over the back burnerIn Japan clothing fires kill about 100 people a year, mostly older adults, and in Tokyo homes the stove starts about 80% of clothing fires that hurt someone; sleeve-sensing stoves now exist, but for stoves already in homes the official advice is a flame-resistant arm cover, which few older people in a Tokyo-area survey knew of or owned, and a warning video shown to students raised perceived risk but not intentionsbuilt without the people who'd live with itsource: primaryfield: health & medicineside effects that arrive years laterBreast implant long-term safety signals emerge years after implantation — and passive surveillance cannot measure thembuilt without the people who'd live with itsource: primaryfield: society & learningwhy the lecture survives the evidenceEvidence-based teaching raises grades, but most STEM faculty haven't adopted it, and student pushback discourages those who trybuilt without the people who'd live with itsource: primaryfield: health & medicineforty-two degrees feels like thirty-nineJapan advises older people to warm the changing room, keep bath water at 41 °C or less, soak no more than ten minutes and check with a thermometer and timer, but in laboratory tests older bathers felt no difference between 39 °C and 42 °C water, surveys found bathroom heating in only 3% to 24% of homes, and the water-heater remotes that watch the bath, whether they alert family or tell the bather when to get out, are described only for particular heater modelsbuilt without the people who'd live with itsource: primary

Sources

  • FDMA (Fire and Disaster Management Agency) study group, "ユニバーサルデザインを踏まえた火災警報設備等の導入・普及のあり方に関する報告書 ~聴覚障がい者に対応した火災警報設備の検討~(抄)" (聴覚障がい者に対応した火災警報設備等のあり方に関する検討会(総務省消防庁), 平成23年3月; in our translation, "Report on introducing and spreading fire alarm equipment based on universal design: fire alarm equipment for people with hearing impairment (excerpt)," March 2011), accessed 2026-11-17. Link ↗
  • 消防庁予防課長「『聴覚障がい者対応型住宅用火災警報器普及支援事業』の実施について」消防予第366号, 平成23年9月29日 (FDMA notice launching the support scheme), accessed 2026-11-17. Link ↗
  • 消防庁「住宅用火災警報器の普及における消防庁の取組について」資料3 (FDMA slides reporting the scheme's results), accessed 2026-11-17. Link ↗
  • Thomas I, Bruck D. "Strobe lights, pillow shakers and bed shakers as smoke alarm signals." Ninth International Symposium on Fire Safety Science, 21-26 September 2008, Karlsruhe, Germany (Victoria University repository record; PDF), accessed 2026-11-17. Link ↗ Link 2 ↗
  • 株式会社東京信友「[AT-0101] 住宅用火災警報器補助装置 外付無線送信器」 (Tokyo Shinyu Co, Ltd, maker's product page for an external wireless transmitter for home fire alarms), accessed 2026-11-17. Link ↗
  • Bruck D, Thomas IR. "Smoke Alarms for Sleeping Adults Who are Hard-of-Hearing: Comparison of Auditory, Visual, and Tactile Signals." Ear & Hearing 2009;30(1):73-80, doi:10.1097/aud.0b013e3181906f89 (full text read through library access), accessed 2026-11-17. Link ↗
  • 目黒区「聴覚・言語機能障害対象の日常生活用具」 (Meguro City ward page on daily living equipment for hearing and speech impairment, updated 2026-07-09), accessed 2026-11-17. Link ↗
  • 江戸川区「日常生活用具給付対象用具一覧(聴覚障害、音声・言語障害、そしゃく機能障害)」 (Edogawa City ward page listing daily living equipment grants for hearing, speech and chewing impairments, updated 2024-09-19), accessed 2026-11-17. Link ↗
  • 港区「障害者(児)日常生活用具の給付など」 (Minato City ward page on daily living equipment grants for disabled people and children, updated 2026-06-22), accessed 2026-11-17. Link ↗
  • Geofire, "Agrippa Pillow Fire Alarm" (maker's product page), accessed 2026-11-17. Link ↗
  • Geofire, "Deaf Alert Home Pillow Alarm" (maker's product page), accessed 2026-11-17. Link ↗
  • Geofire, "Agrippa Deaf Alert User Manual" (maker's manual, linked from the Deaf Alert page), accessed 2026-11-17. Link ↗
  • Fireco, "Deafgard | The alarm for the deaf & hard of hearing" (maker's product page), accessed 2026-11-17. Link ↗
  • 消防庁予防課長「『聴覚障がい者対応型住宅用火災警報器普及支援事業』に係る支援対象者への周知及び申請支援の強化について」消防予第264号, 平成24年7月4日 (FDMA notice on strengthening outreach and application support for the scheme, 4 July 2012), accessed 2026-11-17. Link ↗
  • Tsutsukawa T, Yoshida T, Hokari M (筒川拓磨, 吉田智哉, 穂苅真樹). "Development of residential fire alarm sound transmission system for the elderly with hearing loss" (「聴力低下を伴う高齢者のための住宅用火災警報器音伝達システムの開発」). The Proceedings of the Symposium on sports and human dynamics 2018, D-26, doi:10.1299/jsmeshd.2018.D-26 (full text in Japanese), accessed 2026-11-17. Link ↗
  • Du Bois J, Ashley E, Klassen M, Roby R. "Waking Effectiveness of Audible, Visual and Vibratory Emergency Alarms on People of all Hearing Abilities." Accessible Emergency Notification and Communication: State of the Science Conference, Gallaudet University, Washington, DC, November 2-3, 2005 (conference paper), accessed 2026-11-17. Link ↗
  • Lifetone, "LIFETONE HL Bedside Fire Alarm & Clock: User's Manual, Model HLAC151" (maker's manual), accessed 2026-11-17. Link ↗
  • Health Canada, "Lifetone Technology, Inc. recalls Lifetone HLAC151 Bedside Fire Alarm and Clock" (recall notice, 19 December 2018), accessed 2026-11-17. Link ↗
  • Ronk株式会社「聴覚サポート機器」 (maker's page for the indoor signalling device 知らせる君), accessed 2026-11-17. Link ↗
  • 福ナビ「自立生活支援用具|福祉用具|福ナビ」 (Tokyo welfare-equipment listing with reference prices), accessed 2026-11-17. Link ↗
  • MaxiAids, "SafeAwake Smoke Alarm Aid with Bed Shaker" (retailer page), accessed 2026-11-17. Link ↗
  • Albert DE (InnovAlarm Corp). "Enhanced bedside sound monitoring and alarm response method, system and device." US patent US7522035B2, priority 2004-07-23, filed 2006-09-18, granted 2009-04-21, accessed 2026-11-17. Link ↗

Also cited in the brief

  • Tasmania Fire Service, "Deaf and hard of hearing alarms" Link ↗
  • Bellman & Symfon, "Fire & Rescue Service UK" Link ↗
  • National Seniors Australia, "Why many smoke alarms don't wake older Australians," 27 March 2026, sponsored story. Link ↗
Revision notes

This brief gets checked against its sources and contains the following revisions. Future revisions get added here.

Worth knowing

  • The figure that about 1 home fire alarm in 10 could connect to a light or shaker is for alarms shipped in 2006 to 2009; no newer figure was found.
  • What today's equipment grants cover comes from three Tokyo ward pages, not a survey of Japan.

Full check

Sources read

  • Japan Fire and Disaster Management Agency, study-group report on fire alarms for people with hearing loss (excerpt, 2011): full text (read in the original Japanese)
  • Fire and Disaster Management Agency, notice launching the support scheme (2011), with its application rules: full text (read in the original Japanese)
  • Fire and Disaster Management Agency, follow-up notice on the scheme (2012): full text (read in the original Japanese)
  • Fire and Disaster Management Agency, slides reporting the scheme's results: full text (read in the original Japanese)
  • Okayama Prefectural University, alarm-listening circuit driving a pillow vibrator (2018): full text (read in the original Japanese)
  • Meguro, Edogawa and Minato ward pages on disability equipment grants: full text (read in the original Japanese)
  • Ronk, maker's page for its sound-learning alert system: full text
  • Tokyo welfare-equipment listing (Fukunavi): full text
  • Tokyo Shinyu, maker's top page: full text (top page only; its product pages did not download at this check)
  • Thomas and Bruck, strobe lights and shakers as smoke alarm signals (symposium paper, 2008): full text
  • Bruck and Thomas, smoke alarms for sleeping adults who are hard of hearing (Ear and Hearing, 2009): full text (its tables are images)
  • Smedberg, Ronchi and Hutchison, review of alarm technologies for deaf and hard-of-hearing sleepers (Fire Technology, 2022): full text
  • Du Bois and others, waking study with deaf and hard-of-hearing sleepers (Gallaudet, 2005): full text
  • Lifetone bedside unit manual and the Health Canada recall notice: full text
  • MaxiAids retailer listing for the SafeAwake unit: full text
  • US patent 7,522,035, bedside sound monitoring and alarm response (patent record): full text
  • Geofire, Fireco, Tasmania Fire Service, Bellman and National Seniors Australia pages: full text
  • Safelincs retailer listing for one UK listening unit: could not be opened

What changed

  • Not found in the sources, replaced with what they say: the claim that most hard-of-hearing older people fall outside the grants; the report gives about 450,000 hearing certificate holders against about 6 million people with hearing loss, so the page now says most people with hearing loss hold no certificate; who the old scheme covered besides certificate holders; the notice's rules define this as people whose certificate application was pending, and the page now says so; the lines on what pattern a Japanese home alarm sounds; no source describes it, so the page now says the US listening units accept only one US pattern and that any listener must be re-tested on Japanese alarms.
  • Stated more strongly than the source: the grant rule was said to cover households judged unable to notice a fire and escape; the ward pages say households that find this very hard (one ward: hard); the government study group's idea of combining the alert with everyday devices such as an intercom or alarm clock is now given with the report's own purpose, helping such devices spread; the point about being plugged in and trusted is marked as the page's own reasoning.
  • Contradicted by the source, corrected: one UK listening unit was said to be tested only with the alarm sold with it; its maker's manual lists that alarm and two other tested models.
  • Credit: the claim that a UK unit learns the sound of a particular fire alarm now belongs to the maker's pillow fire alarm, not to its listening unit, as the maker's pages say.
  • Title: the full title now says the finding on today's equipment grants comes from three Tokyo wards, not Japan's towns and cities in general (display title unchanged).
  • Entry points: vibrating pillows are now called pillow shakers, the device the government papers name; the make-a-thing door now says which parts are already sold (a UK unit that learns one alarm and drives a pad under the pillow; a US unit that warns of an unplugged shaker, lost power and low batteries) and that what is still open is one unit that does all of this for alarms sold in Japan, with nothing worn and measured counts of missed and false alarms.
  • Other: the US bedside unit's warnings for an unplugged shaker, lost power and low batteries are now described from its manual.

Left open

  • the trigger rule for one UK listening unit (the retailer listing that gives it could not be opened)

Separate check of specific claims: A maker's add-on radio transmitter that passes a home fire alarm to wristwatch and light receivers was added, with the note that it works only with alarms that have a relay terminal, and citations for the sources the page uses were added to its source list.

Earlier full check: Thirteen statements were brought in line with the sources: the full title's 'but not all' was dropped because one study's bed shaker woke every sleeper, 'most' night-time fire deaths became 'many', the low-pitched sound result was tied to the one study that found it and its gap over shakers marked as not statistically significant, the US low-frequency alarm was named as a different tone, and smaller wording fixes about the recall, a survey question and the old scheme.

Separate check of specific claims: The page now notes that a sound-learning alert system and a wristwatch system are already sold in Japan, a second US bedside unit, and an expired patent on learning an alarm's sound; the page and the make-a-thing door now start from what these leave open.

Separate check of specific claims: The entry points were rewritten into four doors with safety lines (alarms sounded only with their test button, never smoke or fire; students run no sleep or waking tests), and two UK bedside listening units with pillow pads were noted.

Earlier full check: The display title changed from 'The Alarm That Cannot Reach the Pillow' to 'Getting the Fire Alarm to the Pillow', because linked pillow shakers and light units exist and some wards fund them; the full title was rewritten to match, today's grants in three Tokyo wards and the cost and landlord-consent rules were added, and the sleep studies' positive results are now stated as the papers state them.