Build a Smart Home Fall Detection System for Elderly Parents

A fall detection setup for an elderly parent has to survive the awkward cases, not just the product demo. The bathroom is exactly where a camera feels wrong. The bedroom is where a pendant may be sitting on the nightstand. The front walk is where a room sensor has no idea what happened. Wi-Fi can be flaky, a phone can be in another room, and the person who needs help is not the person who should have to debug the system.

That is why the strongest consumer-grade answer in 2026 is not one “best” device. It is a layered system: mmWave radar sensors in the rooms where falls are most likely to become private and undiscovered, plus a wearable fall detector for mobile and outdoor coverage, all tied to an alert chain that turns detection into action.

The urgency is real without needing to exaggerate it. The CDC says about 1 in 4 adults age 65 and older falls each year, falling once doubles the chance of falling again, and roughly 37% of falls lead to an injury that requires medical treatment or restricts activity.[1] The point of these sensors is not to make a home risk-free. It is to shorten the time between “something happened” and “the right person knows what to do.”

Start With Coverage Gaps, Not Device Categories

A fall system has four jobs: notice a likely fall, decide whether it is urgent enough to escalate, notify someone who can respond, and keep working when the first assumption fails. A single device rarely covers all four well.

This is also where privacy enters the design. Cameras may be useful in some eldercare contexts, but they are a poor default for bathrooms and bedrooms. mmWave radar is attractive because it can sense motion, presence, distance, and posture-like patterns without producing an image of the person. That does not make it medically perfect. It makes it a better starting point for rooms where dignity matters.

The Hybrid Architecture

A practical build uses one or two room sensors, one wearable, and one automation hub. The room sensors cover high-risk indoor areas. The wearable covers movement between rooms, outdoor activity, errands, and the moments when the parent is not inside a monitored zone. The hub makes the response predictable instead of leaving each device inside its own app.

• Room layer: mmWave radar in the bathroom, bedroom, or another high-risk room where a camera is inappropriate.
• Mobile layer: smartwatch, pendant, or medical alert wearable with fall detection.
• Hub layer: Home Assistant, Apple Home, or an Alexa-based care setup depending on the devices chosen.
• Response layer: lights, smart speaker audio, caregiver push notification, optional door unlock, and a documented call sequence.

The global fall detection systems market was valued at $516.9 million in 2025 and is projected to grow at a 7.8% compound annual growth rate through 2033.[2] That number mainly says the category is mature enough to have real choices. It does not tell you whether a sensor will notice your father sliding beside a bed at 2 a.m., or whether your sister will hear the alert.

Choose the Room Sensors

For room coverage, mmWave radar has a useful advantage over basic motion sensors. A passive infrared motion sensor can miss a person who is still. A contact sensor only knows whether a door opened. mmWave radar can detect subtle movement and presence, which is why it has become interesting for room-level fall detection and aging-in-place automations.

The first placement decision is not “how many sensors can I buy?” It is “where would a delayed discovery be worst?” For many homes, that means the bathroom first and the bedroom second. A living room sensor may be useful, but it usually comes after private, high-risk rooms unless the parent spends most of the day there.

Aqara FP2: convenient, but do not overread the feature

The Aqara FP2 is a 60GHz presence sensor sold around the $80 range, with up to 30 zones, stated coverage up to 40 square meters, and support paths that matter in normal homes: Apple Home, Alexa, and Home Assistant.[3] For a family that already uses Apple Home or wants a relatively polished device, it is the easiest mmWave option to understand.

The caution is just as important: the FP2 fall detection capability is a firmware-updated feature of a presence sensor.[3] That does not make it useless. It means it should be treated as one layer in a response system, not as a medical alert service wearing a smart home costume.

Seeed MR60FDA2: cheaper, deeper, more responsibility

The Seeed MR60FDA2 kit sits closer to the maker side: around $25, pre-flashed for ESPHome, and designed for direct Home Assistant integration through ESPHome YAML.[4] If you are comfortable editing configuration files and validating sensor states yourself, this route can be powerful and inexpensive.

The tradeoff is that low parts cost is not the same as low system cost. A DIY setup excludes the Home Assistant server, the time spent configuring entities, and the ongoing responsibility to notice when an integration breaks after an update. A bathroom sensor that silently disappears from the dashboard is not a bargain.

Care-oriented radar: Vayyar Care and Essence MDsense

Vayyar Care and Essence MDsense sit closer to eldercare than hobby automation. Vayyar Care uses 4D radar and has been associated with Alexa Together integration, with pricing around $200 plus subscription assumptions.[5] Essence MDsense is a battery-operated radar fall detection device positioned for commercial care settings.[6]

These are not interchangeable with an $80 room presence sensor. A care-oriented product may be easier to justify when monitoring workflow matters more than tinkering flexibility. A maker board may be easier to justify when the household already runs Home Assistant and someone will maintain it. The wrong choice is the one nobody will own after setup weekend.

Place Radar Where It Can Actually See the Fall Pattern

A radar sensor is not a magic ceiling tile. It has a field of view, mounting assumptions, reflection behavior, and blind spots. Before writing any automation, walk the room and identify the places a person could fall and remain hidden: beside the toilet, outside the shower, between the bed and wall, at the foot of the bed, or near a favorite chair.

• Bathroom: prioritize the toilet and shower exit area, while avoiding direct water exposure and unstable temporary mounting.
• Bedroom: cover the bed edge and common walking path, not just the center of the room.
• Living room: use zones carefully if pets, fans, recliners, or multiple occupants create confusing movement.
• Hallway: consider this only if the parent has a history of falls during transfers or nighttime walking.

Multi-person homes are harder. So are homes with pets, moving curtains, oscillating fans, cluttered furniture, and frequent visitors. This is where lab sensitivity numbers stop being comforting and start becoming calibration hints. The only test that matters is whether the sensor, in that room, can trigger the right automation without burying caregivers in false alarms.

Choose the Wearable for the Gaps Radar Cannot Cover

The wearable is not redundant. It covers the driveway, garden, garage, mailbox, grocery store, and the hallway if the parent leaves the radar zone. It also gives the parent a manual help button, which matters because not every emergency is a clean fall.

Apple Watch fall detection has shown about 93% accuracy in controlled study conditions, but that figure should travel with the obvious caveat: a watch on a charger detects nothing on a wrist.[7] For a parent who already wears one every day, it may be the least stigmatizing option. For a parent who removes watches at home or forgets to charge devices, a pendant with monitoring may be more realistic.

Bay Alarm Medical and Medical Guardian are more traditional medical alert choices. SafeHome.org’s June 2026 testing gives Bay Alarm Medical a 9.3/10 score, lists pendant range above 1,000 feet, and describes pricing around $30 per month plus $10 for fall detection.[8] Medical Guardian is listed with a 1,400-foot range, fall detection that distinguishes soft and hard impacts, and plans around $25–$60 per month plus a $10 fall detection add-on.[8]

The choice is less about elegance than compliance. If the parent refuses pendants but likes a watch, choose the watch. If the parent will not charge a watch but accepts a simple medical pendant, choose the pendant. If the parent refuses every wearable, build the best multi-room radar setup you can, but write down the gap in plain language: there is no outdoor fall coverage.

Build the Automation Around Response, Not Detection

A sensor event is not help. Help begins when the right person is notified, can verify what happened, and knows whether to call, visit, unlock a door, or request emergency services. The automation should make those steps fast and boring.

Home Assistant route

Home Assistant is the deepest route if you want radar entities, wearable integrations, conditional logic, and local-first automations. It is also the route that most clearly assigns maintenance to you. That is fine if you are willing to own it.

alias: Elderly fall detected - bathroom
mode: single
trigger:
  - platform: state
    entity_id: binary_sensor.bathroom_fall_detected
    to: "on"
condition:
  - condition: time
    after: "00:00:00"
    before: "23:59:59"
action:
  - service: light.turn_on
    target:
      entity_id:
        - light.bathroom
        - light.hallway
    data:
      brightness_pct: 100
  - service: media_player.volume_set
    target:
      entity_id: media_player.bedroom_speaker
    data:
      volume_level: 0.75
  - service: tts.speak
    target:
      entity_id: tts.home_assistant_cloud
    data:
      media_player_entity_id: media_player.bedroom_speaker
      message: "A fall may have been detected in the bathroom. Are you okay?"
  - service: notify.mobile_app_caregiver_phone
    data:
      title: "Possible fall detected"
      message: "Bathroom radar reported a possible fall. Check camera-free status, call parent, then follow the response plan."
      data:
        push:
          sound: "critical"
  - delay: "00:02:00"
  - condition: state
    entity_id: input_boolean.fall_alert_acknowledged
    state: "off"
  - service: notify.mobile_app_backup_caregiver_phone
    data:
      title: "Unacknowledged fall alert"
      message: "No caregiver acknowledged the bathroom fall alert within 2 minutes."

That YAML is a pattern, not a finished medical system. The important parts are the acknowledgement path and escalation delay. A push notification that vanishes into a sleeping phone is a weak endpoint. Use critical alerts where available, a second caregiver, or a monitored wearable service if nobody in the family can reliably respond.

For an ESPHome radar board, expose a clear binary sensor or event entity instead of burying logic in a dashboard card. For an Aqara FP2, keep zone naming boring: bathroom_fall_zone is better than “Dad Safety Magic.” Six months later, clarity beats charm.

Apple Home route

Apple Home makes sense when the family already uses iPhones, HomePods, and HomeKit-compatible accessories. The Aqara FP2’s Apple Home support makes it relevant here, especially for presence and room-based automations.[3] The limitation is that Apple Home is less flexible than Home Assistant for complex escalation, acknowledgement states, and custom fall logic.

1. Create a room automation triggered by the fall or relevant sensor state exposed to Apple Home.
2. Turn on bathroom, bedroom, and hallway lights immediately.
3. Use HomePod audio where practical, but do not rely on audio alone if the parent has hearing loss or the speaker is in another room.
4. Send caregiver alerts through the device app, a companion app, or a parallel wearable service when Apple Home cannot provide the escalation logic you need.

Apple Home can be the comfortable control layer, but it should not hide the alerting question. If the device app sends the only fall notification, test whether it arrives during Focus modes, overnight, and when the phone is locked.

Alexa Together route

An Alexa-based route is strongest when the parent already talks to Echo speakers and the family wants a care workflow more than a custom automation lab. Vayyar Care’s Alexa Together integration is relevant for that style of setup.[5] The benefit is less tinkering. The tradeoff is that you accept the service model, supported devices, subscription assumptions, and less local control.

If you choose this path, place the radar device where it solves the specific privacy problem, then use Alexa routines and care alerts for escalation. Do not assume an Echo announcement in the room is enough. The caregiver still needs the external alert.

Cost: One-Time Sensors vs. Three Years of Monitoring

The cheapest setup on paper is usually not the cheapest setup to trust. Radar sensors can be one-time purchases. Monitored wearables become an operating cost. DIY builds save subscription fees but shift installation, testing, and support to the family.

For a family choosing between a monitored wearable and a DIY-heavy system, the real question is who is on duty. A subscription can buy a monitoring workflow. A local Home Assistant setup can buy privacy and control. Neither automatically buys a person who answers at 2 a.m.

Accuracy Numbers Are Calibration Signals, Not Guarantees

Controlled studies are useful, but they answer a narrower question than families usually ask. A study may show whether a sensor can classify simulated falls under test conditions. A home system has to deal with bath mats, pets, visitors, furniture changes, firmware updates, and a caregiver who may be asleep.

Research on radar, thermal, and IoT fall detection approaches has reported sensitivity in the 90–98% range under controlled conditions. A 2023 MDPI Sustainability IoT study reported up to 98% sensitivity and 99% specificity.[9] Wearable fall detection has also been reported around 93% accuracy in controlled study settings.[7] Those are encouraging engineering signals. They are not promises about a specific parent’s bathroom.

Treat every impressive number as a reason to test, not a reason to relax. The practical question is not “does this technology ever work?” It is “does this installation detect likely falls, avoid constant false alarms, and notify the right people quickly enough?”

Test the System Like a Household, Not a Lab

After installation, test the boring failures. Walk the parent’s normal routes. Leave the pendant on the nightstand. Turn off Wi-Fi if the system depends on cloud alerts. Put the caregiver phone in Do Not Disturb. Ask what happens if the first caregiver is driving, asleep, or out of service.

• Room test: simulate safe, non-injurious posture changes near the bed, toilet, and shower exit, and confirm whether the sensor state changes as expected.
• False alarm test: run fans, close doors, let pets move through the room, and watch for unwanted triggers.
• Notification test: confirm the alert appears on locked phones, overnight, and for backup caregivers.
• Power test: verify what recovers automatically after a router reboot, hub restart, or sensor power loss.
• Access test: decide whether a smart lock should unlock automatically, unlock only after caregiver confirmation, or simply provide emergency access instructions.

Be careful with automatic door unlocks. In some homes, unlocking the door after a high-confidence alert may be reasonable. In others, it creates a security problem or a false-alarm risk. A safer pattern is often “notify caregiver, ask for confirmation, then unlock” unless the household has a very clear emergency plan.

A Practical Build for 2026

For most technically comfortable families, the balanced build looks like this: one mmWave radar sensor in the bathroom, a second in the bedroom if budget and layout justify it, a wearable fall detector the parent will actually wear, and a hub automation that escalates beyond a single app notification.

The best setup is the one with the fewest unowned assumptions. If the wearable depends on charging, name who checks it. If the radar depends on Home Assistant, name who maintains it. If the first alert goes to one caregiver, name the backup. If the parent will not wear anything outside, write that down instead of pretending the room sensor follows them.

A layered, privacy-respecting fall detection system can reduce obvious gaps: radar watches private high-risk rooms without filming, the wearable follows the parent beyond those rooms, and the hub turns detection into lights, audio, notifications, and access support. It remains imperfect. The responsible finish is a short household document: what is covered, what is not covered, who receives alerts, what they do first, who is backup, and what changes if the parent stops wearing the device.

References

1. Facts About Falls, CDC.gov
2. Fall Detection Systems Market Size, Share & Trends Analysis Report, Grand View Research
3. Aqara Presence Sensor FP2, Aqara
4. MR60FDA2 60GHz mmWave Fall Detection Sensor, Seeed Studio Wiki
5. Vayyar Care, Butlr
6. MDsense, Essence SmartCare
7. Accuracy of Apple Watch for Detection of Falls: A Systematic Review, JMIR, 2019
8. Best Medical Alert Systems with Fall Detection, SafeHome.org, June 2026
9. A Privacy-Preserving IoT-Based Fall Detection System for Smart Homes, MDPI Sustainability, 2023