--- title: "A Chinese lab is watching 1.27 million elderly people for falls — with radar, not cameras" date: 2026-09-28 category: AI in Action site: NeuroAI canonical: https://neuroai.site/a/na-app-ai-eldercare language: en --- # A Chinese lab is watching 1.27 million elderly people for falls — with radar, not cameras > As China's over-60 population crosses 310 million, mmWave-radar AI is being deployed to detect falls and track vitals without wearables or cameras. The clearest example is a Huazhong University platform serving 1.27 million seniors; in 2025 the ministry also published a national elderly-products catalogue recognising such devices. An elderly man lives alone in Wuhan. He slips in the bathroom. There is no wearable on his wrist and no camera on the wall — but a small radar notices the fall within seconds and alerts his daughter and a community worker. This is not a concept video. In parts of China, it is running today. The trigger is a number no policy can ignore: China is ageing faster than almost any large country has before. ## The problem is a demographic wall China's National Bureau of Statistics (国家统计局) reported that at the end of 2024, **310.31 million people were aged 60 or older — 22.0% of the population** — and **220.23 million were 65 or older (15.6%)**. That is the first time the over-60 group passed 300 million. Because most Chinese elderly age at home (居家养老), the risk is quiet: a fall, a long lie on the floor, a stroke with no one nearby. Traditional bedside monitoring exists but is awkward — a wired ECG monitor can cost **hundreds to thousands of yuan a day (≈ US$40–400)** and tethers the patient to a bed. ## Radar that sees without a lens The technology turning the corner is **mmWave radar (毫米波雷达)**. It emits radio waves and reads the echoes of movement and even breathing — no camera, so no footage of a person undressing; no wearable, so nothing to forget to charge. An AI model on the device then judges what the signal means: - **Fall detection** — a sudden collapse or long motionless spell. - **Vitals** — heart rate and respiration, tracked passively. - **Sleep and bed-exit** — how long someone stayed in bed, when they got up. A fall or an abnormal heartbeat triggers an alert to family and, in community deployments, to a local worker — the "golden minutes" that decide outcomes. ## The deployed example: a university platform The most concrete Chinese case is the **Shuzhi Chuangxin Yiyang Health Platform (数智创新医养健康平台)**, built by the team of Academician Ding Lieyun (丁烈云) at Huazhong University of Science and Technology (华中科技大学). According to Wuhan municipal government reporting, over five years of deployment the platform has served **1.27 million elderly people (127万)** with health management, vitals monitoring and fall recognition. In **December 2024** it entered Wuhan Union Hospital (协和医院) for clinical cardiology use, screening for silent cardiovascular disease and supporting post-discharge monitoring. In comparison tests against traditional ECG monitors, the team reported accuracy of about **95%** — while transmitting data only every **7–8 seconds**, a deliberate cost tradeoff the researchers say suits wards, care homes and home use rather than high-frequency ICU monitoring. ## The policy tailwind This is not just research. In 2025, China's Ministry of Industry and Information Technology (工业和信息化部, MIIT) published the **Elderly Products Promotion Catalogue (《2025年老年用品产品推广目录》)**, and mmWave fall monitors were among the recognised categories — companies such as Quanthium (旷时科技) and iCare (爱牵挂) had devices listed. A national catalogue does not mandate adoption, but it signals which technologies the state wants scaled. ## What it cannot do yet The radar approach is privacy-friendly but not magic. Accuracy on falls depends on placement and environment, and a 95% figure from one team's comparison test is not an independent audit. The 7–8-second data cadence means it is a safety net, not continuous ICU-grade telemetry. And "smart elderly care" still needs a human on the other end of the alert. ## What readers can do now - **If you care for an elder in China**, ask care providers whether they offer non-contact (非接触式) radar monitoring instead of cameras or wearables — privacy and compliance both improve. - **If you build hardware**, the differentiator is mmWave plus edge AI plus privacy; the MIIT catalogue is a market signal worth tracking. - **If you follow policy**, ageing is now a stated national strategy, and passive monitoring is exactly the kind of tech being encouraged and funded. ## Honest limitations The ageing figures (310.31 million over-60, 22.0%; 220.23 million over-65, 15.6%, end-2024) come from China's National Bureau of Statistics (stats.gov.cn) and CCTV. The Huazhong University platform statistics (1.27 million seniors served, ~95% comparison accuracy, December 2024 Union Hospital deployment, 7–8-second data cadence) come from Wuhan municipal government portal / Changjiang Daily (长江日报) reporting and are a single government-media source, not independently audited; the 95% figure is the team's own comparison-test claim. The MIIT 2025 Elderly Products Promotion Catalogue is real per MIIT and company releases. No other RMB amounts beyond the ECG cost range shown. Analysis is current to 28 September 2026. --- Published by NeuroAI (https://neuroai.site/) — https://neuroai.site/a/na-app-ai-eldercare Free to quote with attribution and a link to the original.