Fall Warning Dual-Model: Raspberry Pi Vision + 60GHz Radar Monthly Package, Wellness Facility Monthly Revenue of 15k CNY

💰 Revenue

① Small wellness facilities billed per facility monthly (main revenue): Facilities pay a monthly service fee of 3,000 CNY/facility × 5 facilities = 15,000 CNY/month, including equipment rental, algorithm updates, and 7x12 hours of operations, accounting for about 100% of monthly revenue (the sole revenue item in this profile; this figure is derived by multiplying the unit price per facility by the number of facilities, case study data, independently unverified); ② Government elderly care subsidy project annual payment: Interfacing with civil affairs subsidy projects billed annually, with single facility annual payments exceeding 100,000 CNY, annual revenue can reach 300,000 to 500,000 CNY (profile case source, independently unverified), the specific proportion for this path is not provided; ③ Per-bed pricing tier: 50 to 100 CNY per bed per month × 10 beds = 500 to 1,000 CNY/month/facility, provided as an alternative to option ① for small facilities under 10 beds, the proportion of this tier is unstated; ④ Opportunity items - Dual-model effectiveness premium and family value-added: Charging facilities a guaranteed effectiveness premium using a 78% reduction in false-negative rates as a selling point (case study figure, independently unverified), and selling activity heatmap subscriptions to families; pricing for premiums and subscriptions is undisclosed, and their revenue share is also missing.

💸 Cost

One-time hardware investment is about 1,200 CNY/set (Raspberry Pi 5 approx. 600 CNY, radar module approx. 400 CNY, camera approx. 200 CNY), monthly cloud service cost per set is about 50 CNY (used for data backup and remote operations), annual algorithm iteration and maintenance cost is about 2,000 CNY, software tools basically use free open-source versions, with no other fixed costs.

⏱ Time Investment

Initially 2 hours per day for model iteration and hardware debugging; during customer acquisition, 5 hours per week are invested in interfacing with wellness facilities; after stable operations, only half a day per week is needed for algorithm tuning, customer check-ins, and device maintenance.

🔑 Keys to Success

• ✅ Vision + radar multi-modal fusion, increasing fall recognition accuracy to over 95% and keeping the false positive rate below 5%
• ✅ Local offline inference processing, elderly activity data remains on the device, complying with privacy compliance requirements
• ✅ Annual/monthly subscription billing model to establish stable recurring revenue, with reusable equipment to lower marginal costs
• ✅ Interfacing with local civil affairs elderly care procurement directories to lower customer acquisition costs and increase closing rates

⚠️ 风险

• ⚠️ Wellness facilities have extremely low tolerance for algorithm false positives. Daily movements such as bending over, sitting down, and picking up items can be easily misjudged as falls. Caregivers need to assist in labeling scene data in the initial stages. If a fall is missed, civil disputes may arise; product liability insurance must be purchased in advance and liability exemption clauses clearly defined in contracts.
• ⚠️ Complex scenarios in elderly activity areas such as wheelchair occlusion, insufficient lighting, and crowded groups will lead to a drop in recognition accuracy. Continuous collection of real-world scene data is required to iterate the model, resulting in relatively high algorithm optimization costs.
• ⚠️ Elderly care environments are humid and high-temperature, making hardware equipment prone to failure. Radar modules and cameras require regular replacement as they age, generating additional maintenance costs, and an inspection mechanism must be established.
• ⚠️ Elderly video and radar data collected constitute sensitive personal information, requiring compliance with the Personal Information Protection Law, including local encrypted storage and anonymized processing to avoid compliance risks.