1) What it is

A privacy-first, room-level safety monitoring system for dementia and elder-care. It uses a wall-mounted IR/thermal sensor with edge AI to detect critical events and immediately alert caregivers—without using RGB cameras.

2) What it detects (core functions)

• Fall detection: sudden collapse/posture change events
• Bed-exit detection: especially valuable during night hours
• Prolonged inactivity / no-movement: possible fall or medical distress
• Presence & basic activity patterns: to support caregiver context (optional analytics)

3) How it works (workflow)

1. Sensor continuously monitors the room using thermal/IR data (non-identifying).
2. On-device AI classifies events (fall / bed-exit / inactivity).
3. When thresholds are met, the system pushes an instant alert to caregivers.
4. If not acknowledged, it triggers escalation (next caregiver / supervisor / nursing station).
5. Events are logged as metadata (time, room, event type, confidence), with privacy controls.

4) System components

• Hardware: wall-mounted IR/thermal sensing unit, facility-ready power/network options (Wi-Fi/PoE as applicable), device health monitoring
• Software: edge inference engine, alerting service, caregiver mobile notifications, and a web dashboard for room-wise status and event logs

5) Alerts and caregiver experience

• Mobile alerts (configurable severity and escalation)
• Nursing-station dashboard showing: room status, active alerts, acknowledgement, device uptime, event summaries
• Configurable rules per room/resident (e.g., night-only bed exit alerts; inactivity duration thresholds)

6) Privacy, dignity, and data governance

• No RGB video and no identifiable imagery
• Edge processing by default to minimize sensitive data movement
• Role-based access to dashboards and audit logs
• Configurable retention of only necessary event metadata for operations and quality improvement

7) Performance targets (pilot-ready)

• Fast detection-to-alert latency (near real-time)
• High sensitivity for falls and high-risk events
• Low false alarms through threshold tuning and workflow confirmation logic (critical for caregiver trust)

8) Deployment model

• Pilot-first: 3–5 rooms, validate accuracy/false alarms, tune thresholds, document impact
• Scale-out: room-by-room rollout with centralized monitoring
• Support: installation guidance, caregiver onboarding, periodic calibration/tuning, remote diagnostics

9) Roadmap (optional extensions)

• Restlessness / sleep disturbance insights (night-time patterns)
• Pre-fall risk indicators (gait instability proxies where feasible)
• Integration with nurse-call / facility SOP workflows and reporting dashboards

If you paste the exact fields shown under “Product Details” (some forms ask for tech stack, pricing model, deployment requirements), I’ll tailor this into the exact format and length the portal expects.