Predictive algorithms detect free-fall signatures before impact, classifying severity from minor tumbles to critical falls.
"The bracelet can feel when you're about to fall — even before you hit the ground! If it's a big fall, it tells your parents right away. If you just tripped a little, it knows that too and doesn't bother them."
The 6-axis IMU (accelerometer + gyroscope) samples at 400Hz, feeding a real-time neural network trained on 50,000+ fall signatures from children aged 3-12.
The algorithm detects the characteristic free-fall phase (near-zero G) that precedes impact. This gives a 30-50ms warning window before the child hits the ground.
Impact severity is classified by peak G-force: Minor (<2.5G, logged only), Moderate (2.5-5G, guardian notification), Severe (5-8G, urgent alert), Critical (>8G, emergency alert with auto-911 option).
The system distinguishes between falls and normal play activities (jumping, running, climbing) by analyzing the full motion signature: pre-fall trajectory, impact vector, post-impact motion, and recovery pattern.
| IMU | 6-axis, ±16g accelerometer, ±2000°/s gyro |
| Sample Rate | 400Hz continuous |
| Pre-Impact Warning | 30-50ms before impact |
| Minor Fall | <2.5G — logged, no alert |
| Moderate Fall | 2.5-5G — guardian notification |
| Severe Fall | 5-8G — urgent alert |
| Critical Fall | >8G — emergency alert + auto-911 |
| Training Data | 50,000+ child fall signatures |
| False Positive Rate | <2% (play vs fall discrimination) |
| Post-Fall Monitoring | 60s heart rate + motion tracking |
Select an activity to see the real-time G-force visualization and how the algorithm classifies each movement type.
SafeStep's 6-axis IMU (accelerometer + gyroscope) samples at 200 Hz to distinguish normal play from real falls. The AI algorithm analyzes G-force patterns, freefall duration, and post-impact movement to eliminate false positives.
1.0G
Peak Force
1.00G
Average
200
Hz Sample Rate
Standing / Idle
SafeNormal resting state. Bracelet detects minimal movement with steady ~1G from gravity.
No action — baseline monitoring
6-axis IMU captures acceleration and rotation at 200 Hz — 200 readings per second across X, Y, Z axes plus gyroscope.
On-device ML model classifies motion patterns in real-time. It learns the difference between play (rhythmic) and falls (asymmetric).
A fall signature is: sudden 0G freefall (>100ms) → high-G impact (>5G) → post-impact stillness. All three must be present.
Mild falls get a 15s check-in. Hard falls trigger immediate alerts. False positive rate: <0.1% thanks to multi-axis analysis.
200 Hz
Sample Rate
<200ms
Detection Time
<0.1%
False Positive Rate
±16G
G-Force Range
AI Danger Prediction
Microphone, pulse, and emotion sensors fuse together in real-time to classify threats before they escalate.
29-Sensor Intelligence
Twenty-nine sensors work together to distinguish between safe play and real danger — swimming vs drowning, roughhousing vs assault.
Hot Car Detection
Automatic temperature monitoring with escalating alerts and auto-911 dialing when a child is left in a hot vehicle.
Ready to protect your child?