Bahçeşehir University · Faculty of Engineering and Natural Sciences · Istanbul, May 2025

Smart Parachute System I

A functional, low-cost autonomous parachute deployment system designed to safely recover a 0.5 L water bottle payload from variable, unpredictable heights. The system uses real-time sensor fusion, accelerometer + barometric pressure, running on an ESP32-S3-Zero to detect free fall and trigger a servo-actuated release mechanism with less than 150 ms end-to-end latency.

Project Overview

Objective

Design and build an autonomous parachute deployment device that detects free fall in real time, across any launch angle or trajectory, and deploys a ripstop nylon parachute at the optimal descent point to protect a 0.5 L water bottle payload without false triggers during normal handling.

Key Features

  • • 4-state FSM (Disarmed → Armed → In-Flight → Deployed)
  • • Dual-sensor free-fall detection (IMU + barometer)
  • • Wi-Fi manual override via SoftAP HTTP interface
  • • Audible buzzer alerts across all flight states
  • <150 ms threshold-to-servo actuation latency
  • • 90 % success rate over 10 drop trials at 3.5 m

Technical Stack

  • • Waveshare ESP32-S3-Zero (MCU)
  • • MPU-9250, 9-axis IMU (accel / gyro / mag)
  • • BMP388, barometric pressure sensor
  • • SG90 servo motor, release actuator
  • • C/C++ firmware · Arduino ESP32 framework
  • • I²C sensor bus · PWM servo control

Team

Abdulla Ahmed Mohammed

Mechanical Systems Design & Integration

Amr Nasser Benhalim

Sensor & Power Distribution Wiring

Ghiath Abdul Aziz

Communication & Control Systems

Habiba Hassan Ahmed

Data Analysis & Documentation

Advisor: Asst. Prof. Beste Bahçeci, Mechatronics Engineering

Mechanical Assembly

Interactive Assembly View

Click a part to isolate it · Drag to rotate · Scroll to zoom

SLDPRT
STL
SLDPRT
STL
STL

Component Weights

Minimising mass was critical for parachute aerodynamics. The final PLA structure achieved a total system weight of 145.57 g, printed with 1 mm wall thickness after iterative drop-test refinement.

PartSectionMass
Right ShellParachute Compartment22.75 g
Left ShellParachute Compartment22.75 g
Deployment PinRelease Mechanism2.88 g
GearRelease Mechanism1.04 g
Top PlateRelease Mechanism9.43 g
Main BodyMain Body86.72 g
Total145.57 g

Electrical Subsystem

Microcontroller, ESP32-S3-Zero

Chosen for its compact form factor, built-in Wi-Fi and BLE 5.0, 4.5 MB Flash, and 2 MB PSRAM. Runs the FSM firmware in the Arduino/C++ framework. SoftAP mode exposes an HTTP interface for wireless telemetry streaming and manual override commands.

IMU, MPU-9250

9-axis sensor (accelerometer, gyroscope, magnetometer) sampling at 1 kHz over I²C. Detects free fall across all axes and throw angles by monitoring acceleration magnitude falling below the low-g threshold. Only 9.3 µA average current consumption.

Barometer, BMP388

Provides a pressure resolution of 0.18 Pa, equivalent to sub-metre altitude resolution. Used in conjunction with the IMU to confirm sustained descent before triggering deployment, eliminating false positives from brief jolts or vibration.

Servo, SG90

PWM-controlled actuator delivering 1.6–1.8 kgf·cm torque. Achieves mechanical release in 0.10–0.12 s per 60° rotation. Connected to a snap-in deployment pin that separates the left and right shells on command to release the parachute.

Power, Dual Li-ion + LM2596

Two 3.7 V / 2200 mAh rechargeable Li-ion cells in series provide 7.4 V input. An LM2596 buck converter steps this down to a stable 5 V rail. The servo supply is electrically isolated from the ESP32 logic rail to prevent brown-out resets during peak servo current draw. Average system draw under load: <120 mA.

Communication, I²C + Wi-Fi

Both the MPU-9250 and BMP388 share a common I²C bus, reducing wiring and allowing unique-address multi-device operation. Wi-Fi SoftAP broadcasts a lightweight HTTP server for real-time telemetry and authenticated manual deployment commands.

ComponentSpecificationCost
Waveshare ESP32-S3-ZeroWi-Fi + BLE 5.0 · 4.5 MB Flash · 2 MB PSRAM320 ₺
MPU-9250 IMU9-axis · 1 kHz ODR · 9.3 µA · I²C450 ₺
BMP388 Barometer0.18 Pa resolution · sub-metre altitude · I²C280 ₺
SG90 Servo Motor1.6–1.8 kgf·cm torque · 0.10–0.12 s / 60°50 ₺
Buzzer ModuleAudible status alerts during all flight phases15 ₺
LM2596 Buck Converter7.4 V (2× Li-ion 3.7 V) → 5 V regulated
2× Li-ion Battery3.7 V · 2200 mAh each · rechargeable
Other materialsRipstop nylon parachute, PLA body, fasteners1 100 ₺
System Total2 215 ₺

Control System, Finite State Machine

The ESP32 firmware implements a 4-state FSM in C/C++. Transitions are guarded, the system can only advance to the next state after the corresponding sensor condition or user input is validated. This prevents accidental deployment during handling, launch, or turbulence.

DISARMED

Default power-on state. Actuator mechanically locked. Transitions to ARMED only via push-button input.

ARMED

Both sensors actively monitored. IMU and barometer polled at 1 kHz. System ready to detect free-fall signature.

IN-FLIGHT

Low-g event confirmed by IMU and sustained altitude drop confirmed by BMP388. Deployment imminent.

DEPLOYED

PWM signal sent to SG90 servo. Parachute released. Buzzer activated. Monitoring halts. System enters safe mode.

Click a state to highlight its transitions · Initial state: DISARMED (power-on)

Push button (arm)Push button (disarm)IMU: low-g event+ BMP388: descentFalse positiverejectedSustained descentor Wi-Fi overrideSystem reset / power cycleDISARMEDARMEDIN-FLIGHTDEPLOYED

Performance Metrics

Key measured values from bench-level and integrated drop testing.

< 150 ms

End-to-End Latency

threshold detected → servo actuated

1 kHz

IMU Sample Rate

MPU-9250 acceleration & orientation

0.18 Pa

Altitude Resolution

BMP388 · sub-metre accuracy

< 120 mA

Average System Draw

active sensing + processing

0.10–0.12 s

Servo Release Time

per 60° rotation under nominal load

1.8 kgf·cm

Servo Torque

SG90 release mechanism

90 %

Drop Test Success Rate

9 / 10 trials at 3.5 m height

145.57 g

Total System Mass

1 mm PLA + electronics

Live Telemetry Simulation

Simulated sensor readings based on actual project hardware specs · updates every 0.6 s

IMU Acceleration

1.02g
nominal

MPU-9250 · free-fall threshold < 0.3 g

0 g4 g

Barometric Altitude

3.50m
nominal

BMP388 · 0.18 Pa resolution (~2 cm)

0 m5 m

System Voltage

5.00V
nominal

LM2596 regulated rail · nominal 5.0 V

4.5 V5.5 V

Deploy Latency

112ms
nominal

Threshold detected → servo actuated

0 ms200 ms

Sensor stream active, ESP32-S3-Zero · MPU-9250 @ 1 kHz · BMP388 · LM2596 5 V rail

Drop Test Results

10 trials conducted from a constant height of 3.5 m (first-floor building). Three deployment modes were tested: automatic free-fall detection, perturbed release (lateral push), and Wi-Fi manual override. Descent time measured by stopwatch (±0.5 s human error). Peak impact acceleration measured by the onboard MPU-9250.

TrialModeTime (s)Peak (g)OutcomeNotes
1Auto (normal)1.452.1Soft landingClean deployment, stable descent
2Auto (normal)1.482.0Soft landingVery similar to Trial 1
3Auto (normal)1.472.2Soft landingSlight canopy sway
4Perturbed1.502.5AcceptableLateral oscillation, still controlled
5Perturbed1.522.7Acceptable/harshStronger swing, higher peak g
6Perturbed1.492.4Soft landingGood recovery despite disturbance
7Manual Override1.552.3Soft landingManual deploy triggered on time
8Manual Override1.603.0Harsh (late)Override pressed late, shorter canopy time
9Auto (normal)1.462.1Soft landingConfirms repeatability
10Perturbed1.512.6AcceptableModerate oscillation, no structural damage

Descent range: 1.45 – 1.60 s · Peak acceleration range: 2.0 – 3.0 g · All electrical components intact after all 10 trials.

Summary & Conclusion

The Smart Parachute System I successfully demonstrates an autonomous, compact, low-cost parachute deployment platform. Starting from a clearly defined need, safe recovery of a 0.5 L payload from unpredictable launches, the team systematically worked through requirements analysis, component selection, mechanical CAD design, firmware development, and multi-phase testing.

The final prototype achieved a 90 % drop-test success rate from 3.5 m, with consistent deployment latency below 150 ms and peak impact accelerations between 2.0 and 3.0 g, well within the structural safety margins for both the electronics and the payload. The 1 mm PLA shell reduced total system mass to 145.57 g without sacrificing structural integrity, as confirmed by repeated drop and vibration tests.

The dual-sensor fusion approach (IMU + barometer) proved robust against false triggers during handling and normal vibration. Wi-Fi manual override via SoftAP performed reliably in all authenticated test states, with command gating ensuring that unauthorized or out-of- sequence commands were rejected. Future work (Capstone II) will focus on 3D printing cost optimisation, wider drop heights, and parachute size scaling.