Model Rocket Active Drag Control System

This repository contains the Arduino flight computer code for a model rocket active drag control system. The system reads sensor data, filters it, predicts the apogee using numerical integration, and determines when to deploy or retract airbrakes to hit a target altitude.

Hardware Components

• Arduino compatible microcontroller

• MPU6050 6-DoF IMU

• Adafruit BMP3XX Precision Barometric Altimeter

• Micro SD Card Module

• Servo Motor

Pin Configuration

• Servo Pin: 10

• SD Card CS Pin: 4

• BMP3XX: I2C address 0x77

• MPU6050: Standard I2C

Core Features

Sensor Fusion and Filtering

The codebase uses a 1D Kalman filter to smooth the altitude readings from the BMP3XX sensor. It calculates velocity and acceleration based on the change in these smoothed altitude readings over time. The MPU6050 data is processed using a Mahony AHRS algorithm to determine the roll, pitch, and yaw of the rocket.

Apogee Prediction

A numerical integration loop simulates the remainder of the flight path during every cycle. It uses the current velocity, altitude, air density, rocket mass, and aerodynamic drag coefficients to predict the final apogee at 100ms time steps.

Deployment Logic

The target apogee is set to 750 feet. If the rocket is above the minimum deployment altitude of 300 feet and the predicted apogee exceeds 750 feet for 5 consecutive cycles, the system triggers the deployment state. If the predicted apogee falls 5 feet below the target for 5 consecutive cycles, it triggers the retraction state.

Flight 1 Testing Mode

The current codebase is configured for a passive test flight. The drag coefficients for both deployed and retracted states are set to 0.559. The servo actuation commands within the deployment and retraction logic are commented out, and the servo is locked to 0 degrees during setup. This allows the system to log the theoretical behavior of the control loop without physically altering the flight profile.

Data Logging

The system records flight data to the SD card at 50ms intervals.

The main data file records the following metrics in a comma-separated format: Time in milliseconds, Altitude in feet, Velocity in feet per second, Acceleration X, Acceleration Y, Acceleration Z, Roll, Pitch, Yaw, Predicted Apogee in feet.

A secondary apogee log file records the maximum detected altitude once the rocket reaches apogee, detects a negative velocity, and begins descending.

---

Optimization

To ensure real-time performance and prevent mid-flight hardware crashes, this flight computer software was heavily optimized before deployment. We focused on reducing cycle times and managing the strict memory limits of the microcontroller.

Efficient Algorithms (Compute Time Reduction) The original apogee prediction algorithm utilized a while loop to numerically integrate the remaining flight path using 100ms simulated time steps. This created a variable execution time that scaled with altitude, risking loop overruns and delayed sensor readings. We replaced this O(n) simulation with a closed-form kinematic approximation that runs in O(1) time. This ensures the main control loop always executes well within the 50ms hardware delay limit, resulting in a much more responsive airbrake deployment and perfectly consistent polling rates.

Reduce Memory Usage (SRAM Optimization) Microcontrollers have extremely limited RAM, making memory fragmentation fatal during a launch. We reduced dynamic memory allocation by wrapping all static string literals in our SD logging functions with the F macro, forcing them to be stored in Flash memory instead of the precious SRAM. We also optimized our data structures by strictly defining variable scopes and removing redundant global arrays, which reduced our overall memory footprint, avoided memory leaks, and yielded a noticeably smaller compiled bundle size.

---