asiiapine / simple_imu_reading

The repository comtains everything for IMU data reading via micropython firmware for ESP32

Table of Contents

• Table of Contents
• Install python packages
• ESP32 part :
  • Flash the firmware
    • Micropython Installation instructions
    • Program uploading instruction
• PC part
  • Read messages from USB device:
    • Via picocom
    • Via python script
  • Code part
    • Calibration
  • IMU data reading
• PC part
  • Data visualization
    • Why do we use Redis?
    • Now we can allow us to work with pitch-yaw-roll
  • Redis workflow structure
• 3D visualization

Install python packages

Create python environment with all dependencies and activate it:

python -m venv /venv
source /venv/bin/activate  
pip install -r requirements.txt

ESP32 part :

All the code is stored in esp32_part Remember that the program entry point is main.py

Flash the firmware

Micropython Installation instructions

If you are putting MicroPython on your board for the first time then you should first erase the entire flash using:

Check the port connected to the device:

ld /dev/ttyUSB*

esptool.py --chip esp32 --port /dev/ttyUSBn erase_flash

Then upload the needed micropython firmware instruction all available firmware is here.

esptool.py --chip esp32 --port /dev/ttyUSB0 --baud 460800 write_flash -z 0x1000 FIRMWARE.bin  

Program uploading instruction

Load the program via the instruction:

ampy --port /dev/ttyUSBn put esp32_part/main.py
ampy --port /dev/ttyUSBn put esp32_part/micropython_mpu9250

PC part

Read messages from USB device:

Via picocom

Install picocom

sudo apt-get install picocom

Check the port connected to the device:

ld /dev/ttyUSB*

Use the port name when start reading data via picocom. Use default baudrate with -b flag

 sudo picocom /dev/ttyUSBn -b 115200

Via python script

sudo python pc_part/IMU_serial_reading.py 

Code part

All the code is stored in esp32_part Remember that the program entry point is main.py

Calibration

To calibrate sensor, use the calibration function for gyroscope and magnetometer.

sensor = MPU9250(i2c)

sensor.ak8963.calibrate()
sensor.mpu6500.calibrate()

main.py example with sensor calibration:

import utime
from machine import I2C, Pin
from micropython_mpu9250.mpu9250 import MPU9250
import json
i2c = I2C(scl=Pin(22), sda=Pin(21))
sensor = MPU9250(i2c)

print("MPU9250 id: " + hex(sensor.whoami))

sensor.ak8963.calibrate()
sensor.mpu6500.calibrate()
print("Calibration complete")
while True:
    print("Gyro offset", sensor.mpu6500._gyro_offset)
    print("Mag offset", sensor.ak8963._offset)
    print("Mag scale", sensor.ak8963._scale)
    utime.sleep_ms(1000)

Store the results and use them later after sensor initialization:

sensor = MPU9250(i2c)

sensor.mpu6500._gyro_offset = ...
sensor.ak8963._offset = ...
sensor.ak8963._scale = ...

# start data reading

IMU data reading

example for data reading:

import utime
from machine import I2C, Pin
from micropython_mpu9250.mpu9250 import MPU9250
import json
i2c = I2C(scl=Pin(22), sda=Pin(21))
sensor = MPU9250(i2c)

print("MPU9250 id: " + hex(sensor.whoami))

while True:
    data = {"acc": sensor.acceleration, "gyr": sensor.gyro, "mag": sensor.magnetic, "tem": sensor.temperature}
    print(json.dumps(data))
    utime.sleep_ms(1000)

PC part

==To start the code on ESP, open the port via picocom and press Ctrl+d==

The simplest example contains in simple_serial_reading_to_csv.py. We collect data from the /dev/ttyUSB0 port and then writes it to .csv file for further exploration.

Data visualization

It is almost impossible to check if the sensor works correctly without visualization, therefore lets start with graphs.

The base for graph plotting script: simple_visualize_graph.py.

Why do we use Redis?

The data buffer of the serial port use queue to store data, therefore (since the Matplotlib is slow) the visualization will have a growing delay. So, in several seconds it will be hard to map the graph with real inertia of IMU.

So, now we will use Redis.!.

Redis installation guide

After installation you are ready to start redis service:

sudo systemctl start redis

Updated version of graph plotting script: visualize_graphs_vs_redis.py.

So, to start reading the data from IMU and store it to Redis, run the script: Terminal 1:

sudo python pc_part/read_serial_vs_redis.py

Open the new terminal and run the visualization script: Terminal 2:

sudo python pc_part/visualize_graphs_vs_redis.py

Also, we can speed up calculations, because of splitting up it to several processes.

Now we can allow us to work with pitch-yaw-roll

Redis workflow structure

3D visualization

Terminal 4:

sudo python pc_part/3d_visualization_vs_redis.py

The visualization works if it runs via Debugger. I know, I know, vpython sucks. But easy <3