Nannk/nekomimi-wo-kakkidzukete

猫耳を活気づけて ("Animated Cat Ears")

TODO List

new

• 1. Switch to eps8266-rtos-sdk 💀
  • 1.1. merge new project with the old repo(stop using platformio, use idf.py instead).
  • 1.2. determine what components i need to implement.
• 2. Implement components
  • 2.1. Poses (see poses.cpp)
  • 2.2. Servo output (PWM)
  • 2.3. SPI
  • 2.4. FFT
  • 2.5. That Magic part of pose selection i still have no idea how to implement
  • 2.6. Kalman Filter
  • 2.7. Attitude + Gyroscope rotation to angle adjustment
  • 2.8. [optional] Add thermometer or utilise the built-in one for temperature-dependent reactions
• 3. Hardware
  • 3.1. Find a way to fix Axles (Büroklammern) in place
  • 3.2. Wire second AMS1117 on the prototyping board
  • 3.3. Model all the hardware into the CAD project
  • 3.4. Make a board in KiCAD (most definitely for a PCB manufacturer)
  • 3.5. Decide on the Fur
• 4. Misc
  • 4.1. Clean up README (make it actually readable)

old

• Hardware

  • general structure (Hinges, servo placement and Head-to-ears connection) - still have to think where to put that paperclip wire.
  • Microcontroller & IMU
    • Choose the IMU and MCU - it looks like ESP8266 has not enough umpf to read data, do 3 ffts on that data and send the result to serial. I want to use RPi0W and ESP8266, as RPi has only 2 PWM outputs.
      • MCU: NodeMcuv2 (ESP8266) for servo control, RPi0W for data processing.
      • IMU: GY-91 IMU Board through SPI (MPU9250, BMP280) 6 Axis Gyro+Accelerometer, 3 Axis Magnetometer AK8963, Temperature and Pressure sensor BMP280
    • change Voltage regulator to ams1117
    • Make a power distribution board (2 ams1117 in parallel for 3 small servos)
    • Model all hardware into the Project (uncluding headband, CU usw)
  • Fluff (Fur) - this? (a combination of white & black? pink insides?)

• Software

  • Poses
    • List of poses
    • Poses as map of pose codes to desired_angle variables.
  • IMU Data
    • Find a Library (depends on hardware IMU)
      • hideakitai/MPU9250 - with what appears to be a Kalman Filter.
      • MPU9250_asukiaaa - Library that can return raw data, i am using it right now.
    • Get raw IMU data
    • Kalman Filter for attitude aquisition (there was already a lib for that, might just use it. although i need raw data for FFT)
    • How to detect the direction of a movement? From averaged Kalman Filter data?
  • IMU Data processing
    • FFT
    • IMU Data samples to compare to
    • Temperature and/or dTemp/dTime dependent reactions?

1.2.1 RTOS Task graph

graph TD;
A[Task: 
  Read data from SPI] -- IMU data --> B[Task: 
  Perform FFT];
A -- IMU data--> C[Task: 
  use Kalman Filter];
B -- Peaks: 
Magnitude, Frequency --> D[Task: 
  decide on Pose
  choose_pose
  adjust angles]
C --Attitude--> D;
A -- IMU data--> D;
D --servo angles--> E[Task:
  set pwm duty];

Loading

1.2.2 Compute schema

graph TD;
A[Accelerometer] --> KF[[Kalman Filter]] ;
KF --> C[Attitude];
B[Gyroscope] --> KF;

C --> D[Attitude dependent Position];
B --> E[Response to Rotation];
D --> F[Ear::set_ear_position];
E--> F;
F --> G[Ear::move_to_set_angles];

A --FFT--> I[Acceleration Spectrum];
I --> J[Determine Peaks and their Magnitude];
B --FFT--> K[Rotation Spectrum];
K --> L[Determine Peaks and their Magnitude];
L --> M[Compare with predefined Values];
J --> M;
M --> N[choose_pose];
N --> G;

Loading

2.1 List of poses

 # | pose                               | ear left            | ear right           |
   | direction | opennnes               | left | main | right | left | main | right |
---|-----------|------------------------|---------------------|---------------------|
20 | sides     | fully open             | 106  | 90   | 26    | 106  | 90   | 26    |
21 | sides     | fully closed           | 0    | 90   | 132   | 0    | 90   | 132   |
22 | sides     | half open              | 53   | 90   | 79    | 53   | 90   | 79    |
23 | sides     | half closed to forward | 132  | 90   | 132   | 0    | 90   | 0     |
24 | sides     | half closed to back    | 0    | 90   | 0     | 132  | 90   | 132   |
---|-----------|------------------------|---------------------|---------------------|
10 | forward   | fully open             | 106  | 30   | 26    | 106  | 150  | 26    |
12 | forward   | half open              | 53   | 30   | 79    | 53   | 150  | 79    |
13 | forward   | half closed to inside  | 132  | 30   | 132   | 0    | 150  | 0     |
14 | forward   | half closed to outside | 0    | 30   | 0     | 132  | 150  | 132   |
---|-----------|------------------------|---------------------|---------------------|
31 | back      | fully closed           | 0    | 180  | 132   | 0    | 0    | 132   |
33 | back      | half closed to inside  | 0    | 180  | 0     | 132  | 0    | 132   |
34 | back      | half closed to outside | 132  | 180  | 132   | 0    | 0    | 0     |
---|-----------|------------------------|---------------------|---------------------|
and some intermediates

Note: First digit: Ears direction

• 1 - Forward
• 2 - Sides
• 3 - Back

Second digit: Ears opennnes

• 0 - fully open
• 1 - fully closed
• 2 - half open
• 3 - half closed to inside
• 4 - half closed to outside

3.3 CAD Model

servos

2g servos

Brand:Surpass Hobby S0002P 2g plastic gear digital servo Voltage: 3.7~4.2V DC Torque *:≥0.4kgf·cm at 4.2V Speed * : ≤0.06sec/60° at 4.2V Operating Angle: 180°±10°(500~2500μs) Mechanical limit Angle: 360° Weight: 2.2 g Size: 16.7 × 8.2 × 17mm Shell material: ABS Motor type: hollow cup motor

9g servos

TowerPro sg90 (i guess knock-offs)