mrc

MicroPede Robot Controller 🤖🖥👾

The MRC lets you control 6 DOF/axis robotic arms using the g-code like MRIL language. The istructions can either be transferred directly to the controller using screen, or any other serial monitor, or with the help of a library, such as the MRCL (MicroPede Robot Control Library)

Zu den Robotern hier entlang.

This project is work in progress. Feel free to fork, update and improve it 😃

All configurations and settings are based on the robot model depiced below.

bildschirmfoto 2017-05-15 um 12 11 43

Getting started

Clone the repo

git clone https://github.com/glumb/mrc.git

Set you microcontroller in the platformio.ini.

Uncomment #define EXAMPLES 1 in the main.cpp file to build the calibration example.

In CalibrateServos.h set the pin_servo_n according to your setup.

Build the project.

platformio run

Upload the program to the microcontroller.

platformio run --target upload

Calibrate your servos (see section Calibration).

Change the configuration to match your robot (see section Configuration).

Optionally attach a 128x64 display to the pins defined in main.cpp

Comment #define EXAMPLES 1, then build and upload the project.

Use the MRCP/MRIL to control the robot.

Send a few commands via serial to check if everything is working. B should return the free buffer size (300).

PlatformIO

This project was developed with the help of PlatformIO (PIO). PIO is an open source cross-platform build system and Atom based IDE. It is used to build and upload programs to a microcontroller. I highly recommend using the PIO IDE or at least the PIO build-system to upload the program to your Arduino/Teensy.

Configuration

Most configuration is located in the config.h file.

Aspects that have to be configured:

Config	Description
Servo Pins	Set which PWM pins you want to use on your μC for driving the servos
Servo Config	Assign pinNumber, maxAngularVel degree/sec, calibMin, calibMax, angleDegMin, angleDegMax and home position to each servo. This configuration can be generated using the `CalibrateServos.h` in the examples directory.
Additional Axis Pins	Set which PWM pins to use for additional Servos. (e.g. grippers)
Geometry	Configure the robot geometry based on the convention depicted above. The base coordinate system (left, bottom) is used as reference. V3 must have no offset in Y, Z. V4 must have no offset in X, Y.
Pin Map	The PinMap is used for setting the IO pins, used in the MRIL I, O commands (I2 1 wait for pin 2:high, O7 0 set pin 7:low). A PinMap {8, 3, 24}, maps the physical pins 8, 3, 24 to the logical pins 0, 1, 2.
Logical to Physical Angles	A method to transform logical to physical angles. (See section kinematics for more detail.)

Servo Calibration

The servos have to be calibrated to relate PWM pulses to physical robot positions. Therefore build and upload the /examples/CalibrateServos.h file. Open a serial monitor and connect to the microcontroller. Use the keys 0-9 to select a servo. Press n to move a servo without changing the config. w increments, s decrements the current value.

Workflow

You basically alter the servo frequency to move the servos to certain positions and then relate those positions to physical angles. You can always press h to display all commands and explanations.

select a servo using keys 0-9
press n to set rapid movement mode
move the servo to the maximum position (using w,s) according to the joint rotation direction depicted above.
press z to enter "set maximum PWM value" mode.
again use using w,s for small adjustments. Presseing w,s now sets the maximum PWM value
use a ruler or similar device to measure the current angle of the joint. The home position for all joints is also depicted in the figure above.
press r to enter "set maximum angle" mode.
use w,s to set the angle.
repeat the steps for the minimum angle in the same way.
one max/min angle and PWM value are set for one joint, press q to move to home (0°) position. If the robot does not move to 0° according to the above diagram, adjust the angles or PWM values accordingly.
when done, press p to print the configuration. Copy and paste it into the config.h file.

 INCREMENT 'w'
 DECREMENT 's'
 CHANGE_MIN_ANGLE 'e'
 CHANGE_MAX_ANGLE 'r'
 CHANGE_MIN_FREQUENCY 't'
 CHANGE_MAX_FREQUENCY 'z'
 MOVE_TO_ZERO 'q'
 SET_MOVE 'n'
 PRINT_CONFIG 'p'
 CHANGE_MODE 'm'
 PRINT_HELP 'h'

Kinematics Logical - Physical Angles

The MRC uses two types of angles - logical and physical angles. Logical joint angles are used internally by the robot controller and are the result of the IK calculation. The robot is therefore modeled as a series of joints and links. However, a real physical robot may use mechanical links to move some axis. Therefore moving R1 may also move R2, because of the kinematic linking.

To account for mechanical kinematics, the methods logicalToPhysicalAngles and physicalToLogicalAngles are used. On the robot depicted below, both motors for J1 and J2 are at the position of J1. A link is used to drive J2. Therefore moving only R1 also decreases the angle of J2. To not move J2, R2 has to move the same amount as R1.

This is represented using the methods like this:

void logicalToPhysicalAngles(float angles[6]) {
    angles[2] += angles[1];
}

void physicalToLogicalAngles(float angles[6]) {
    angles[2] -= angles[1];
}

Tests

For testing and mocking gMock/gTest is used. To run the tests /tests/test.h.

Debug

increase the log-level in Logger.cpp to see whats happening in the controller.

Troubleshooting

MRC does not respond unexpected data in the EEPROM may freeze the MRC. In main.cpp uncomment Eepromstorage.clear(); in line 75.

can not send commands make sure your serial monitor sends a '\r' when pressing return. Otherwise in MRCP.h change the MRCP MRCP_END_FRAME = '\r' to any unused char.

todo

implement circular interpolation methods
implement proper logging
dependency inject the logger
implement interpolation of poses (quaternion slerp?)

glumb / mrc

readme

mrc