byapparov/robot

2 Motor Robot with Raspberry Pi - Named Olie

robot's Issues

Get distance to the obstacle in meters

Here I would like to have a simple way to measure the forward distance:

 # this should return distance to an obstacle in meters
distance  = robot.distance()

Loop where robot moves forward till next obstacle and changes direction randomly to get to the next one

Robot moves straight when forward() or reverse() commands are given

Robot moves to a desired relative point by turning toward that point and going straight

Desired input would be something like this:

# both parameters will be in meters
robot.move(
  x = 1.5, 
  y = 2
)

Error reading mpu6050 data

Sometimes gyro data is not read correctly and an error is coming from mpu6050 library:

Traceback (most recent call last):
  File "robot.py", line 53, in <module>
    turn(360.0)
  File "robot.py", line 26, in turn
    angles = gyroscope.update(sample_time)
  File "/home/pi/robot/src/gyro.py", line 40, in update
    gyro_data = self.sensor.get_gyro_data()
  File "/home/pi/.local/lib/python2.7/site-packages/mpu6050/mpu6050.py", line 213, in get_gyro_data
    z = self.read_i2c_word(self.GYRO_ZOUT0)
  File "/home/pi/.local/lib/python2.7/site-packages/mpu6050/mpu6050.py", line 71, in read_i2c_word
    high = self.bus.read_byte_data(self.address, register)
IOError: [Errno 121] Remote I/O error

Switch to Ubuntu Mate so ROS could be installed and maintained easily

Make robot turn to a degree specified in turn() command

To track the angle MPU6050 with three axis gyroscope is used in this project.

Z axis provides Yaw velocity in degrees which will allow implementation of precise turning to a given number of degrees with velocity integrator and a PID controller.

Example of code that is required:

import robot # we also might rely on gpiozero, it is just an example. 
import yaw_control # this is the package that can contain PID for yaw

r  = robot()
# Turn robot left 60 degrees:
r.turn(60)

# Turn robot right 90 degrees
r.turn(90)

• Update readme
• Add gif or video demonstrating this works

Known challenges that need special attention:

• MPU6050 is not calibrated which means that measurements must be adjusted
• It is not clear how often signal should be read to get precise angle estimate
• There is about 0.5 d/s noise in the readings which might require kalman filter to be added