Autonomous Robot Part 2 – Acrobot

Group Members: Brian Yang and Luke Hottinger
Roles: Brian Yang and Luke Hottinger as Scribes, Designers, Integrators, and Tutors

Introduction

The power of robotics is in creating autonomous robots that exhibit a behavior and responds to its environment. This is a study on creating an autonomous behavior. The Acrobot is a robot that understands it’s own position and responds accordingly by waving its arm back and forth. The Acrobot is also human-like as it does get tired and requires breaks once in a while. The Acrobot will continues to exhibit the same behavior in various environments autonomously or with human interactions.

Video

Technical Notes

The Acrobot uses a variety of sensors to attain its behavior.  An accelerometer mounted on the center beam of the device keeps track of the pitch of the robot, which is fed to a servo that controls the pivoting arm.  The pivoting arm creates additive motion, allowing the robot’s spin to accelerate.  As the Acrobot flips, a hall-effect sensor tracks the number of rotations as the sensor passes over a neodymium magnet mounted on the stand.  The Arduino program sets a random number of rotations and once the rotations have completed, the robot rests.  The amount of rest is proportional to the number of rotations it just completed.

First Iterations

Completed Second Iteration

The Completed Second Iteration

The Servo Mount

The Arduino Mount

The Acrobot is weighted by a steel plate

The Accelerometer/Gyro

The arduino gets its power from a 9V battery and supplies the power for all of the sensors and the servo.  The structure is made entirely out of OpenBeam providing a very rigid and sturdy structure, allowing the Acrobot to spin freely without worrying about the structural integrity of the base.

Schematic and Code

Link to the Full Arduino Source Code

Flipping Algorithm Code Snippet

void setup()
{

myservo.attach(9);

randRot = random(3, 20);

}

void loop()
{

y = accel_t_gyro.value.y_accel;
z = accel_t_gyro.value.z_accel;
x = accel_t_gyro.value.x_accel;
gx = accel_t_gyro.value.x_gyro;
gy = accel_t_gyro.value.y_gyro;
gz = accel_t_gyro.value.z_gyro;

complementaryFilter(x, y, z, gy, gz);  // Filter pitch data from accelerometer

if (analogRead(A0) == 0)  // Hall Effect Sensor
{
checkPrevious();
}

magnetState = analogRead(A0);

if (rotations > randRot)
{
myservo.write(95);
delay(randRot * 1000);
randRot = random(3, 20);
rotations = 0;
}

if (abs((int)pitch – 0) > 10)
{
myservo.write(95); // Vertical position
}

myservo.write(95 + (int)pitch);
delay(10);

}

void checkPrevious()
{

if (magnetState > 500)
{
rotations = rotations + 1;
magnetState = analogRead(A0);
}
else
{
magnetState = analogRead(A0);
}

}

About lhotting

CMU Information Systems Class of 2017

View all posts by lhotting →