Cosi119a

* READING: <a href="/content/topics/homeworks/hw_readings_119/110_hw_119_read_ai_robot/">Read Artificial Intelligence – An Introduction to Robotics</a>

*Reminder: "Readings" are your responsibility. You will be expected to come to class prepared, having read the material, and ready to participate in the discussion*

<slide_break></slide_break>

<h2 id="robot-hardware">Robot Hardware</h2>

<ul>
  <li>Subsystems are ‘represented’ in the software</li>
  <li>Locomotion (e.g. wheels)</li>
  <li>Actuation (e.g. arms)</li>
  <li>Sensing (e.g. cameras, lidar, buttons, encoders)</li>
  <li>Computing (e.g. single-board computer SBC, microcontroller)</li>
</ul>

<slide_break></slide_break>

<h2 id="locomotion">Locomotion</h2>

<ul>
  <li>Holonomic - non-Holonomic
    <ul>
      <li>Holonomic - can move in any direction</li>
      <li>Non-holonomic - can’t move in any direction</li>
    </ul>
  </li>
  <li>Degrees of freedom</li>
  <li>Ambiguity in defining DoF</li>
</ul>

<slide_break></slide_break>

<h3 id="wheeled-vehicles">2 wheeled vehicles</h3>

<ul>
  <li>Differential drive: Turtlebot3, Platform, Bullet, Branbot</li>
  <li>Turning by different speed on two or more wheels</li>
  <li>Sometimes there is a caster</li>
  <li>Can only move forward and backward or pivot in place</li>
</ul>

<slide_break></slide_break>

<h3 id="wheeled-vehicles-1">4 wheeled vehicles</h3>

<ul>
  <li>Not up and down or (directly) sideways</li>
  <li>With more than 2 wheels, or a track - skid steering</li>
  <li>Cars have “Akkerman” steering (4 wheels, 2 that can steer)</li>
  <li>4WD “Mechanum” steering</li>
</ul>

<slide_break></slide_break>

<h3 id="quadruped">Quadruped</h3>

<ul>
  <li>Gait - trot, gallop, etc</li>
  <li>Sensing vs. Non Sensing issues</li>
</ul>

<slide_break></slide_break>

<h2 id="common-case-differential-drive">Common case: Differential drive</h2>

<slide_break></slide_break>

<h3 id="lower-level">Lower Level</h3>

<ul>
  <li>Subscribes to <code>cmd_vel</code>
    <ul>
      <li>Desired translation and rotation speeds</li>
      <li>Reverse kinematics tells us desired rotations per second for each wheel</li>
    </ul>
  </li>
  <li>Publishes to <code>odom</code>
    <ul>
      <li>Forward kinematics tells us estimated actual motion (position and speed)</li>
      <li>‘Dead Reckoning’</li>
    </ul>
  </li>
</ul>

<slide_break></slide_break>

<h3 id="supporting-hardware">Supporting hardware</h3>

<ul>
  <li>Two wheels + 1 or 2 casters</li>
  <li>Each wheel has motor and encoder</li>
  <li>Motor will go “faster” when given more current</li>
  <li>Encoder counts up for each rotation</li>
</ul>

<slide_break></slide_break>

<h3 id="algorithm">Algorithm</h3>

<ul>
  <li>PID control to minimize delta between desired and actual</li>
  <li>Forward kinematics</li>
</ul>

<slide_break></slide_break>

<h2 id="sensors">Sensors</h2>

<slide_break></slide_break>

<h3 id="lidar">Lidar</h3>

<ul>
  <li>Rotates and uses a laser to measure distance to nearest obstacle</li>
  <li>Returns a vector of values corresponding to the directions</li>
  <li>LiDAR’s can operate in 2d (scan) or 3d (point cloud), ours is 2d</li>
</ul>

<div class="callout callout-small">
  <span class="callout-badge">Limitation</span> Operates in a plane, like a disk. Can't detect an obstacle above or below the plane!
</div>

<slide_break></slide_break>

<h3 id="visual-cameras">Visual Cameras</h3>

<ul>
  <li>Webcam sees a color picture</li>
  <li>Matrix of dots, each dot has a color</li>
  <li><code>Dot[x,y] = {r,g,b}</code></li>
  <li>Data can be processed for example with <code>opencv</code></li>
  <li>A lot of data and a lot of processing</li>
  <li>Bandwidth limits</li>
</ul>

<slide_break></slide_break>

<h3 id="depth-cameras">Depth Cameras</h3>

<ul>
  <li>Like a Microsoft Kinect</li>
  <li>In addition to <code>{r,g,b}</code> also has distance so we get <code>{r,g,b,d}</code></li>
  <li>Useful for getting range measurements of an object detected through camera</li>
</ul>

<slide_break></slide_break>

<h3 id="computation">Computation</h3>

<ul>
  <li>Distributed environment, typically:
    <ol>
      <li>roscore running onboard. on Raspberry pi</li>
      <li>Some additional ROS nodes running on Raspberry Pi</li>
      <li>More ROS nodes running on a laptop or other computer</li>
    </ol>
  </li>
  <li>Workload is truly distributed across all of those</li>
  <li>Eventually we may have to have a much beefier computer running all three</li>
  <li>This would make the robot much more independent</li>
</ul>

<slide_break></slide_break>

<h2 id="our-actual-robots">Our actual Robots</h2>

<ol>
  <li>Differential Drive - 2 powered wheels and one (or two) casters</li>
  <li>Arduino or Arduino-like computer to control the motor.</li>
  <li>Motor Controller and IMU directly connected to Arduino</li>
  <li>Raspberry Pi which runs Ubuntu and ROS, connected to Arduino via a USB cable</li>
  <li>Lidar connected to Raspberry Pi connected via USB cable</li>
</ol>

<slide_break></slide_break>

<h3 id="diagram">Diagram</h3>

<slide_break></slide_break>

<h2 class="shadow p-3 bg-white rounded">Thank you. Questions?<img class="img-fluid w-100" src="https://picsum.photos/800/100.jpg" /> <small> (random Image from picsum.photos)</small></h2>