Tuesday, April 12, 2016

EE Projects: Obstacle Sheets

In my electrical engineering classes, student pairs are hard at work tinkering with Arduino sketches that they think will have some utility towards their ultimate goal of building a robot to defeat their landscape's "obstacles." (project summary here)

Here is some of their project work - I call these the "Obstacle Sheets," where they outline the challenge that their robot will face and provide some ideas for overcoming them. It's just broad strokes at this point; they needed to generate some ideas to explore, and the next task for each of them will be building three Arduino projects (we have a couple of books - The Sparkfun Inventor's Kit and the Make Arduino Book), modifying them to change their functioning, and documenting that process in a Youtube video.


Balloon-Cushioned Robot

Task:

-We will initially drive the robot off of some moderate height

-We are going to use CO2 canisters to inflate balloons

-We are having the robot activate an airbag (the balloons) so that it survives the fall

-This will be motion activated- It will notice that it is in freefall

-Needs to right itself after it hits the ground

-Needs to escape the balloon (Overinflate them)

-Needs to find goal


Items/Research:

-Research Mars Pathfinder

-We also need a 9v gas valve

-c02 canisters

-balloons

-regulator to control pressure


Scared of Everything Robot (meant to be used in an indoor environment, normal floor)

Objective:
Find the most optimal spot
 -Search for darkest spot, only stay if spot meets other sensor requirements
  -quiet, not too hot
Make comments about its surroundings based on sensors
 (optional if rest of project is finished)

 Obstacles:
-Getting around objects that are in the way of robot finding best spot
 Algorithm for getting around object, avoid getting stuck circling room
-Avoiding light (photo sensor)
 Photo Resistor, pg. 41
-Avoiding noise of certain decibel (directional microphone research)
-Avoiding temperature that’s too hot
-check dark spot for space heater
-only turn on sensor when dark spot has been reached
-Getting the robot to speak for certain light levels (speaker and photo sensor)
  Play certain sound files when sensors reach certain levels

Actively search for darkest place (mission)
 -multiple sensors
Secondary move away from loud noises and avoid objects (avoid obstacles)
Final: Turn off when optimal hiding place is found

 Projects Pertaining to Robot:
P. 64 light sensors
P. 65 (SIK Guide) Spinning a Motor:
P. 54 thermostat sensor

 Robot Course of Action

Scan for darkness w/ multiple photosensors, travel using random search (like Roomba)
Constant sensors
Noise (Directional Microphone)
Shake (pivot back and forth) when it detects sound of certain decibel until sound goes away
Physical Obstacle
Move along edge, until turned 180 degrees (close enough)
Sonar
Avoiding walls, follow step 1 until travelled x centimeters in straight line, then give up and turn 90 degrees and restart search
Darkness Reached
Check if good enough
If it’s dark enough
If it’s quiet enough (small speaker)
If it’s not too hot (space heater)

Solar Robot

     The robot runs off of solar power, and when a photocell detects a lack of light it wirelessly triggers a lightbulb above it that is powered by a wall outlet. The robot will have ultrasonic proximity sensors to detect and avoid boundaries. Its search algorithm will initially be random (like that of iRobot’s Roomba).

Goal
- Get the vehicle to central location
- Need voltage regulator

Robot Car: Ditch/cliff Avoidance

Obstacle: Ditches and cliff in the terrain
 Solution: Ultrasonic sensor will be in front of the cart to detect the change of distance between the cart and ground. The distance for it to work is 1 inch to 10 feet.
(page for the sensor)   http://learn.parallax.com/KickStart/28015

 Large objects as barriers
 Solution: multiple flex sensors on different sides of the car

 How does it navigate to the Goal?
 It will have a navigation algorithm. In the algorithm the bot continues to move forward until it reaches an obstacle. At the obstacle, it will back up and turn a certain amount of degrees depending on which flex sensors are touched. It will repeat this until it reaches the goal.

 How does it know that it’s in the goal?
The goal will be under cover like in a cave, so it will be dark instead of light out at the goal. The robot will have a photocell to detect the change of light to identity the goal

 Sources to help us
http://learn.parallax.com/node/235

http://blog.miguelgrinberg.com/post/building-an-arduino-robot-part-iv-a-not-so-basic-robot-firmware


Robot Arm Upgrade (expansion of Science Olympiad project)

 Obstacles
Moving quickly and efficiently
Grabbing objects
Pencils
Legos
Dice
Ping pong balls
Programming movement
Increasing precision of movement
Grabbing objects of different textures and moving them.
Robot does not grunt enough or say enough old man things.

 Solution ideas
By building an Arduino that can use programmed macro movements we can move arm to the area of the objects quickly and then human controls will be in charge of the micro movements.

 By using stepper motors controlled by the Arduino, we can move the arm more quickly and more slowly as we need to. We will be able to move the arm specific distances by programming the motors.

 We will shift from belt operated motors and actuators to more precise rack and pinion mechanisms.

 We will reduce the amount of clamps as currently the arm is lousy with clamps.

 We will build more precise actuators, reducing the amount of syringes we have, which is currently a [90's movie reference] amount.

 We will have parts on the robot be of uniform length and ensure the even level of the arm in order to increase precision.

 We will add a speaker that will speak randomized, programmed sounds related to the activity of the arm in order to let users know his feelings and to demoralize the enemy.

 Experiments Related

Circuit #11
Circuit #12
Coding pg. 94, 137
Driving bigger loads pg. 72

Tuesday, April 5, 2016

Electrical Engineering: A Project-Based Course

This year, with the new Electrical Engineering elective (prerequisite: electronics), I'm trying out a fully project-based course. With the background knowledge that they have and a whole lot that they pick up along the way, these students are going to build a robot to respond to this central prompt:

Robots overcome obstacles in many different ways; some are similar to the methods that humans would use and some are markedly different. The goal here is to design an "obstacle" and a robot to navigate the obstacle. Teams will build both the "arena" in which the robot operates and its goal location, and their robot must overcome the obstacle to reach the goal. "Obstacle" could mean many different things, definitely not limited to physical obstacles - a solar-powered robot may have darkness as an obstacle, a robot on Venus would have to overcome high temperatures, and a rescue robot would have to overcome uneven terrain.

Along the way, teams will need to meet several intermediate goals, producing several 'deliverables,' which demonstrate planning, incremental progress, and proof-of-concept for their robot.

Deliverables:

  • "Obstacle" description, along with ideas (plural!) about how the robot might overcome it (team evaluation )
  • Specifications sheet: details your robot's inputs (information from sensors), outputs (expected behaviors, actions, etc.), and expected exceptions (problems that can occur) (team)
  • Three Arduino projects from the texts that could pertain to your problem (individual evaluation; three projects different from your partner's three). For each, summarize how you think it might pertain to your project and show how you modified the project/sketch to change how it functions in some way. Present these as a Youtube video, with commented code (showing especially the modifications) linked
  • Program flow chart (detail the sequence of sensor readings, calculations, and outputs that will take place in the loop, as well as the preliminary variables that need to be set) (team)
  • Contribution to the WCGW? (What Could Go Wrong?) meeting: brainstorming unexpected exceptions for other projects - if you help them figure out the potential issues, they'll be able to design around them. You'll get the same help (individual)
  • Schematic: Arduino and all associated electronics (sensors, motors, LEDs, etc.) (team)
  • Sensor and output validation: show, with isolated snippets of code, that you can accurately measure whatever sensors are measuring and accurately control any output devices (individual - one partner designs, executes, and videos illustration of inputs, the other does the outputs). Present these as Youtube videos
  • Landscape, including the "obstacles" and a goal. The robot needs to be able to detect when it's in the goal! (team)
  • The robot, fully functional (team)
  • Reflection on the process and the big question of how robots overcome obstacles and how that is similar to or different from how humans do (individual)

Each of the deliverables will be evaluated on the 11-point scale. Teams/individuals must earn at least 7 on a deliverable in order to proceed, with revision increasing the grade. The final evaluation will take place together, with the five arenas and robots moving towards their goals simultaneously. This is their 'exhibition,' and I'm planning to invite as large a committee as I can to make it an authentic experience.