Robot Sensor Platform Meeting Notes 2012 09 16

From Artisan's Asylum

Jump to: navigation, search

Meeting notes

In attendance: Richard, Brandon, Andrew, Justin, Jeremy

  • General why we are here; build cool robot and people will know of this cool thing
  • Vending machine fix
  • old robot issues; line following; sensors; use of dolly for navigation testing
  • Initial goals
    • better base
    • wiring trays/channels
    • battery box
    • E-stop
    • cover chainstays
    • diagram and documentation (most of these details are found in Robotics_Intensive)
  • Vending machine goes back together; not part of this group
  • GOAL - drive to Gui's office in telerobotics mode; need a wireless connection, web camera, netbook
  • Schematics - use Eagle
  • TODO LIST (by next Sunday, 2012-09-23)
    • Brandon - get materials for better base and possible grid top
    • A2 - wikify things/document
    • Richard - post pictures; vending machine issue; contact Charlie
    • Justin - N/A
    • Jeremy - N/A


Richard's more detailed notes:

Public Robotics Lab Meeting/Work Session #1 Sunday, September 16, 2012

In Attendance:

  • Andrew
  • Justin
  • Jeremy
  • Brandon
  • Richard
  • Colin
  • (Avi also stopped by a little later in the afternoon)

--

Quick summary:

We met in the electronics area starting about 12:30. We introduced each other and indulged in some big-picture planning. Around 2 o'clock we moved to the robotics area to inspect the motorized base that was built for the Robotics Intensive class. We cleaned out the storage shelves, took notes about and pictures of the base, and decided to get it moving again next week. We also agreed as a near-term goal to rewire and enclose some of the electronics (especially the power electronics) and to build a platform atop the base for mounting laptops, sensors, etc, with a possible goal of remotely driving from the robotics area to Gui's office.

---

Longer notes:

The group that met includes 2 former FIRST mentors, a couple automation engineers, and a few people who have done robotics-related work in college or grad school. Brandon is at the Asylum during working hours (and quite a few other times, I'm sure!) to work on his Rascal micro project. Colin, Avi, and I are alumni of the Robotics Intensive class from last spring.

Quite a few people said that they showed up in order to work on the kind of nontrivial robotics/engineering project that they can't work on at their job, moreover that they would be pleased to use their new skills professionally. It seemed that everyone wanted to tackle the harder aspects of autonomy -- such as vision, mapping, navigation, planning, and state estimation -- and not simply engineer workarounds for the problems those techniques attempt to solve.

I expained briefly how the current "base" came to be, and that it was designed to carry a vending machine around the Asylum. Discussion turned to the safety issues involved in having such a heavy base (never mind the vending machine) trundling around an unstructured industrial space like the Asylum. Although I can't do justice to the discussion here, we spent a while discussing: what would be the purpose of having an autonomous robot wander the asylum (PR value?), and did it make sense to deliver a human-in-the-loop, semi-autonomous system that could be watched overhead and/or through a first-person webcam.

(Somewhere in there, I proposed that we consider trying to succeed at some kind of "safety rodeo", and proposed the name LAARS -- the Large Autonomous Asylum Robot Safety project--because what we have is heavy metal and bangs on things.)

Some technical ideas that came up in discussion: Andrew had done thesis (?) work on "stars", or "artifical constellations" -- small strobes which blink at different frequencies and which are mounted on the ceiling in order to allow robots to triangulate their position. Kiva System's use of QR codes for its warehouse robots -- and the fact that people aren't allowed in their way -- was mentioned. Justin mentioned that researchers at WPI had mounted sensors and laptops to unmotorized pushcarts in order to test mapping algorithms without spending months building a robotic base first. Brandon discussed a little bit about the ARM-based, python-programmed "Rascal" microcontroller he is designing ( http://rascalmicro.com/ ) and his interest in reducing reliance on traditional C and assembly programming in the automation industry.

Andrew expressed an interest in building unsupervised/fully autonomous systems that do something useful and speculated about whether we could solve the safety problem by finding an open outdoor environment where they could be let loose. However, the point was raised that it is much easier to work on a robot if you don't need to travel to its test environment, and that Asylum members would be likely to help out and suggest clever ideas if the working robot was highly visible to them.

Justin expressed an interest, by contrast, in robots that work well together with humans. He argued that it would be an interesting challenge to navigate a cocktail party without getting in the way of the humans, and brought up avionics companies' attempts to build UAVs that can share uncontrolled airspace with piloted aircraft that are following "see and avoid" rules. Justin and I both expressed an interest in AR systems and natural user interfaces that use sensing and computer vision.

We all agreed that, whatever we build, that it is very important to carefully document what we do. We wrapped up the blue-sky discussion by agreeing that we should aim to build a platform that can support many different types of sensors, allowing us to try many different ideas in collaboration and friendly competition.

---

After that, we moved to the robotics area and everyone took a look at the "base" that was built by the Robotics Intensive class. There was some skepticism at first about some of the construction techniques, but eventually we agreed that the base was structurally sound and that, while some of the wiring should be redone and enclosed, that we could work with the base.

  • We decided as a goal for next week's work session to get the robot

moving again.

  • Brandon proposed, and we agreed, that we should build a lightweight

platform on top of the base, ideally with a grid of mount points for securing laptops and sensors. Brandon volunteered to take charge of acquiring the material for the surface of the platform.\

  • We agreed to make CAD drawings (tool unspecified) and a schematic

(probably using Eagle) of the base in its current state

  • We agreed to begin assembling a wishlist of items that members or

friends of the Lab and Asylum could donate

  • Andrew took notes about all the components he found, and volunteered

to create a wiki page for next week with part numbers, datasheets, and specs

  • I took pictures of the base:

https://www.dropbox.com/sh/t97vez9z67jis8g/2bMEfhTgYr

  • I mentioned that I had a copy of the Arduino program that was used

to drive the base previously -- https://github.com/rklancer/SnackAttack/blob/caf11f6d5ba676381794abe849d1a73ba8a1d385/SnackAttack.ino -- and that I had studied it well enough to attempt to refactor it along the lines of Charles' original intent -- https://github.com/rklancer/SnackAttack/blob/caf11f6d5ba676381794abe849d1a73ba8a1d385/SnackAttack.ino (although I didn't have access to the base during the refactoring, so I have no idea if the new code works at all, or how much of it is really needed)

  • We cleaned up the shelves that were dedicated to Robotics Intensive

stuff. We got rid of lots of assorted packaging material and other junk, returned generic electronics stuff to the electronics area, and saved any robot-specific components we could find.

  • We reassembled the sweeper that was originally intended to be part

of a second Robotics Intensive robot

Personal tools
Wiki Maintenance