Balancing on One Leg

I modified Keith Code to operate the 12 degrees of freedom in the legs. I was playing around with the robot to find its stability when standing on one foot. I was successful on getting to shift and stand as you can see in the video. It became apparent that Keith's method of coding was not good for expanding, the code got very large and messy. I also had to do all the code heavy lifting by hand, this needed to be changed to be generated on the fly.

Torque ON and LED

After the success with Keith Code, I knew I needed to test my knowledge I gained to see if I actually learned how the communication was broken down. I used an example that was in the Dynamixel RX-28 manual, turn the Torque On/Off and turn the LED On/Off. I used Keith Code structure, and with some tinkering around I managed to make a successful communication of the functions. Here is the screen shot of the front panel (user interface) of the code and a video of the operation (my version of Hello World).

Mini-Hubo First moves with Keith's Code

The best way to start a LabVIEW program is to start from existing code. I knew that Keith Sevcik one of my colleagues from DASL made a basic communication with Dynamixels using LabVIEW. Keith's program was built for a specific reason actuating the 3 axes of the Gantry Pan, Tilt, and Yaw camera mount. I am also a very visual learner, so I dissected his code probing it to understand the the basics of Dynamixel communication. To find out more on the Dynamixel I wrote a post on it when I was in Korea Designing the ATLAS Humanoid. I ran my phone bill through the roof for January, picking Keith brain on the way it worked (Thanks Keith!) I created a basic Tutorial in layman's terms on how the 8 bit communication works and the Dynamixel control table. I guess the biggest challenge for me was understanding the CheckSUM and calculating the length. At the end of the communication string of characters, an area adds up the data verifying that the full length was received by the device. Also calculation of the length had a formula that counted the motors and another length parameter. I was coming up short when I tried to calculated it by hand, took me a while to wrap my head around this concept. These two points in mind pushed me in my future code revisions to make the code calculate these internally to allow for expandability and elimination of error.

~Keith's code gave me a Control that I could reference anytime I made new changes. His code worked first time with communication with the Dynamixel!

Lets Play Catch Up

I have let a bit of time pass from my last post, so let's give a few hints to the future topics that came about the last month.

• Mini-Hubo First moves with Keith's Code
• Torque on and LED
• Balancing on one Leg
• FitPC2 Arrived, Dr. T Frequents the Cube, with the High Ups
• The fall - and Burn out....
• The birth of the Gantry
• FitPC2 Running LabVIEW 2009!
• First Code Review
• Code rework, no more Local variables!
• Code Rev_01
• Academic Mini-Hubo Presentation
• Welcome a Girl to Engineering Day (Mini-Hubo Wears a Dress)
• Code Rev_02
• Final Code Review (Perfect)
• Code Rev_03 ( I'm a perfectionist)
• ARCS-10 Paper ( First Paper... not going to well) (Present)
• Mini-HUBO Mark2 Pre-CAM for DASL
• Developer Zone Documents - FitPC2 with LabVEIW, Dynamixel Sync Write (Future)
• Jeannie Falcon, Red Team, Verification!
• RJ leaves NI March 12th :(

Now that is a List!

Weekend 1: ASME Managment Divison

This weekend started off with a simple question I placed on Facebook. "First weekend in Texas... what to do". I got one response and It turned out to be very serious one. A Drexel Alum posted saying there was a ASME Management retreat being held in Austin this past weekend. He left his number, and just said call for details. I called him on Saturday morning and he gave the address and gave a forward message to the chair asking if I could sit in as a guest. The retreat was being held on University of Texas Campus at the A&TT Conference center. Each engineer sitting on the board was very welcoming and surprised that I wanted to spend my weekend with them! The group discussed action items and the largest topic was the Global Engineering Management Conference (GEMC). This conference will be held in Dallas in April of this year, many keynote speakers geared on managing global groups and engineering type tours. I actually liked the management prospective that was conveyed, the group of professional people could move from one topic to the next, dispute civilly and conclude, as well as have fun. This was a huge change than what originally pushed me away from Drexel's ASME student section. This gave me enough interest to pursue this group when I become a professional in a few years.

Anyway, Saturday evening we went to a Texas style ribs place called Stubs. We got the restaurants endless plate of ribs...it was amazing! I feel that my Korea Co-Op has helped me in more ways than one. Especially in the case of talking to industry professionals, I can engage confidently in a conversation and I can arm myself with a large variety of stories geared on international collaboration or just pure technology discussions. After dinner, the younger folk (22-88 year olds) decided to hit 6th street bar/live music district. I made an interesting connection with the the volunteer retreat chair, he is Alum president of my best friend's fraternity and a good friend of my best friend as well. So odd that we would make a connection like that when both of us are from the east coast. We played the evening away in a shuffle board bar, with some very intense competitiveness included! I have never seen these boards before, long tables with pucks the size of your fist that you slide down the smooth surface covered with sand. The puck that stops closest to the opposite side wins the point of all combined pucks. It

really reminded of full sized shuffle board (stick one) or Bocce Ball.

The group invited me back for the second day, I gladly accepted. I arrived in the morning and had breakfast with the group. More meeting until noon, then finished out the day with a tour of

University of Texas (UT) campus. The Host chair, a senior lecture from UT, asked me to return in February to do a guest Lecture focusing on my international research/travel to his engineering students.

So the moral of the story? Posting your thoughts on Facebook, may turn out to be a fantastic networking opportunity!

National Instruments: Robotics Module

The day I started, there was a large development that was happening, National Instruments was releasing the Robotics Module for LabVIEW. "A new version of its graphical system design software that provides a standard development platform for designing robotic and autonomous control systems." This was quoted from the official press release that can be found here. From what I see of it so far it is chock full of robotic like programming. My first thoughts where that it was going to be more for First Robotics, geared for elementary programming, but I have been quite surprised. The programs or "VI's" are on the advanced levels, systems that I have seen my colleagues program from scratch with C or other softwares. Things such as LIDAR, GPS, IMU, Inverse Kinematic solvers etc. Plus, the most basic things like Sharp IR and Ultrasonic sensors to name a few. There is also a huge base for drive/steering control for wheeled robots, including omni wheel/mecanum vector drives. Also, they have a simulation engine for obstacle avoidance and path planning that can be fed into real world drives listed above. I am looking forward in the next few weeks to dig into these VI's, in a later post I will do a demo program with my serial Garmin GPS to show what I can do with this new found Power!

Waterloo Labs

On my first day of work I had a run in with an Application Engineer, Eric, from the second floor. This floor is mainly the 1-3 years out of college employees, that do tech support for the NI software/Hardware. I kind of had a Deer in the headlight look when I first walked into the cafeteria and this fellow helped me out with ordering/ telling me what's in included with my meal. After, he brought me up to his floor and pointed out key people and what they do. It looked like a lot of fun, everywhere we turned there was a new College banner hanging from the ceiling and some homemade type banner made from random electronics/cardboard characters you find in party stores. Eric asked me if I have ever heard of the Waterloo Lab guys, I was unsure. He explained that this group from Application Engineering makes these DIY projects using NI hardware and other cool gadgets. Below is a video of this group with their latest project, an iPhone controlled full size car! Eric introduced me to the guy Hunter, you will see him in the video. Eric said "I think this RJ kid has a unique set of skills that may help you with your new project". That is all I can divulge for now, but they asked me to be a consultant to their group based upon my expertise (for you guys out there who know me, you can piece it together..lol)

National Instruments: Week 1

I arrived to NI to discover first a very different place than I expected. I packed with the thinking that the employees would be in nice slacks and button down shirts. Almost everyone is in jeans, to the point that I stick out like a sore thumb with my designer clothes...I brought a pair of jeans, but for the weekends, I was not intending it to be like academia. But don't let the dress fool you, the employees are serious engineers. Each person is an expert in their own region of NI hardware, software, and engineering concept. The second thing I noticed is that the NI employees are taken care of very well by their employer. There are cafeterias for employee meals, full gym and out door basketball courts. The campus is nestled into a very nice piece of property filled with trees an

d small shrubs. Many side walks are enclosed by arbors, and meander throughout the campus. NI has three business complexes and 2 parking garages. I am located in the largest of the three building C, 8th floor. I have my own cubical that is located between metrology and academic sales. Since I am here during an odd cycle for Interns, I am the only inhabitant. It seems like they used to pack 3 or 4 interns in here so it offers me some room to spread out. The computer that I was provided, is an antiquated piece of machinery, IT spent a few hours remotely fixing the Client when I first tried to access my network mail. I decided that the best course of action was

to bring in my Mega Machine of a computer that I lugged down here from Pennsylvania. I tell you I am the envy of the floor with my dual monitor setup, I tell them you should have seen it before my late crash when I had to down size to just one graphics card...

Anyway, the first week I hashed out what I proposed to do during my 3 month stay. The main goal for my stay to qualify a target computer, attach and employ a basic position posing for the Mini-Hubo robot (hardware & basic LabVIEW Software). I have made my first task to produce a an online tutorial to fill a gap that exists on the NI web and google searches etc. There are a few pictures and videos of the robotis Dynamixels on the Web using labVIEW, but no baseline code to build a platform from. This is a solved problem as

we know, Virgina Tech for may years has been using Labview to run their Humanoids. Though, NI states that the

best way to start programing in LabVIEW is to start from an existing piece of code. So I feel

for the sake of Mini-Hubo, being an open ended project, it would be fitting to produce a basic

instruction on how to communicate with these motors. Later, I'll post

the continuation of my work on

Mini-Hubo, for a base for other universities choosing to make their own Mini-Hubo, if they chose to use LabVIEW as their programming base.

Introduction: Mini-Hubo

Mini-Hubo is a miniature version of the humanoid robot Hubo developed by the Korea Advanced Institute of Technology. Mini-Hubo is an affordable and easy to manufacture model to offer a broader audience an opportunity to build a humanoid robot of their own. The robot is 18" tall, weighs about 6.5lbs and uses Robotis Dynamixel servo motors for actuation. The cost to build Mini-Hubo is about $5,000 and takes about 40 hours for manufacture and assemble. All documentation for the robot is available to the public. The documentation includes Users, Fabrication, Assembly manuals. Also all CAD and NC code files are available.

The robot was Developed by Virgina Tech's RoMeLA group under the funding of the National Science Foundation PIRE grant. More information on Drexel/VT/NSF involvement can be found here:

http://www.nsf.gov/awardsearch/showAward.do?AwardNumber=0730206

Pre-National Instruments

The first three months of my 6 month Co-Op I spent it at Drexel Autonomous Systems Lab. During that time I fabricated a working Hubo2 KHR-4 right leg that I successfully tested on Drexel's Jaemi Hubo. During the three months I produced a innovative tooling system, and produced CAM based on my drawing that I created during my 08-09 KAIST Co-op. More in depth explanation will be available on my webpage in the next few weeks.

http://www.pages.drexel.edu/~rjg48/index.html

Below is my posted youtube video showing the validation of my working Hubo leg.