mikeornstein | project RSS

MonkeyBot, Ready to Swing

Mike Ornstein — 2010-11-02T13:01:11-04:00

The structural failures have been remedied and MonkeyBot back in action. We’ve had the bot swing around some and the servos are better than I could’ve hoped! Their feedback is more than adequate for our application and the software side of the project should have a good time dealing with them.

The following are details of MonkeyBot as he stands today.

Full Passive Assembly

The robot utilizes a pair of Dynamixel RX-28 servos, a CM-700 microcontroller and a serial and power tether, all pictured above. In the near future, the robot will be outfitted with simple force sensors in the claw and an IMU in the body.

Claw/Arm Interaction Detail

We created an aluminum plate that fit our current plastic parts such that this interface is much more robust.

Servo Connection Detail

The servo connection consists of 12 2-56 button head screws holding the servo into the body and 8 M2 socket head cap screws holding the arm to the servo horn.

Full Assembly and Failure

Mike Ornstein — 2010-10-28T12:48:37-04:00

After getting in a few of the remaining pieces, the team put together the remainder of MonkeyBot.

Arm Assembly Detail

We were eager to see the bot swing for the first time and we quickly ran into failure:

Failure Detail

A resolution has been devised and we’re moving quickly to get MonkeyBot swinging again.

Quadrotor Landing Pad

Mike Ornstein — 2010-10-23T12:45:35-04:00

Here’s a landing pad made for Quadrotor:

Flight Logging

Mike Ornstein — 2010-10-28T13:51:46-04:00

The group will be instituting flight logging from this point forward. The log will be available as a Google spreadsheet.

Flight Logs

Quadrotor Renders

Mike Ornstein — 2010-10-22T04:12:09-04:00

I took some time to model the quadrotor in SolidWorks and rendered a few shots. The following are the results:

For later branding purposes:

CMUQuadrotor Introduction Video

Mike Ornstein — 2010-10-25T04:08:41-04:00

We’ve spent the last week working away on the quadrotor, getting stable mode up and running and getting some serious flight time in.

Here’s what we’ve got to show for it:

MonkeyBot, in Pixels

Mike Ornstein — 2010-10-14T15:16:02-04:00

This Thursday’s meeting was particularly productive. The body group and passive claw group finalized their designs for laser cutting tomorrow. The Mechanical Engineering Shop got a hold of a brand new laser cutter and we’ll be the first students to produce parts with the machine.

Here is what the bot will look like:

Also the active claw group has modeled a feasible bi-stable, simply actuated claw.

MonkeyBot, in Plastic

Mike Ornstein — 2010-10-15T15:07:56-04:00

The body and passive claw for MonkeyBot have been laser cut and partially assembled.
The following are photos of the parts and the passive claw.

All the parts laid out.

The passive claw: construction required only the 4 (2 visible) fasteners at the base of the claw.

The body, securely capturing the two joint motors.

Fence Flying

Mike Ornstein — 2010-10-13T12:22:07-04:00

With stable hovers possible, we took a little bit of time between classes to take the quadrotor out to The Fence for some fly time. The project got a lot of attention!

[Fence Flight Testing video here]

We were doing a bit more testing including payload maximization and aerial videography.

[Lifting Things video here]

PID Tuning and Stable Flight

Mike Ornstein — 2010-10-13T02:12:18-04:00

As demonstrated by the “First Flight” video, the quadrotor was not very flyable. The group adjusted the PID coefficients in the AeroQuad configurator to improve the craft’s flight performance considerably.

The following video includes snippets from our testing and the current result:

[PID tuning video here]

The quadrotor just turned into everyone’s favorite new toy! It is clear that CMUQuadrotor will be undergoing very extensive flight testing...

We’ll be putting on some LEDs to indicate direction during night flight in the future to improve night time flying capabilities.

[First Stable (Night) Flight video here]

First Flight!*

Mike Ornstein — 2010-10-10T18:07:08-04:00

CMUQuadrotor has left the ground under its own power for the first time!

All videos will be stored on the quadrotor media page for convenient reference.

Although the quadrotor’s flight is not terribly impressive, nor terribly graceful, today marks an important milestone for the project. The next step will be making the craft more controllable.

Making a Case for Quadrotor

Mike Ornstein — 2010-10-08T14:02:24-04:00

To protect the battery and electronics from the elements and hard landings, the team has made an ABS plastic shell for the flying machine.
We used a vacuum forming tool to create the shell over a mold made from insulation foam.

And the final product, top and bottom- trimmed and tabbed. The result is not quite as polished as we would like to have eventually, but at least the guts are mostly protected.

RoboClub Grant and Domain

Mike Ornstein — 2010-10-03T16:12:45-04:00

In an effort to further the quadrotor cause, the group is applying for a RoboClub grant (with the hope of receiving a $500 award). The initial proposal looks like this:

Robotics Club Funding Proposal: CMUQuadrotor Group members: Jeff Cooper, Priya Deo, Jitu Das, Dan Jacobs, Mike Ornstein, Harrison Rose, James Wahawisan, Alex Zirbel. Abstract: Quadrotor helicopters are a popular research platform for aerial work. These helicopters have proven useful in search and rescue, surveillance, mapping, and communications applications in controlled environments. The particular research goals of our project include terrain mapping in unstructured environments as well as navigation in GPS-denied conditions. Both of these goals rely heavily on accurate perception of the quadrotor’s surroundings. The first steps toward these goals have been taken by purchasing and assembling the components necessary to make an effective autonomous hovering platform. The next step is to outfit the newly constructed quadrotor with the electronics necessary to sense and act on its surroundings. We expect to implement computer vision, inertial measurement, and range finding capabilities on the quadrotor to meet our current goals. Future work will include improvements to the aerial platform in addition to enabling HD aerial video and photography using teleoperation.
To increase the visibility of this project, the group has purchased CMUQuadrotor.com

MonkeybarBot Prototyping

Mike Ornstein — 2010-10-01T13:49:47-04:00

In this week’s hardware meeting, the group moved forward with the insight gained from the previous meeting’s activities.

Last week the group prototyped basic ideas to get a sense of proportions and required degrees of freedom to accomplish the monkey bar crossing task. Here are two photos of some of the concepts that were constructed:

A monkey bar set up with 2 bars and a passive, double hook at the end of each arm.

A full scale model of what could be an active grasper.

More prototypes of the active graspers were made this week; examples are shown below.

A possible (prototypical) implementation of a palm triggered grasper.

A simple model demonstrating the functionality of the palm triggered concept.

This week we moved forward in the prototyping stage and also developed a timeline for the project and split our project into tasks. Group members elected to be in different parts of the project, a breakdown follows:

Active Gripper: Yannick, Rodrigo, Carl
Passive Gripper: Chad
Robot Body: Nathan, Priya
Monkey Bars: Andy
Servos/Electronics: Andy, Jeff

Veteran members including Pace, Harrison and me will fill in gaps as needed. The software aspect of MonkeybarBot will be dealt with at a task level by Jitu.

The deadline for a functioning physical robot is October 15, the Friday of Mid-Semester Break. This project will be concluded by the end of the semester, at which time we will have a finalized design for the quadrapedal HyLo robot. Construction of the HyLo robot will begin at the start of the second semester.

Flying This Weekend

Mike Ornstein — 2010-09-29T11:40:41-04:00

The construction of the Quadrotor is 95% complete. The only thing left to do is secure some landing gear and secure the GPS and Receiver modules. In the future, a small amount of work will need to be done to mount additional sensors, but the center mounting plate has an easily extensible design.

Following is a photo of the Quadrotor hanging upside down in the club:

The transmitter and battery should be in later this week in preparation for the maiden flight on Saturday. We picked up a 4000mAh 3 cell LiPo pack from HobbyKing and a 7 channel Futaba Spread Spectrum transmitter/receiver pack from RobotMarketplace with a discount thanks to Dan Curhan.

Quadrotor Construction

Mike Ornstein — 2010-09-25T12:15:40-04:00

I started getting the Quadrotor manufactured and assembled. Using the Robotics Club CNC mill, I made a polycarbonate plate with the help of Harrison Rose to hold the carbon fiber spars.

CNC in progress.

Finished CNCed part

Part in place, holding the spars due to friction.

The Arduino Mega and AeroQuad shield mounted on MDF and the polycarbonate part.

Detail of the AeroQuad shield.

Mechanical Design Underway

Mike Ornstein — 2010-09-23T01:23:09-04:00

I’ve made some decent headway on the mechanical design aspect of the Quadrotor project. I tested some methods for securing the cross spars and input my results to the CAD model.

Here is a preliminary rendering of the vehicle:

Electronics Delivered

Mike Ornstein — 2010-09-22T13:35:31-04:00

The majority of the electronic components for Quadrotor have been delivered.
The order consisted of:

Hardware	Source
Triple axis gyro	http://www.sparkfun.com/commerce/product_info.php?products_id=9801
Triple axis accelerometer	http://www.sparkfun.com/commerce/product_info.php?products_id=9723
Barometer	http://www.sparkfun.com/commerce/product_info.php?products_id=9694
Triple axis magnetometer	http://www.sparkfun.com/commerce/product_info.php?products_id=9371
Arduino Mega	http://www.sparkfun.com/commerce/product_info.php?products_id=9152
GPS	http://www.sparkfun.com/commerce/product_info.php?products_id=9436
GPS adapter	http://store.diydrones.com/product_p/br-0008-01.htm
GPS adapter cable	http://store.diydrones.com/product_p/ca-0001-03.htm
AeroQuad Shield	http://carancho.com/AeroQuad/store/index.php?main_page=product_info&cPath=2&products_id=63

Conceptualization

Mike Ornstein — 2010-09-17T08:47:38-04:00

At the Thursday meeting this week we discussed some high level goals and brainstormed about the simplest way to go about independently solving each problem.

Below is a snapshot of of a some tasks that the robot might eventually complete.

Listed with red bullets are:

Monkey Bars
Tight Rope ~cargo net
Crevice Crawling
Transfer [of] Small Objects
Drive on Wheels

We then went into more detail about how to solve the individual problems:

Monkey Bars:

Tight Rope:

Group members also got a chance to handle the newly arrived Dynamixel RX-28 servo motors.
Next week we will be moving forward with construction of some of the simple robots we conceived during the meeting and put the groups newly found SolidWorks modeling skills to use.

HyLo Software Meeting

Mike Ornstein — 2010-09-15T00:41:19-04:00

This meeting introduced members to ODE.

ODE (Open Dynamic Engine, www.ode.org) is an open source dynamics simulation package. In HyLo, this software will be used to simulate simplified versions of actual robots so that it is possible to distribute a controllable platform.

In the future, we will move the control techniques employed in simulation onto the physical robot.

For now, you can find technical, software related notes on Jitu Das's site.

First HyLo Meeting

Mike Ornstein — 2010-09-09T21:24:24-04:00

HyLo Meeting 9/9/10

Topics

Project Intro
Documentation Past
HyLo's New Goals
Current Direction and short term Objectives
Resources

Introduction
Welcome to the HyLo Project. This project group is about making robotic platforms that explore interesting ways to move around and do useful things. Originally the group set out to construct a relatively small robot capable of climbing up steps. This was seen as a problem for robots of similar size who only rolled along on wheels. To solve the issue of climbing stairs we built a robotic platform that has four legs as well as powered wheels to traverse large obstacles. Typically a robot can not climb anything taller than about half the diameter of its wheels, but by being able to pick up a foot and place it over an obstacle the usefulness of the wheel is extended. To further the idea of unbounded mobility we will be progressing the project upwards, literally. HyLo's new challenge will be to climb a tightrope and a ladder as well as be able to walk and roll.

Documentation Past
What documentation? Scattered and inaccessible. http://mikeornstein.com/projects/files/category-stairbot.php http://mikeornstein.com/stairbot/ http://www.roboticsclub.org/StairBot We will be making use of Redmine's wiki system to keep track of weekly progress (eventually).

Hylo's New Goals
Capable of walking up steps (eventually) Capable of balancing itself Strong enough to perform climbing tasks Built with modular feet in mind

Current Direction and short term Objectives

Mechanical Stuff

Let's make something to play with ASAP. Build a 3 link planar robot capable of brachiation (swinging from branch to branch). Make Environments to test. Learn ODE- Dynamics simulation software (C++). http://sourceforge.net/projects/opende/files/ODE/0.11.1/ode-0.11.1.zip/download

Electrical Stuff

Implement sensors. Learn about IMUs and their associated gospel (more resources to com): http://en.wikipedia.org/wiki/Inertial_measurement_unit Hardware level control- talk to microcontrollers and servos. Learn ODE- Dynamics simulation software (C++). http://sourceforge.net/projects/opende/files/ODE/0.11.1/ode-0.11.1.zip/download

Software Stuff

High level things. Learn ODE- Dynamics simulation software (C++). http://sourceforge.net/projects/opende/files/ODE/0.11.1/ode-0.11.1.zip/download

Resources
Examples to draw from:
Related to rolling and walking http://www.youtube.com/watch?v=6y4PVjh0XBs&feature=related http://www.youtube.com/watch?v=p1xwF6y8MOs&NR=1
About Sloth locomotion (tightrope walking) http://news.bbc.co.uk/earth/hi/earth_news/newsid_8972000/8972024.stm
Paper about brachiation and dynamic control http://kodlab.seas.upenn.edu/uploads/Kod/Nakanishi00.pdf

Quadrotor Begins!

Mike Ornstein — 2010-08-08T16:50:15-04:00

I’ve finally gotten my hands on a quadrotor helicopter! In addition to the bones of the beast I’ve convinced two CS students, Alex Zirbel and Jitu Das, and an ECE student, James Wahawisan (to whom the quadrotor initially belonged) to join my quest for aerial dominance. The project will be an exercise in control and vision in addition to just being an awesome flying platform.

Stairbot -> HyLo

Mike Ornstein — 2010-08-08T16:46:41-04:00

With the upcoming semester looming, there is a need to get the Hybrid Locomotion project back on its feet. Stairbot is now named HyLo. Our goal for this iteration of the design is to use more powerful, accurate and sensitive servo motors at each joint and simplify the mechanical design to make the robot easier to program. Pace Nalbone is going to be in charge of our initial CAD work, designing a new robot of similar form factor that utilizes Dynamixel Servos. More on the new design soon.

Circular Vibratron

Mike Ornstein — 2010-06-24T00:32:38-04:00

The Vibratron project is picking up steam. The following are some fresh renderings.

You can read all about the design changes on Andrew Burks’ blog.

Vibratron Under Way

Mike Ornstein — 2010-06-13T23:15:04-04:00

The Vibratron Project
Having been approved for a SURG (Small Undergraduate Research Grant) for the the Fall of 2010 members of the RobOrchestra team have begun the design and prototype stage for Vibratron. Vibratron came to be thanks to the generous donation of a Vibraphone from Rich Pantaleo Sr, father of recent RobOrchestra graduate, Richie P.

Here are a few teasers:

Priliminary Rendering of the New Vibratron Key Arrangement via Andrew Burks.

Cryptic Scribblings.

Chutes.

I wrote up some more on my random stuff blog.

Meeting of the Minds: RobOrchestra

Mike Ornstein — 2010-05-06T15:02:17-04:00

Roborchestra was recognized by the Studio for Creative Inquiry for the fourth year in a row. This time around we were granted the opportunity to perform in the Studio in addition to receiving a cash prize. We brought the instruments and put on some demonstrations in addition to displaying our poster and entertaining the questions of passer-byes. Here is a small version of the poster.

Click for a full size version (.png).

Meeting of the Minds: StairBot

Mike Ornstein — 2010-05-06T15:00:38-04:00

The Meeting of The Minds presentation was yesterday, and I have to say I’m quite satisfied with how our years worth of work panned out. For this presentation we set up StairBot on a little display table that demonstrated his extreme maneuverability.

In this photo are (from left to right) Wil Hamilton, Harrison Rose and Mike Ornstein.
Our poster is hosted on the site over here.

Mobot Race

Mike Ornstein — 2010-04-25T03:01:42-04:00

The mobot races have finished. Mobert came in 4th passing through 4 gates. The night before was a successful run of 8 gates where our only issue was that we reached the 4 minute limit. Next year will have an all new mobot fast enough to beat the course record.

Parts and Progress

Mike Ornstein — 2010-04-24T20:07:13-04:00

In this post are some photos that were taken prior to the video in the It Moves! post, sorry things are a little out of order! Also included are some reasonably current cell-phone photos of the wholly assembled robot. All that’s missing are the electronics and wiring! Colin Haas has agreed to assist in gait programming of Stairbot, so with his experience in quadrapedal robots we should be walking around in no time!

Notice the extra upper leg segment for insurance. Also not pictured are two extra lower leg segment pieces (4 shown in the top right).

Our freshly laser cut ABS parts. Excellent quality and great fitment, thanks Pololu! All said and done the cost for material, machine time and shipping was about $120, a little pricey but very rapid turn around. This ABS will perform much better than the acrylic we were originally going to use.

A shot of one of the legs assembled prior to the arrival of our additional high-power servos. Looking good! Those spikes do stir up a bit of controversy... You can see this level of assembly wiggle around in the It Moves! post.

It’s excellent news that the robot is statically stable in this position without power. This bodes very well for the robot under power. Stairbot is also statically stable with certain layouts of only 3 of the 4 wheels on the ground, which means that a very stable gait can be made quite simply.

Table mode! Having fully assembled the robot has shown that there is truly a lot of potential for other applications. There will surely be short comings with this iteration of the design but eventually we’ll have a product that can be highly effective at many mobile tasks. Climbing, perching, self-righting and many more things are clearly within reach of this form factor.

It Moves!

Mike Ornstein — 2010-04-10T16:51:40-04:00

Stair Bot Leg Test from Wil Hamilton on Vimeo.

Stairbot: Final Part Count

Mike Ornstein — 2010-03-29T22:46:56-04:00

After much deliberation, we’ve finally got our (hopefully) last part order in. 10 servos are in the mail and soon we’ll the laser cut parts in. We’ve moved from acrylic to ABS in favor increased durability and are outsourcing, likely utilizing Pololu’s laser cutting service (http://www.pololu.com/catalog/product/749). With that said, renderings should be as close to possible as reality sans electronics. Pictures of that, and possibly an animation of Stairbot walking up these glass steps soon.

Body Progress

Mike Ornstein — 2010-03-26T14:13:18-04:00

A series of renderings that are indicative of the final state of the body.

Mobot Renderings

Mike Ornstein — 2010-03-26T13:37:45-04:00

A new name! The mobot is now Mobert.

Here are some renderings of Mobert. These demonstrate the final product compared to the design reference.

The only difference here is a little missing hardware and the electronics (and mess of wires!).
The final mobot will have a lid to protect from sun and rain.

Roborchestra Photo Shoot

Mike Ornstein — 2010-03-25T01:24:44-04:00

Lots of pictures of Roborchestra. The whole band sans snare bot.

Stairbot Photo Shoot

Mike Ornstein — 2010-03-25T01:20:22-04:00

Had a little photo shoot today for Roborchestra so I brought Stairbot out to join the party. Here’s how he ended up:

Guts of Mobot, Complete!

Mike Ornstein — 2010-03-25T01:02:38-04:00

Mobot is finished! (From a mechanical and electrical standpoint).

Realine has been realized.

Here are some photos of the unshelled creature. Design renders will be coming along shortly along with a more detailed explanation of the components pictured here.

Stairbot Update

Mike Ornstein — 2010-02-18T10:48:25-05:00

Stairbot is getting closer and closer to completion. The above render is what the final version should look like. We’ve decided to laser cut a lot of acrylic to build most of the leg and body structure.

We’ve ordered all of our electronics, and we should be receiving them soon. Then we can start actually programming and getting gaits set up. Our goal is to have the construction mostly completed by the end of the month, and start working on some basic motion control.

Below are some images from the build process.

Layout of prototype parts. Hardboard upper and lower chassis halves on the bottom left, upper segment blanks and standoffs top, some failed CNC attempts and a jig for drilling the hole patter for our servo horns on the left.

Shoulder joint aluminum pieces. These need to be optimized yet by milling them down to a thinner final thickness and drilling clearance holes for the servo wires.

Assembled shoulder joint, partially through optimization stage (hole drilled).

Finishing Mobot

Mike Ornstein — 2010-03-24T01:20:26-04:00

Went into overdrive last week, and we’re looking at a nearly complete robot. All that’s left to do on that front is to drill out the brackets to secure the chassis to the wheels. The sensor bars and electronics have been handled and all that’s left is a little calibration and some implementation of code.

More Stairbot Progress...

Mike Ornstein — 2010-03-24T00:46:11-04:00

I created a fresh rendering today to include some steps in v3.4 of the design.

Unfortunately, the shiny acrylic is going to get scrapped in favor of ABS for increased durability.
We still have to place an order for the remaining 10 PowerHD 9150MG servos, and pick up a few odds and ends from McMaster. The only parts besides those being laser cut are those meant for mounting sensors and 3 additional ‘C’ brackets for the wheels. I’ll be getting the order in soon for the servos so we can assemble the rest of the legs with material on hand.