KnuthLab Exploration Rover with Researchers A. Fischer and N. Malakar

The LEGO Exploration Rover is powered by six NXT Standard Motors in a Rocker-Bogie suspension system used in all of the NASA Mars rover designs. The rover is approximately 1.5 ft high with a 1 ft x 1.5 ft base. It is larger than the NASA Sojourner Rover, which was part of the Pathfinder Mission to Mars in 1997, and smaller than the Mars Exploration Rovers Spirit and Opportunity. It can safely carry a payload of 8 pounds.

KnuthLab LEGO Exploration Rover

The white frame mounted on top of the rover is the Bayesian Vision-Based Navigation System being developed by Autonomous Exploration Inc. for NASA.

Check back, as we will be posting videos of its operation and discussing some of the important design features.

They are:
1. LEGO Star Wars Obi-Wan’s Jedi Starfighter (10215)
2. LEGO Cars Big Bentley (5828)
3. LEGO Star Wars Snow Trooper Army Pack (8084)
4. LEGO Mindstorms NXT 2.0 (8547)
5. LEGO Star Wars Clone Trooper Battle Pack (7913)
6. LEGO City Police Helicopter (7741)
7. LEGO City Fire Truck (7239)
8. LEGO Green Building Plate (10″ x 10″)
9. LEGO Harry Potter Quidditch Match (4737)
10. LEGO Architecture White House (21006)
11. Lego Star Wars Millennium Falcon (7965) – 2011 Release
 

New Rotocaster Omni-Directional Wheel

While these are often used on robots with three wheels where all three rotate at different rates allowing the robot to go in any direction,

Three-Wheeled Omni-Wheel Prototype by Xander Soldaat at Bot Bench (click on image to see more)

There are several different options that one can consider. LEGO purists may consider making their own omni-directional wheel designs out of only official LEGO parts. Another option is to purchase manufactured omni-directional wheels. These come in two classes: those that are designed to be LEGO compatible, and those that are not LEGO compatible. In the latter case, one would have to construct some kind of coupling mechanism to enable the wheel to connect to LEGO parts.

Here are some of the options that I have found.

Omni-Directional Wheels Constructed from LEGO Parts

Omni-Directional LEGO Wheels by Xander at Bot Bench
These wheels use the LEGO pulley to get a six-fold symmetry. Each “wheel” uses 12 LEGO wheels. For archiving purposes, here are three pictures. Please visit Xander at Bot Bench for larger images and more details.

Omni-Directional LEGO Wheels by Xander Soldaat at Bot Bench

Details of Omni-Directional LEGO Wheels by Xander Soldaat at Bot Bench

Details of Omni-Directional LEGO Wheels by Xander Soldaat at Bot Bench

This amazing tank by Peer Kreuger also sports omni-directional wheels made from LEGO parts

LEGO Tank with LEGO Omni-Directional Wheels

The Omni Bot v2 by jason701802 also sports omni-directional wheels

Omni Bot v2 by jason701802

LEGO-compatible Manufactured Omni-Directional Wheels

This LEGO-compatible wheel is made by the School of Robotics

School of Robotics Omni-Directional Wheels

Rotacaster makes a LEGO-compatible omni-directional wheel:

Rotacaster Robot Wheel (LEGO-compatible)

Rotacaster is also coming out with a new design

New Rotocaster Omni-Directional Wheel

Rotacaster also produces several non-LEGO compatible designs.

Tetrix also makes LEGO-compatible Omni-Directional Wheels

Tetrix Omni-Directional Wheels

General Manufactured Omni-Directional Wheels

Mecanum Omni-Wheel Designed by Bengt Ilon in 1973 at the Mecanum AB Company

Mecanum Wheel

Mecanum Omni-Directional Wheel

Vex Omni-Directional Wheel

Vex Omni-Directional Wheel

Damon WH-01 Omni Directional Wheel (hexagon hole)
(Outer Diameter:60mm, Inner Diameter:11mm, Material: Nylon)

Damon WH-01 Omni-Directional Wheel

Kornylak Omni-Directional Wheel

Kornylak Omni-Directional Wheel

Kornylak Transwheel

Kornylak Transwheel

Kornylak Omni-Directional Wheel

Kornylak Omni-Directional Wheel

Here are the MATLAB packages that I know of.  The first two are specifically geared toward the NXT; whereas the last is a general robotics package.

• LEGO MINDSTORMS NXT Toolkit for MATLAB and Simulink
  http://www.mathworks.com/programs/mindstorms/
• Robotics Toolbox for MATLAB (Release 7.1) (P.I. Corke)
  http://petercorke.com/Robotics%20Toolbox.html
• RWTH Mindstorms NXT Toolbox for Matlab
  http://www.mindstorms.rwth-aachen.de/

I have finally compiled building instructions for my Little Rover, which can be seen above in a 3D Rendering courtesy of POVRay.  An earlier version of this rover can be seen in this YouTube video:

Little Rover Prototype Video

Rover Design

The complete detailed building instructions can be found here in this 94-page pdf file.
Warning: it is about 9MB in size.  The design is not entirely compatible with the standard NXT Mindstorms Kit.  This design relies on two touch sensors, several 1×9 bent liftarms, and as far as I can tell from Peeron, the NXT Kit has only two.  This may require a little redesign.  Other compatibility issues and their solutions can be found in the Parts List in the instructions.

Remember to download the software DriveSmart here as well.
Installation instructions can be found in the zip file.

DriveSmart Code

The main file is called DriveSmart.rbt.  Drive Smart runs four threads:

Drive Thread
The Drive Thread (lowest one of the four) drives until a warning flag is set by one of the other
threads. It then waits until it gets an all clear message via the Wait Until Free block, and then
it starts driving again.

Bumper Threads
There are two threads that monitor the bumpers.
The reaction is only activated if nothing else is currently commanding the robot.  In this case the
bumper has been pressed and the robot will veer away from the hazard.

Ultrasound Thread
This thread monitors the ultrasound rangefinder.
The reaction is only activated if nothing else is currently commanding the robot.  When the robot
comes too close to a hazard, the robot is commanded to stop.  It then looks both ways and then turns
in the direction with more room.  If the robot is within 10 cm of a hazard on both sides, it then
backs up.

The robot can roam about a wide variety of rooms and not get stuck.
He does not detect stairs though!  So be careful.

Download: instructions and code.

Enjoy!
Kevin Knuth

The NXT cable has six wires. Below I list a table with the wires and their colors:

The WHITE and BLACK wires (Motor 1 and Motor 2) deliver power to the motor.
If standard batteries are used, the potential difference will be 9 volts, otherwise the NiMH rechargeable batteries provide 7.2 volts. If the white wire is positive and black is negative, the motor will turn one way. If you reverse the polarity, the motor will turn the other way.

The RED wire is connected to the ground (GND). Note that in the sensors, RED and BLACK are connected to one another. This is not the case in the motors.

The GREEN wire is connected to the +4.3 NXT power supply.

The YELLOW and BLUE wires are connected to the quadrature encoder, also called an incremental rotary encoder.

As shown in the figure from Wikipedia above, (http://en.wikipedia.org/wiki/Quadrature_encoder) the wires return square wave pulses that are 90 degrees out of phase. If the rising pulse on TACH00 leads the rising pulse of TACH01 by 90 degrees, then the motor is going forward. If it instead lags by 90 degrees, the motor is rotating backwards. One complete square wave cycle corresponds to 2 degrees of rotation. In the diagram above, if TACH00 refers to A and TACH01 refers to B, we can see that the motor is going backwards as TACH00 is lagging TACH 01.

By measuring the frequency of the square wave oscillation, one can compute the rotational velocity. Since one cycle corresponds to 2 degrees of rotation, one cycle per second (1 Hz) corresponds to 2 degrees/sec. If you record a frequency of X Hz, then the rotation rate is 2X cycles/sec.

Note also that by tracking both square waves, you can identify quarter cycles, which gives you a resolution of 1/4 of 2 degrees, which is 0.5 degrees.

The motor speed is controlled by pulse-width modulation (pwm), which works by driving the motor with a variable duty cycle square wave. This effectively turns the motor on and off, fast. The longer it is on, the more torque it will generate and the faster it will go.

These details and more can be found in the excellent book: Extreme: NXT with a sneak peak here.

Additional details can be found in the excellent book Extreme NXT: Extending the LEGO MINDSTORMS NXT to the Next Level (Technology in Action) by Michael Gasperi, Philippe E. Hurbain, and Isabelle L. Hurbain.

Philo uploaded a comment, and reminded me that “Note that there are some internal photos of the NXT motor here: http://philohome.com/nxtmotor/nxtmotor.htm and schematics here: http://www.brickshelf.com/cgi-bin/gallery.cgi?i=1846577”

Happy Hacking!

So for those interested in some LEGO electronics hacking, here is a list of supplies that will get you up and running fast for about $275… just a but more than the cost of a single Mindstorms kit.  Plus you’ll now get to learn electronics!

First, check out the book:
Making Things Talk: Practical Methods for Connecting Physical Objects

This book explains how to wire, program and interconnect various microcontrollers, some of which are very closely related to those used by the NXT Brick.

Supply List

Item Number	Description	Quantity	Unit Price	Total
Amazon.com
	Making Things Talk	1	$19.79	$19.79
Jameco.com
19166	Desoldering Pump	1	$4.95	$4.95
159291	Wire Stripper	1	$10.15	$10.15
161411	Diagonal Cutter	1	$7.49	$7.49
35474	Needlenose Pliers	1	$5.49	$5.49
127271	Mini Screwdriver	1	$1.89	$1.89
681002	Helping Hands	1	$8.75	$8.75
159611	Power Connector	2	$1.79	$3.58
10444	Alligator Test Clip Leads	2	$4.39	$8.78
103377	Header Pins	10	$0.16	$1.60
119011	Push Button (PCB Type)	10	$0.27	$2.70
29082	Potentiometer	2	$1.05	$2.10
242115	LM1117T-3.3 Voltage Regulator	3	$1.39	$4.17
51262	7805T 5v Voltage regulator	3	$0.32	$0.96
38236	2N2222A Transistor NPN	5	$0.41	$2.05
32993	TIP120 Power Transistor	5	$0.45	$2.25
643488	3.3V Zener Diode	5	$0.03	$0.16
35991	1N4004 Diode	5	$0.04	$0.20
152792	LED Yellow	10	$0.17	$1.70
152805	LED Red	10	$0.21	$2.10
153139	LED Orange	10	$0.35	$3.50
156962	LED Green (567 nm)	10	$0.20	$2.00
334529	LED Bargraph Red	1	$1.31	$1.31
334537	LED Bargraph Yellow	1	$1.23	$1.23
334511	LED Bargraph Green	1	$1.28	$1.28
17187	7-segment LED Display	3	$0.88	$2.64
38818	4-switch DIP	4	$0.48	$1.92
38842	8-switch DIP	2	$0.89	$1.78
103166	Resistor Refill	1	$12.95	$12.95
15270	0.1 uF	10	$0.15	$1.53
94161	1 uF	10	$0.12	$1.20
29891	10 uF	10	$0.06	$0.60
158394	100 uF	10	$0.11	$1.08
MPJA
4443 TE	Solderless Breadboard	1	$4.95	$4.95
4447 TE	Large Solderless Breadboard	1	$22.95	$22.95
7027 TE	Jumpers	2	$3.95	$7.90
14213 TE	Digital Multimeter	1	$14.95	$14.95
15860 TL	Mini Soldering Station	1	$14.95	$14.95
Sparkfun
Wiring Platform	DEV-00744	1	$84.95	$84.95
Radio Shack
64-025	Lead Free Solder	1	$3.89	$3.89

Note that the light gray items are optional, and not necessary.

Also, this list does not include some sort of power supply. Pulling one out of an old computer is an easy option. Or rechargeable batteries work well too (in which case you will need battery holders).

Last, there are special items in the book Making Things Talk that you may decide to purchase separately, such as flex sensors, or bluetooth boards, etc.

You can store your electronics in much the same way you store your small LEGO parts. Check out the article on Storage.

Enjoy Hacking!

There are several ways to find the Center of Mass of the motor.  The most straightforward way is to hang the motor from an axle placed in one of the holes.  The motor will orient itself so that the Center of Mass lies directly below the axle.  By hanging a mass on a string from the axle, the Center of Mass must lie somewhere along the line defined by the string. 

After performing this experiment, I placed a small piece of Scotch tape over the string so that I can keep track of where that line is.  I then cut the string off of the axle.

Now to find the precise point, we simply perform the experiment again, but place the axle through a different hole.  This gives us a second line.  Since the Center of Mass must be on both the first line and the second line, it is located at the intersection of these two lines.

The Center of Mass is very close to being aligned with the holes on the motor.  Below is an MLCAD image of the NXT motor (from Philo).  I have overlayed a Cartesian coordinate system that corresponds to that used to define the 3-D image file.  The origin of this system is at the center of the axle hole on the motors drive axis.  This is perfect for me since I will be rotating the motor and trying to compute the position of the Center of Mass after the motor has rotated through some arbitrary angle.

This image not only helps with identifying the Center of Mass of the NXT motor, but also in understanding the dimensions of the NXT motor overall.

Kevin Knuth
Albany NY

The LEGO Mindstorm NXT robotics system is an excellent testbed for research in machine learning and artificial intelligence.  At Knuthlab Robotics at the University at Albany, we are developing intelligent instruments using LEGOs.

Our first instrument is a robotic arm that is designed to locate a characterize a white circle on a black background using the LEGO light sensor.  It relies on Bayesian inference, which is implemented using a technique called Nested Sampling, which was developed by John Skilling.  This software allows the robot to learn the characteristics of the circle using the light sensor data that it has collected.  The real advance here is the inquiry engine, which uses Bayesian adaptive exploration to decide which measurements to take next.  It does this by considering all the possible measurements that it could take, and computes the expected gain in information from each possible measurement.  It then chooses to take the measurement with the greatest expected information gain.  The process then repeats as the robot learns about the circle.

The system is easily generalized to solving other problems, such as exploring rooms, interpreting people’s emotions, and doing real science.

We recently presented our research at the MaxEnt 2007 workshop in Saratoga Springs NY.  Below are links to a video of the talk, my slides, and our research paper.

Video: Designing Intelligent Instruments, K.H. Knuth

Slides: Designing Intelligent Instruments, K.H. Knuth

Research Paper:
Knuth K.H., Erner P.M., Frasso S. 2007. Designing intelligent instruments. K.H. Knuth, A. Caticha, J.L. Center, A. Giffin, C.C. Rodriguez (eds.), Bayesian Inference and Maximum Entropy Methods in Science and Engineering, Saratoga Springs, NY, USA, 2007, AIP Conference Proceedings 954, American Institute of Physics, Melville NY, In press.