Let's Make Robots!

Rotary walker entry for Rescue Robot Challenge

Navigates simulated disaster zone and picks up ping pong balls


            The premise of the Rescue Robot contest is to construct a remote controlled vehicle capable of navigating a playing field filled with obstacles in order to retrieve ping-pong balls.  The playing field is 18 feet by 10 feet and the obstacles that must be navigated include a 45-degree slope, an area filled with pea gravel, a 6-inch high wall, and a covered cave area.  Each contestant must collect four ping-pong balls from 4-inch tall pick pylons and place them all on a single drop pylon.  The contestant to move all four balls to the drop pylon first wins the round.     


Solution abstract

            My entry in this challenge has been scratch-built to meet all of the requirements in the most effective, yet inexpensive, way possible.  To traverse the difficult terrain my robot uses a versatile rotary walker locomotion method.  This allows my robot to have the same capabilities of a robot with treads without any added weight or worries about gravel jamming the treads.  The majority of the robot is made using strong, yet lightweight, PVC plastic, and this lightness also aids in getting over obstacles.  To navigate the “cave” my robot is fitted with a camera that transmits its image back to a real-time display.  This lets me fully control the robot even when visual contact is lost.  The final element of the robot is a versatile arm which allows the robot to easily pick up the ping pong balls yet still precisely set them down on the drop pylon.



            One key element to any robot is a versatile platform.  I have chosen not to buy a kit-based platform but instead build a custom one, which is incredibly powerful for its weight.  The construction is mostly PVC plastic, and in order to save weight the frame has holes cut out in areas not key to the structural integrity.  The platform uses a rotary walker mechanism for locomotion.  To my knowledge, there are no well documented robots that use this type of locomotion, and as a result a lot of work went into refining the basic idea before a working model could be built.  The advantages, however, were worth the time.  The rotary walker mechanism gives the robot a large surface area that has contact with the ground.  This spreads the robots weight evenly, allowing it to quickly cross shifting and unstable terrain such as the pea gravel in the playing field.  The extra contact area also aids in climbing up the steep incline on the course.  The robot also has a motorized “tail” which can lift the back of the robot up to help in climbing tall obstacles such as the 6-inch wall.  In addition to impressive capabilities, the platform is also simple and durable.  Because of the design, rocks and debris do not get jammed into the mechanism as they often do with treads.  Also, two motors drive each side, so if one experiences failure the other motor can still drive the mechanism without problems. 

            Mounted to the platform is an arm with an equal amount of innovative touches.  This arm can rotate a full 360 degrees, as well as raise and lower.  The servo that rotates the arm is geared down to allow for precise control, and the 4 joint mechanism which raises and lowers keeps the claw level with the ground at all heights.  To make the robot versatile, the rotating part of the arm contains mounting points that allow the existing claw to be easily swapped for a different one.  The claw for this robot is designed to be able to pick up a ping-pong ball easily without precise positioning, and yet still very precisely drop the ball on the place pylon.  It can do this because the claw opens in two ways.  One swings the claw wide open, which is ideal for picking up a ping-pong ball on a pick pylon because when it closes from this position any ball in the general area will be snatched.  The second way that the claw can open is a more gentle motion where each half of the claw opens slightly then swings up and out of the way.  This is ideal for placing the ball on the drop pylon without knocking all the other balls off.  The dual-opening mechanism is driven by two regular servos with just 90 degrees of rotation; it is a clever hinge mechanism that allows for a complex motion despite limited servo movement.               

            To control all of the robot’s functions a simple Play Station controller is used.  This is familiar to many people, meaning a complete novice can learn to accurately control the robot in a short time.  The real time camera allows the robot to be controlled out of sight and in tight places such as the “cave” where the robot is not visible to the operator.  The image from the camera is displayed on a portable color monitor.  A bright LED light allows the camera to still be useful when the area is dark, and a fish eye lens allows the operator to see more of the surroundings in the display.  To help make controlling the robot over the camera even easier a red colored LED shines on the area below the claw, so if a ping-pong ball is the correct distance away for retrieval the ball will be lit up red.  To power the robot a RC airplane lithium polymer battery is used, which is very lightweight.  One charge lasts about twenty minutes per battery, and because this robot uses a standard connector plug a higher capacity battery can easily by swapped in if more duration is required.  This robot is easy to operate, versatile, and costs under 300 dollars, which is significantly less then the competition.   

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Pretty cool :)

I can't stop thinking if it would not be relatively easy to make it work on water as well - you have these nice long light weight bars along the lifting surface.. Of course everything would have to be light weight, but still - It'd for sure beat the hell out of tank treads if it could "drive on water" as well!

Adding some foam tubes on his previous design would do the job .

Or he could add some extra wheels between those the design has already and do the same thing with foam tubes .


Very innovative design, great job !

I look forward to the next installment

How does it perform when one side spins forward and the other backwards? Sideways in large circles? Could you use this to be faster at alligning towards the target balls?

No, the result is the same as tank treads, it turnes in place.  To reduce toe in (the left and right sides driving in different directions when you try to drive forward) I have made the angle between the bars and wheels small.  This prevents any cool omni-wheel like maneuvers.   

Hmm..  Yes, I guess it's all about that angle.. And just to mention it in a brainstorm, it could be an option to change the angle, if the bars had some form of flexible mount.. You could go from one type of locomotion to the other.. and drive freakin fast to the sides in one state :)

Hmm.. There's some cool options with this, I like what you do, and thanks for sharing, it's quite inspiring :)

BTW; Why have you chosen to have the wide end as front?

That’s a really good question- because it does work fine both directions.  I went with that end because I thought it would give me more room to mount an arm up front.  Also that end is the heaviest- and that’s best for getting over the wall.  I am also thinking of an articulating front suspension that could help it get over obstacles, and that would need to face front to be effective.  Honestly in the end though I could easily change the weight distribution with my battery placement, and the arms holding the “front” wheels take up space, so its mostly because that end looks like the front lol…

That's funny, because I think the narrow end looks like the front - I see it like a spaceship with two Gatling guns or something like that :D

And.. I reckon (but may be wrong, pure guesswork) that it'd be better to scoop material inwards and attacking little rocks with the outer sides pushing you up on top of them as you slide over - than scooping outwards, and sort of "catching" stuff and push it out.

I am not sure, it just appears to me to be a more logic approach to negotiate small stones and rocks if you drive with the narrow end as front :)

But - it may very well be counter intuitive! I'd like to see an A/B test over the same terrain!



Hmmmm!! On the other hand - when driving with the wide end front, you meet obstacles with a wheel that either misses the obstacles, or lifts you over.. Hmm.. Now I actually think the wide end first - approach _is_the best!

Please find an uneven terrain, and do me an A/B test, please! :D Just switch it on for 20 seconds, and see which way it got further!