Frequently Asked Questions
In the interest of providing a fair competition environment, all questions posed to the judging staff will be posted here with answers so any team can reference the information. If questions contain unique team strategies or information, they will be genericized prior to posting.
How is the beginning of the ground robots' movement synced up with the beginning of an autonomous flight?
Team captains declare to the Judges when their system “goes autonomous”. That is when the run starts. The ground robots are started simulataneously by a line Judge via remote control so they will be moving by the time a vehicle's autonomous interactions are taking place in the arena.
Do the ground robots use a compass to rotate, or do they use some less accurate sensor? When the rules state that they will turn 180 degrees, will the turn be very close to actually being 180 degrees, or will friction make the turn more or less than 180 degrees?
Turns are based on time and wheel rotation, so while it turns out that it's pretty accurate, wheel friction is a factor. Robotic maneuvers are supposed to be somewhat random, and we in fact inject randomness into the trajectories on a periodic basis so that their trajectories can not be predicted in the long term. Also, collisions between robots (both target and obstacle) will introduce new random variables.
What portion of the top of the roomba (if not all) is tactile?
Approximately 16 square inches.
What force is required to activate the tactile sensors on the top?
Unknown, but small (simply the weight of any typical quad rotor used in the IARC would activate the switch — we will use a microswitch with a cantilevered paddle).
We read in the rules that a white circle will be put in the center of the field to set up the roombas.. will that circle remain through the 10 minutes? If so, how large is this circle?
Page 21 of the Official Rules says that the circle is 1 meter. As stated on that page, the purpose of the circle is to facilitate the initial robot placement and is not intended to be a navigation marker (although, if you can detect it, feel free to use it). The white circle will be thin and potentially “broken” because it is used to set up the target robots. Reference to the circle is really just to alert teams to the presence of “other features” that they might encounter on the floor, but is not specified as a reliable navigation feature (similar to the "dots" for the starting locations of the obstacle robots).
Does a team need to finish mission 7a to be eligible to compete in mission 7b?
Yes. Teams that can not even do the minimum mission would waste the time of the other teams in Mission 7b because they would be eliminated immediately.
Who is the point of contact for the Asia/Pacific Venue in 2016?
The point of contact for the Asia/Pacific venue in 2016 is GAO Rui. She can be reached via email at firstname.lastname@example.org or via via phone at +86 10-82315732 or +86 13-581856172.
Where can we get iRobot Create ground robots?
The iRobot Create is no longer in production and is being replaced by the Create 2. The interface is different form that of the Create, but it can still be programmed to perform to the IARC-specified behaviors. The following links provide some information about the Create 2:
Are teams allowed to operate more than one robot (at the same time) as part of their entry?
The intent is that there be only one (1) aerial robot during Mission 7a. The reason is that while swarms are compelling, we really want to push the "behavior" and "planning" functions which require the aerial robot to prioritize its actions. With more aerial robots, this workload can be divided (e.g., 4 aerial robots could just divide up the arena into 4 sectors and concentrate only on those ground robots coming into their sectors). That would reduce the workload on each aerial robot.
Are teams allowed to operate a moving ground robot as all or part of their entry?
Are teams allowed to land in the arena to block the progress of the ground robots?
Yes, landing is acceptable. However, the IARC is an "aerial robotics competition", not a "ground robotics competition". As such, the "aerial robot" must principally fly. While it is acceptable to briefly land to block ground robots, or to even move along the ground temporarily to reorient itself, locomotion around the arena must be by means of out-of-ground-effect flight. It is clear from the definition of "AIR VEHICLE" found on page 11 of the Official Rules, that the "air vehicle" must fly the majority of the time. Specifically, the Official Rules state, "“Air Vehicles” are considered to be those capable of sustained flight out of ground effect while requiring the earth’s atmosphere as a medium of interaction to achieve lift (as such, pogo sticks and similar momentary ground-contact vehicles are not considered to be flying air vehicles)." Teams with robots that are designed so that sustained flight is not possible for at least 10 minutes in the opinion of the Judges, or aerial robots that spend an inordinate amount of time involved in ground locomotion in the opinion of the Judges, will be eliminated from the IARC.
Are teams allowed to (intentionally) have anything drop off of their robots (onto the ground and/or the ground robots, that move, or are fixed)?
The rules indicate that there will be "at least three competition cycles". What does that mean? This seems to imply that it is not possible to complete Mission 7 until 2016.
Exactly "when" Mission 7a transitions to Mission 7b will be the call of the Judges and the Organizers… Mission 7b can not begin until there are enough teams that accomplish Mission 7a. Logistically, this is a planning problem, but the Organizers feel that Mission 7a will not be totally completed by many teams in the first year. From past experience and familiarity with the technologies involved, we would not expect to have teams showing sufficient proficiency until competition cycle 3 (2016), so the earliest that Mission 7b would be considered would be 2017, but this is just an estimate. As a minimum, two teams would be necessary for the conduct of Mission 7b, but the desire of the Organizers is to see several teams reach this level of capability (at least 4 teams).
Will there be a circle painted in the center of the arena?
To facilitate the placing of the ground robots at the beginning of a run, a circle of start positions will be marked at the center of the arena for use by the Judges. Do not depend upon these markers for navigation because they will intentionally be subdued.
For the four moving pylons, is the center of their circular motion specified or random?
The initial center of rotation for the moving obstacles is the geometric center of the arena, however, due to floor friction and other factors, the center of rotation of the moving obstacles can drift unpredictably.
What happens if the pylons hit each other?
Obstacles will be programmed to attempt to continue their circular paths. Collisions between obstacles and ground robots will normally result in the ground robot reorienting its own trajectory. Collisions between obstacles will become a shoving match until they are free to move unimpeded. Simulations show that collisions between obstacles is unlikely in the 10 minutes allotted to each run.
Will the ground texture be the same in both competition venues?
This can not be guaranteed, so that is why an update was issued to the rules to better specify the arena with its one meter white grid. The interior of the one meter grid squares will be unknown to the teams except that it will not be the same color as the grid lines. With the presence of the grid lines, optical odometry can be used, or optical flow can be used with both the grid lines and the patterns appearing inside the squares.
Regarding flight outside the boundary (or boundary + 2m), does this occur if any part, the center, or all of the vehicle goes outside the limit?
This is a Judges' call. Going outside the boundary is not bad if it seems to be accidental. Staying outside the boundary is undesirable. Going far outside the boundary is a potential safety issue, so in either case, the line Judge will subjectively decide if the run should be terminated or allowed to continue. The precise amount that that a portion of the air vehicle exceeds the boundary will be at the discretion of the Judges.
Can the aerial robot lift the ground robots and carry them directly to green boundary?
No. This is entirely against the spirit of what we are trying to accomplish with Mission 7 where the aerial robot is supposed to exhibit "herding behaviors" over a field of somewhat randomly moving ground objects (robots). Picking up and moving the ground robots to the green line is actually an easier behavior to demonstrate than to have to herd them in the presence of unpredictable factors that create "trajectory noise". Also, we have demonstrated the ability to autonomously acquire and move objects in an arena in Mission 1, 2, and 6, so that is not something that will further advance the state of the art in aerial robotics. Further, Mission 7b would become technically less interesting were the aerial robots allowed to shuttle the ground robots in the air.
What are the exact dimension of the moving ground obstacles?
The obstacles are of unknown dimensions and my be of different sizes on each of the four obstacle robots. All that you know is that the obstacles are cylindrical and no taller than 2 meters in height as per the rules. Your obstacle avoidance system needs to be able to see and avoid the obstacles. “Seeing” them implies that you can detect them.
Will the robot reset the 20 second counters by colliding with another robot?
No. The 20 second timer runs independent of any external stimuli.
Does the robot always turn clockwise upon top contact or collision?
Trajectory noise could result in a small turn in either direction from time to time. When collisions occur, or if the robot is contacted by the aerial robot on top, it will always turn clockwise (either 180° or 45° depending upon the source of contact).
Will the obstacle robots change their orbit after collision with other robots?
The obstacle robots will not intentionally change their circular trajectories after a collision, but it is very possible that the collision can cause an offset in the circular center point so that the obstacle robot begins to orbit a slightly different point (other than the center of the arena). In most cases, collisions will be brief and the target robots will quickly move away from the obstacle robot, but in the event that there is a "90° side" collision, it is possible that the center of rotation of the obstacle robot could shift.
Can anything hang from the vehicle on a cord to interact with the robots or does it have to be attached?
Something (e.g. a weight on a cord or string) can hang below the air vehicle provided it's intent is to interact with the ground robots' top touch paddle. The remainder of the air vehicle must still fit within the size constraints dictated in the rules.
Do we have to trigger the tiop-touch sensor to adjust the ground robot trajectory? Could we do so solely by hitting the bumper and navigate to the goal?
You could influence the trajectory of the ground robot solely by using the collision bumper if you desired. You would still need to do this with an "aerial robot" and not a ground robot that just flies into the arena and spends the rest of the time running around as a ground robot (that would not be in the spirit of what we’re trying to demonstrate in Mission 7).
When will the ground robot be removed from the arena; when any part of it reaches the boundaries of the arena or when the center crosses? How quickly will it be removed after crossing the boundary?
As soon as any part of the ground robot crosses the edge of the boundary. A Judge or staff member would take the robot out of service immediately as it crosses the boundary because in all likelihood, the Judge or staff member would be anticipating the robot going out of bounds and would be near it when it goes out of bounds. The only thing that could delay that would be if an aerial robot were flying in the vicinity and it would be safer to let the robot continue out of bounds to be retrieved later outside the arena. We don’t want to confuse the aerial robots, so Judges and Staff would likely maintain a reasonable distance from the aerial robots. Remember, the arena is rather large. Aerial robots should be not only “bullet proof” from an RF interference standpoint, but they should also be immune to activity outside of the boundaries of the arena (flash bulbs, people moving, sounds, etc.) The only people that should be anywhere near the boundaries would be Judges (3), Staff (3), and the Safety Pilot (1).
Will there be any identifying colors on the ground robots?
Half (5) of the target robots will have the top plate painted with Krylon "Gloss Emerald Green #52016” (RGB = 22, 86, 16) and the other half (5) will have the top plate painted with Krylon “Gloss Banner Red #52108” (RGB = 176, 22, 4). Obstacle robot top plates will be white.
Can the Obstacle Robots be eliminated by blocking them or knocking them over in the arena?
The Official Rules state that, “Should impact occur between the aerial robot and one of the cylindrical obstacles, the run will be terminated.” Here, “cylindrical obstacles” refers to the “Obstacle Robot” as a whole, not just the cylindrical extension above the mobile platform beneath.
Is it possible, or is it against the spirit of the competition, to use the infrared technology to map the space around our drone?
You map anything you can see at any wavelength you want so long as it poses no threat to the Judges, Staff, or bystanders. If you are considering something like a SLAM laser, please be aware that the arena has no 3D features and it will be in a large indoor area like a basketball court or stadium.
Will there be any spectators or visitors near the competition arena? Will the competition take place in a closed space or an open space?
Judges and Staff will be present near, but outside the arena. Spectators will be kept at a safe distance from the arena. The arena has no 3D features and it will be in a large indoor area like a basketball court or stadium.
Will the ground robots use infrared sensors or will they be pre-programmed to move randomly?
The target and obstacle ground robots are pre-programmed and do not use their infrared sensors. Look at the team resources pages to see exactly how the robots will be configured and programmed. The code is provided there.
Will the grid lines be 8cm wide just like the side lines?
This depends upon the duct tape that we can find locally. Don’t key off of the actual exact thickness of lines. The line width is a rough estimate. Even the colors are relative. What is “Green”? What is “Red”? How white is the “White”? Similarly, the line width could be 7cm or 9cm. The resolution of these things should be accounted for in the robustness of your design. Think of it as a human doing the same task. Is the human going to care whether the Green line is 7cm or 8cm? Rather, the human will identify the line as “Green” rather than “Red” or “White” and that it is indeed a “wide" line and not a piece of string on the floor. Your aerial robots should perceive things robustly in this way.