The Team Game Element is an optional Team designed and manufactured game element that will be used in the CENTERSTAGE game. The Team Game Element for the CENTERSTAGE game is called the Team Prop.

Team Props from the CENTERSTAGE Game Animation

See section 7.4 Team Game Element Construction Rules in the game manual part 1

  • <TE01> Inspection – Each Team Game Element must pass inspection
  • <TE02> Color Constraint – Teams will need two Team Game Elements (one red and one blue). Shades of red/blue allowed.
  • <TE03> Material Constraints – same as robot section 7.3.2, but no AprilTags, QR codes, Barcodes, etc.
  • <TE04> Size Constraints – minimum 3x3x3 inches, max 4x4x4 inches.
  • <TE05> Team Number – Team Game Elements must be labeled with their Team number
  • <TE06> Illegal Parts – electronics, robot parts in section 7.3, COTS game elements used in the current (and prior) season’s game. e.g. Signal Cone
  • <TE07> COTS Scoring Elements – The Team Game Element may not resemble any current season’s COTS game elements. e.g. pixels

Team Prop Q&A

See for official questions and answers about game rules. Here are some Team Prop questions and answers:

  • Q15 – light blue and pink are acceptable colors 
  • Q37 – orientation of team prop is not a precise activity
  • Q73 – the RM06 robot material rule applies to Team Props, therefore no Powerplay elements
  • Q77 – retroreflective tape not allowed per TE02 Rule
  • Q79 – solo cups and other disposable cups are allowed.
  • Q92 – placement of team prop is not a precise activity
  • other images such as team logos, sponsor logos, raised areas, cut outs, etc. are allowed, providing no other rule is violated.
    • Q192 – Team Prop Patterns
    • Q197 – Are team logos considered fiducial markers?
    • Q213 – Regarding Team Prop patterns

Team Prop Design

FTC-docs has a web page with suggestions on selecting objects for the Team Prop. Please read that section of the page carefully. They start by pointing out some quirks of TensorFlow:

  • TensorFlow is not so good with geometric shapes.
  • TensorFlow is good at pattern detection, ideally the surface of the object has a pattern to it.

They go on to suggest various shapes that could work well, like a chess rook or a chess knight. They suggest a couple of patterns like wrought iron fencing or leaf patterns. The idea is to create an object with some pattern or profile.

Here are some team prop objects that I’ve tested with.

Cup, Crown and Tower team props

The Crown and Tower team props are made from paper, you’ll probably want to make yours more durable or be prepared to replace them. Robots will drive over them or push them around during the match.

Two of these props were modeled after the chess rook example. The blue Cup is a plastic solo cup with the top lip cut off to get the size under 4 inches (and the cup lip was white and two colors are not allowed for a team prop). This is an image from an actual TensorFlow machine learning toolchain step where you label the objects in the image that you want TensorFlow to recognize. Note: captured with a Logitech C920 webcam that has a wide enough angle of view to see all three spike marks from the starting position.

The cup, crown and tower example props have the advantage that you won’t need to take extra videos that show different profiles of the prop. Note: just taking videos of the props on different spike marks and backgrounds should get some rotation of the prop so that the top points in the crown and tower have some variation.

The solo cup is a simple geometric shape and does not have any surface pattern and barely has any profile. It doesn’t follow the recommendations, but in testing it seems to be ok and blue and red solo cups are available. Likely you just need to trim off the lip of the cup (mine had a white lip and team props must be red or blue). It also needed to be trimmed to fit within 4″.

The other objects I tested with were made from LEGO Duplo. You probably need to glue the blocks together for actual use at an event as they come apart fairly easily. You could also use LEGO bricks. LEGO bricks and Duplo have natural surface texture with the knobs on the top surfaces. I then chose a shape that left a hole through one side and also a profile to the sides of the object. This is similar to the chess knight example. Props like these will require extra videos so that TensorFlow will recognize either orientation of the prop.

Team Props made from LEGO Duplo

Having designed with paper or LEGO you could then create a 3D model and use a 3D printer to create your team props. At that point you could even add a pattern or profile to the sides of the crown or tower props. The Crown and Tower models could use some surface texture in addition to the profile they have at the top.

Don’t forget that you need two props, one red and one blue.

Once you have your team props, the next step is to take videos of them. The videos are used as input to training a model.