Future TomBot Articulations

Tags: think and control
Personhours: 3
Future TomBot Articulations By Cooper

Task: Plan out potential robot articulations to improve game strategy

Getting back from the tournament, we were able to immediately start to think about what was the big problems and possible improvements to the articulations of the robot. Overall, we ended up coming up with several ideas, both for fixing things and for efficiency.

1- Turntable Articulations

In the competition, we realized the extreme convenience that having some articulations for the turret. Not to say that we hadn't tried to make them before the competition, we were having some issues writing them. plus, even if we didn't have those convene, it would have been improbable that we would have gotten them tuned for the competition. Anyways, even though we agreed on needing to have these presets, we could not agree on what they should be. One argument was that we should have them field-centric, meaning that it would stay in one position from the POV of the audience. This was cited to have a good number of use cases, such as repetitive positions, like the left/right and forward of the field. However, another idea arose to have them be robot-centric. This would allow for faster relative turns.

So, what we've decided to do is write the code for both. The field-centric will be turns and subsequent static positions will implement the IMU on the control hub mounted to the turret. The robot-centric version will be based on the tick values of the encoders on the turret's motor. Then, we will have the drivers choose which one they prefer. This we believe is effective, as it will allow for a more consistent use of the turntable for the driver.

2- Move to Tower Height Articulations

this is one of the more useful Ideas, which would be to extend the arm to the current height of the tower. How would we do it? Well, we have come up with a 2 step plan to do this, in different levels of difficulty. The first one is based on trig. We used the second controller to increment and decrement the level of the current tower. That value is then used in the extendToTowerHeight() method, which was written as the following:

public void extendToTowerHeight(){ hypotenuse = (int)(Math.sqrt(.25 * Math.pow((currentTowerHeight* blockHeightMeter),2)));//in meters setElbowTargetPos((int)(ticksPerDegree*Math.acos(.5/ hypotenuse)),1); setExtendABobTargetPos((int)(hypotenuse *(107.0/2960.0))); }

As you can see, we used the current tower height times the height of a block to get the opposite side of the triangle relative to our theta, in this case the arm angle. The .25 is an understood floor distance between the robot and the tower. This means that the arm will always extend to the same floor distance every time. We think this to be the most effective, as it means not only that the driver will have a constant to base the timing of the extension, but we minimize the amount we have to extend our arm. If we assumed the length of the hypotenuse, there would be overextension for lower levels, which would have to be accounted for.

The next phase of the design will use a camera to continue to extend the arm until it doesn't see any blocks. not only will this allow for a faster ascension and more general use cases, It will eliminate the need for a second controller (or at least for this part.

3-Auto-grab Articulation

Finally, the last one that we came up with is the idea to auto-grab blocks. To do this we would use vison to detect the angle and distance that block is away from the robots back arm and extend to it. Then rotate the gripper, snatch it and reel it back.

Next Steps

Use a culmination of drive testing and experimentation to refine the robots movements and ultimately automate the robot’s actions.

Date | December 7, 2019