2.12 robot design: day one, arms, analysis, and the team [2.12 Design Notebook]
So today the goal is to create a design for a robot arm based on the constraints we have been given as designers for the robotic system. Before we get all the way deep into our own design, the 2.12 staff has some robots, standard robots they want us to analyze.
pset 3, part 2: term project arm design
We're going to do this on paper and just paste it here. But the main idea is this. We're going to examine the three types of arms, their work-spaces, benefits, and flaws and use this for part 2 where we recommend an arm design.
And some pictures of the little 5-bar linkage robot I made:
part 2: robot arm design recommendations:
So personally, after looking at the designs above I want to try a regular Articulate Robot Arm where the lengths L1+L2=L3, the reason for this is because this design seems the most versatile as far as reaching points easily goes. It will also be more natural for us to develop a throwing algorithm for this robot arm (as the robot arm has to throw bottles) because it is in similar form to a human arm.
Other things we want to do:
(1) Place all possible motors/weight at the base of the robot arm.
(2) Make the arm links as light as possible so their inertia may be neglected in controlling.
(3) Use smaller reductions/other brushless motors to make the response of the arm more dynamic if possible. The 100:1 harmonics aren't doing us much good to be honest.
(4) May be able to use things like pneumatics to make the arm more articulate and lighter.
This is a fun idea we would never do-an octopus robot arm. It pretty much has three motors and infinite degrees of freedom. It would be a cool experiment in controls and automation. Don't know how it would go though kind of risky. These are just initial ideas. But it could help us with the gripper design.
2.12 team meeting ONE:
Here's the notes from our first team meeting. Lab today has a few goals:
(1) Choose team leads
(2) Two arm designs
(3) Team name!!!!
Notes on Linear Z-Axis:
- Having a linear Z-Axis means you de-couple X&Y from the Z-Axis which also means it separates the pitch and roll from the YAW axis which means it should be easier to control - there's a case for a SCARA arm here.
- It shouldn't be horrible to get the arm to throw according to mason - no harder than it would be anyways.
So here are pros and cons:
So with an articulated arm, we would have a harder set of kinematic equations and X-Y-Z position are all dependent on each other. With a SCARA style we'd have XY positioning and then Z separately (as far as coordinates go). My only worry with a SCARA arm is how do we get the bottle to release accurately in the throwing - but that's more of a gripper thing than it is an arm thing. We're going to look at a few designs of arm.
Mason: "let's not make robot arms like snakes"
DESIGN ONE: SCARA Style Robot Arm
This is a 3DoF SCARA arm where there's two revolute joints and one prismatic. The Prismatic joint is at the end-effector in this design.
- There's a de-coupled XY and Z axis this makes kinematics and positioning easier
- This can go over other bottles on the field
- Need actuators at the end effector, this increases the inertia of the robot
**We could also put the prismatic axis at the base which means the whole arm moves up and down. But then it's harder to go over objects and it might not be such a rigid structure. But that's a separate design as well.
DESIGN TWO: ARTICULATE ROBOT ARM
There's also the articulate robot arm.
- Larger workspace (do we care?)
- Can put all motors at the base easily
- More human-like dynamics for throwing
- Kinematics is more challenging
- Orienting the gripper in 3-Space will be more challenging
So there's also the problem of how do we grip things. There's a few things we are looking at. First there's the UNIVERSAL GRIPPER (Coffee beans) I've heard this is not reliable and we don't want to go with that. Then there's suction gripper and finger grippers.
- Doesn't need to conform around shapes as much
- Usually only has one orientation
- Depends on the seal/the contact and if the seal isn't good we won't be able to pick up things
- Doesn't depend on seal
- It can wrap and squeeze to pick up objects
- They need to be able to wrap around the object
- Orientation of the gripper matters a lot for picking and placing
- More difficult to position --> but on an articulate arm it's not bad to position
IDEA ONE FOR SUCTION:
IDEA TWO FOR SUCTION:
GRIPPER DESIGN THREE:
So this is a 4-Bar linkage robot gripper, and there's pros and cons to this as well of course. So this is a very simple thing to build it has high grip and is difficult for things to fall out of it one it's gripped properly. The problem with it is the linkage needs to move outwards and inwards to grip the device which means your position changes and that needs to be accounted for. A prismatic gripper may be better?
GRIPPER DESIGN FOUR:
So this is a three-prong gripper. This gives us some more stability for grabbing the bottle but it might be very hard to ling up in 3-Space. We probably want to design a gripper with ONE degree of freedom. Really passive mechanism that just picks up the bottle from multiple orientations. We don't think the third point of contact is going to do much for us.
GRIPPER DESIGN FIVE:
So a compliant-wheeled gripper design will spin the wheel in and out to be able to pick up and place the bottle. This design is very passive and can be designed such that the bottle just gets sucked in really easily and line-ups are relatively easy. The problem for our application is that releasing the bottle when throwing is going to be very hard. It also just won't work with certain shapes and if the bottle is crushed or etc.
COMPILED NOTES SIMPLE:
- The granular jam/balloon gripper - having built one before it just doesn't work very well. You need some certain object shapes and to get the abilities you see, you need a lot of pressure.
- If we are doing a mechanical-style finger gripper we don't want three fingers that are symmetric around the gripper because it seems like the third finger wouldn't do much more than to mis-align the bottle instead of gripping it. We'd need something like a grabber where there's one DoF that just wraps around the bottle.
- If we were to use a wheel gripper, we could throw with the wheels, but we'd need to find a good way to grip more than one shape.
- The suction gripper is an option but it seems took risky given that it's going to be very hard to get a good seal on one of these things.
- Articulate robot arm with COMPLIANCE WHEEL Gripper
- SCARA robot arm with UNDER-ACTUATED Gripper w/ three fingers
THINGS TO LOOK INTO:
- Gravity compensation of robot arms like articulate robot arms (torsional spring/weights)
NOTES FROM GROUP DISCUSSION/PRESENTATION:
- Mason - is robot arm design lead
- Adi - is the gripper design/end-effector design lead
So I'm going to be looking into the ideas for a gripper design. Over this weekend we will be CAD-ing and finalizing designs for things like the design review coming up.