mecanum drive

We recommend building a mecanum drivetrain using 3 to 4-inch diameter wheels with ball bearing supported rollers, arranged so that the rollers form an "X" shape when viewed from above. According to Game Manual 0, this specific wheel orientation allows the robot to turn significantly faster because the forces applied to the ground act tangent to the turn radius, while the ball bearings minimize roller friction to optimize strafing performance. Game Manual 0 also advises sticking to 3 to 4-inch wheels because larger 6-inch wheels are more expensive and do not offer significant performance advantages .

When purchasing wheels, we recommend selecting either the 96mm goBILDA Mecanum Wheels v2 or the 75mm REV Robotics Mecanum Wheels, as Game Manual 0 identifies these as the gold standard due to their ball bearing supported rollers and excellent mounting options. Conversely, teams should avoid bushing-based options like TETRIX or fragile wheels like the AndyMark Standard Duty, which Game Manual 0 notes perform poorly and barely strafe .

While a mecanum drive offers fantastic agility and the ability to strafe in any direction, it suffers from a lack of traction and pushing power. As explained in Game Manual 0, the diagonal rollers are prone to slipping, which reduces speed during strafing and makes the robot vulnerable to being pushed around by defensive tank drives. If a game features obstacles or terrain that must be traversed, or if heavy defensive play is expected, teams should pivot away from mecanum to a high-traction drivetrain .

Because of the inherent wheel slippage, teams using mecanum drives should implement a separate dead-wheel odometry system for accurate autonomous tracking rather than relying on built-in motor encoders. Game Manual 0 notes that drive encoder localization on mecanum drives is highly inaccurate and prone to drift during high acceleration. For teleoperated control, Game Manual 0 recommends using a mathematical scaling method to normalize motor powers to the -1 to 1 range, maintaining the correct power ratios even when inputs exceed maximum thresholds .

For teams utilizing FTCLib, the library provides dedicated guides on Mecanum Drive Kinematics (https://docs.ftclib.org/kinematics/wpilib-kinematics/mecanum-drive-kinematics) and Mecanum Drive Odometry (https://docs.ftclib.org/kinematics/wpilib-kinematics/mecanum-drive-odometry) .

If you are using the REV DUO system, the REV Robotics Docs offer tutorials on simplified mecanum programming (https://docs.revrobotics.com/duo-control/hello-robot-blocks/part-4-going-beyond/programming-mecanum-simplified), refined programming (https://docs.revrobotics.com/duo-control/hello-robot-blocks/part-4-going-beyond/programming-mecanum-refined), and physical wheel setup (https://docs.revrobotics.com/duo-build/ftc-starter-kit-mecanum-drivetrain/mecanum-wheel-setup-and-behavior) .

Was this right?

Sources16

Holonomic Drivetrains > Mecanum Drive

Game Manual 0·CC BY-NC 4.0

Mecanum drivetrains consist of four mecanum wheels which are powered independently by one motor. This configuration angles the velocity of each wheel, allowing the robot to strafe.

Term: Mecanum Wheel — Mecanum wheels are a special type of wheel that enable maneuverability and holonomic strafing as opposed to traditional wheels. They consist of a series of rubber rollers rotated 45 degrees to either the left or right.

In a conventional mecanum drivetrain, running the wheels on one diagonal in the opposite direction to those on the other diagonal causes sideways movement. Combinations of these wheel motions allow for vehicle motion in any direction with any vehicle rotation (including no rotation at all).

The primary advantage to mecanum drive is the maneuverability it affords, especially because the robot can strafe instead of turn and drive. The rollers on mecanum wheels form a 45 degree angle with the wheel's axis of rotation, which means that mecanum drivetrains can't strafe as fast as they can drive forward.

This can be explained by discussing the forces involved. When each wheel rotates, it applies a friction force to the ground, which moves the robot. When moving forward, both sets of left wheels rotate in the same direction at the same speed, and both sets of right wheels rotate in the same direction at the same speed, meaning that the forces do not oppose each other. However, when strafing, neither the two left wheels nor the two right wheels are rotating at the same speed. In many cases, they even rotate in opposite directions.

These two opposing forces cause the rollers to slip more and more, which angles the robot's velocity at the expense of traction (more slipping results in a loss of speed). However, the wheels do still slip when moving forward but not as drastically as they do when strafing.

This is the primary disadvantage to mecanum drivetrains: they tend not to have much pushing power and thus, are vulnerable to defense by a sturdy tank drive.

Due to the fact that mecanum wheels are more likely to slip because of the diagonal rollers, an optional addition to mecanum drives is a separate odometry mechanism in order to track the robot's location during autonomous.

Attention: It is important to note that in order to maximize the efficiency and stability of mecanum drives, when viewed from above, the rollers of each wheel should point towards the center of the robot, forming an X shape, rather than a rhombus.

The primary reason for this is that it allows the drivetrain to turn significantly faster than it would otherwise be able to. When using the suggested setup, when viewed from the robot's underside, the rollers form a rhombus. This allows the force applied by the wheels on the ground to act tangent to the turn radius, leading to faster turning.

See this video (https://www.youtube.com/watch?v=xgWf_t8owI0&t=3152s) and this other video (https://youtu.be/YJaX3vQ6kHw?t=123) for a more in depth explanation.

Open Game Manual 0

Holonomic Drivetrains > Mecanum Drive > Advantages

Game Manual 0·CC BY-NC 4.0

Fantastic maneuverability and agility due to strafing, can avoid defense very well
Good acceleration, can have high top speed
Very versatile drivetrain for nearly any game

Suffers in traction, as mecanum rollers have a lower coefficient of friction than traction wheels; cannot traverse terrain
Able to be pushed around on defense
Wheels must be powered independently, so there is no redundancy

CAD Examples of Mecanum Drive (Click to expand) goBILDA COTS Click here to explore this model in OnShape online CAD (https://cad.onshape.com/documents/45549489f570f3694569a2df/w/85ff26b9fca4988ebc4df3b4/e/df9bcea72fcdd7e4aee4134b)

REV COTS Click here to explore this model in OnShape online CAD (https://cad.onshape.com/documents/45549489f570f3694569a2df/w/85ff26b9fca4988ebc4df3b4/e/b8d4a6cfe4f26c5170f40a6c)

goBILDA Custom Click here to explore this model in OnShape online CAD (https://cad.onshape.com/documents/45549489f570f3694569a2df/w/85ff26b9fca4988ebc4df3b4/e/cc2df2fbce0e9ca393696b54)

REV Custom Click here to explore this model in OnShape online CAD (https://cad.onshape.com/documents/45549489f570f3694569a2df/w/85ff26b9fca4988ebc4df3b4/e/40c3266dab2e444011cce79c)

(Figure: Configuration for mecanum wheels, courtesy goBILDA)

Open Game Manual 0

Holonomic Drivetrains > Mecanum Drive > Mecanum Wheels Miniguide

Game Manual 0·CC BY-NC 4.0

One of the most important features of a mecanum wheel is the mechanism that allows the roller to spin---either bushings or ball bearings. Mecanums which use ball bearings strafe better than those which use bushings, as the rollers can spin with less friction.

Note: In the past, some teams have invested in 6 inch diameter mecanum wheels. These are generally much more expensive and generally do not provide significant performance benefits. We highly recommend sticking with wheels that are between 3 and 4 inches in diameter.

Open Game Manual 0

Holonomic Drivetrains > Mecanum Drive > Mecanum Wheels Miniguide > Recommended

Game Manual 0·CC BY-NC 4.0

These wheels are the gold standard of mecanum wheels, if you are buying new mecanum wheels there is very little reason to buy any mecanum wheels not from this section.

goBILDA Mecanum Wheels v2 (https://www.gobilda.com/96mm-mecanum-wheel-set-70a-durometer-bearing-supported-rollers/) ($127.49 with team discount): The goBILDA v2 Mecanum Wheels are the gold standard for teams wanting a quality design in all aspects. They feature a 96mm diameter and a thickness of 38mm, with ball bearing supported, 70A durometer rollers that provide excellent strafing performance. With the use of recessed 16x16mm and 32x32mm holes, these wheels have the largest amount of mounting options and can fit all shaft styles commonly used in FTC®.
REV Robotics Mecanum Wheels (https://www.revrobotics.com/rev-45-1655/) ($127.50 with team discount): The REV Robotics Mecanum Wheels come in at only 75mm in diameter, providing a smaller mecanum wheel option to FTC teams; however, they are not the thinnest option at 40.8mm thick. They feature ball bearing supported rollers that provide exceptional strafing capabilities and traction. The hole pattern featured on these wheels is compatible with Andymark Nubs as well as the REV Robotics Universal Hex Adapter v2 (which is included); however, in order to use other shaft options the use of an adapter or physical modification will likely be required.

Open Game Manual 0

Holonomic Drivetrains > Mecanum Drive > Mecanum Wheels Miniguide > Viable

Game Manual 0·CC BY-NC 4.0

These mecanums have acceptable performance but there is very little reason to buy them at this point, as they have been superseded by better ones. If you already own them, they are a viable option, but consider looking at some of the mecanums from the Recommended section.

Nexus Ball Bearing Mecanum Wheels (https://www.robotshop.com/en/100mm-mecanum-wheel-set.html) ($134.00): These wheels are 100 mm in diameter and 59 mm wide. They strafe excellently because of the use of ball bearings. However, they are more expensive, take up more space, have less traction, and are harder to mount to than the recommended wheels.
goBILDA Mecanum Wheels v1 (https://www.gobilda.com/96mm-mecanum-wheel-set-70a-durometer-bearing-supported-rollers/) (discontinued): These wheels are very similar to the Nexus Ball Bearing Mecanum Wheels, but with a different color scheme and better mounting options.
Andymark Heavy Duty Mecanum Wheels (https://andymark.com/products/4-in-hd-mecanum-wheels) ($225): These are easily the most expensive mecanums on the list. These wheels are 4" in diameter and 1.65" wide. These are bushing based mecanums, so while they strafe decently they still perform worse than bearing based ones. They have a good amount of traction, more than the Nexus bearing or bushing wheels.
Nexus Bushing Mecanum Wheels (https://www.robotshop.com/en/100mm-aluminum-mecanum-wheel-set.html) ($80.00): This is the Nexus Ball Bearing Mecanum Wheel with bushings instead of ball bearings. There is not much to say about them except that they strafe decently but worse than the ball bearing based equivalent.

Open Game Manual 0

Holonomic Drivetrains > Mecanum Drive > Mecanum Wheels Miniguide > Not Recommended

Game Manual 0·CC BY-NC 4.0

There is almost no reason to use these mecanum wheels---they perform very poorly and are not much cheaper than those in the Recommended section. We cannot recommend the use of these on an FTC robot, if at all possible replace them with mecanums from the Recommended section.

TETRIX Mecanum Wheels (https://www.pitsco.com/TETRIX-MAX-Mecanum-Wheels) ($149.00): These mecanums are designed with a bushing based, hard plastic roller which in turn creates poor strafing performance. The integrated hub is a set screw based design with a round bore, causing unreliability as well as taking up extra space.
Andymark Standard Duty Mecanum Wheels (https://andymark.com/products/4-in-sd-mecanum-wheel) ($77.00): These wheels barely strafe and are super fragile.

(Figure: 8103 Null Robotics, Rover Ruckus, using Nexus mecanum bearing)

(Figure: 731 Wannabee Strange, Rover Ruckus, using AndyMark HD mecanum wheels)

Open Game Manual 0

Kinematics > Mecanum Drive > Inverse Kinematics

Game Manual 0·CC BY-NC 4.0

The inverse kinematics of a mecanum drive relate the desired velocity of the robot to the velocity required on the wheels. These are as follows:

vfl = vf - vs - (2r_b · ω)

vbl = vf + vs - (2r_b · ω)

vbr = vf - vs + (2r_b · ω)

vfr = vf + vs + (2r_b · ω)

Open Game Manual 0

Kinematics > Mecanum Drive > Forward Kinematics

Game Manual 0·CC BY-NC 4.0

The forward kinematics of a mecanum drive relate the velocity of the wheels to the forward, strafe, and rotational velocities of the robot, relative to itself. These are as follows:

vf = frac{v_fr + v_fl + v_br + v_bl}{4}

vs = frac{v_bl + v_fr - v_fl - v_br}{4}

ω = frac{v_br + v_fr - v_fl - v_bl}{4*2r_b}

Open Game Manual 0

Mecanum TeleOp

Game Manual 0·CC BY-NC 4.0

Note: Some common COTS mecanum drivetrain kits are the goBILDA Strafer Chassis Kit (https://www.gobilda.com/strafer-chassis-kit-v5/) and the REV Mecanum Drivetrain Kit (https://www.revrobotics.com/rev-45-2470/).

A standard mecanum drive configuration possesses 4 mecanum wheels oriented in an "X" shape. This means that the rollers are angled towards the center when looking at it from above. This configuration allows one to add up the force vectors generated by the offset rollers and derive movement in any direction. It is important to note that because of friction, perfect movement isn't possible in every direction, so a mecanum drivetrain will be able to drive slightly faster forwards/backwards than any other directions. Combining translation and rotation will also result in slower movement.

In the image above, vectors 1, 2, 3, and 4 are the force vectors created by the mecanum wheels when the chassis is instructed to drive towards the top of the image. All motors are driving forward. The blue and red lines are their X and Y components, respectively. Here are a few examples of how the wheels must be driven to achieve different movements:

Attention: It is strongly advised to not hardcode these movements in; there is a much better way described below that allows for true holonomic movement and is much more elegant.

Open Game Manual 0

WPILib Kinematics > Mecanum Drive Kinematics

This source does not publish a licence, so its text is not reproduced here. Open the source to read it.

Open FTCLib

WPILib Kinematics > Mecanum Drive Odometry

This source does not publish a licence, so its text is not reproduced here. Open the source to read it.

Open FTCLib

Dead Wheels > Drive Encoder Localization

Game Manual 0·CC BY-NC 4.0

Pros:

Cheap (the motors you're using most likely already have encoders built in)
Accessible
Very little configuration necessary
Cons:

Drive encoder localization on mecanum drive can be quite inaccurate due to lack on traction on mecanum wheels.
Will drift on high acceleration on mecanum drive. Accuracy will be good enough for basic autonomous modes if acceleration is limited

Open Game Manual 0

Programming Mecanum - Simplified | REV DUO Control System

This source does not publish a licence, so its text is not reproduced here. Open the source to read it.

Open REV Robotics Docs

Programming Mecanum - Refined | REV DUO Control System

This source does not publish a licence, so its text is not reproduced here. Open the source to read it.

Open REV Robotics Docs

Mecanum Wheel Setup and Behaviour | REV DUO Build System

This source does not publish a licence, so its text is not reproduced here. Open the source to read it.

Open REV Robotics Docs

Mecanum TeleOp > Deriving Mecanum Control Equations

Game Manual 0·CC BY-NC 4.0

Now that we have a functioning mecanum driving program, there are a few things that can be done to clean it up. The first of these would be multiplying the left X value by something to counteract imperfect strafing. Doing this will make the drive feel more accurate on non axis aligned directions, and make field centric driving more accurate. In this tutorial, we will use 1.1, but it's really up to driver preference.

Java

double y = -gamepad1.left_stick_y; // Remember, Y stick is reversed!
double x = gamepad1.left_stick_x * 1.1; // Counteract imperfect strafing
double rx = gamepad1.right_stick_x;

The other improvement we can make is scale the values into the range of -1 to 1.

Since the SDK simply clips (limits) the powers to that range, we can lose the ratio we are looking for unless we proactively put all the numbers back in that range while still maintaining our calculated ratio. For example, if we calculate values of 0.4, 0.1, 1.1, and 1.4, they will be clipped to 0.4, 0.1, 1.0, and 1.0, which is not the same ratio. Instead, we need to divide all of them by the largest power's absolute value when it exceeds 1:

Java

// Denominator is the largest motor power (absolute value) or 1
// This ensures all the powers maintain the same ratio, but only when
// at least one is out of the range [-1, 1]
double denominator = Math.max(Math.abs(y) + Math.abs(x) + Math.abs(rx), 1);
double frontLeftPower = (y + x + rx) / denominator;
double backLeftPower = (y - x + rx) / denominator;
double frontRightPower = (y - x - rx) / denominator;
double backRightPower = (y + x - rx) / denominator;

Make sure to set the powers on your motor and update this every loop in an opmode!

Open Game Manual 0