Wheel Odometry

Counts how much the wheels have turned to estimate how far a robot has traveled and how much it has turned.

Wheel odometry is the most basic method of relative localization: an encoder on each drive wheel reads how far it has rotated, which multiplied by the wheel radius gives distance traveled; for a differential-drive base, dividing the difference between the left and right wheels’ distances by the wheelbase gives the body’s turning angle, and accumulating this frame by frame gives the robot’s pose relative to its starting point. It’s cheap, runs at high frequency, is unaffected by lighting, and comes built into almost every wheeled base. The drawback is that error only grows: wheel slip, uneven ground, and inaccurate wheel-radius calibration all push the estimate further off over time — drift — so wheel odometry is usually fused with an IMU, lidar, or visual odometry, for example with an extended Kalman filter, and then corrected further by a SLAM or localization algorithm. ROS 2’s diff_drive_controller computes and publishes the odom topic from exactly this kind of left/right wheel feedback.

ExampleA differential-drive robot with a 0.4-meter wheelbase has its left wheel travel 1.0 meter and its right wheel travel 1.2 meters over some interval: the body center moves forward about 1.1 meters while turning left by (1.2 − 1.0)/0.4 = 0.5 radians, about 29°.

Also called

Wheel Encoder Odometry

Related

Rotary Encoder · Differential Drive Kinematics · Visual Odometry · Leg Odometry · Extended Kalman Filter · Slip

Sources

Wikipedia: Odometry
ros2_controllers: diff_drive_controller user documentation