What is Odometry?

In a search and retrieval challenge like the Robo-Rat competition, it is critical for a robot to be aware of its location. It might not seem like a difficult effort to know where you are, but as you shall see, it can be really difficult. People can locate themselves using their sharp vision, long-term memories, and awareness of their surroundings. However, robots do not naturally possess these sophisticated abilities, especially the ones you will be developing for this class. Odometry is one technique that robots use to locate themselves in their surroundings.

What is Odometry?

The measurement of a robot's position change about a given location using motion sensors is known as odometry. For instance, if a robot is moving straight forward and is aware of the diameter of its wheels, it can determine how far it has come by counting the number of wheel rotations. Robot driving wheels are typically equipped with shaft encoders, which produce a predetermined number of pulses each revolution. These pulses are counted by the CPU, which can calculate the distance traveled.

Examples

• Imagine a robot with rotary encoders on its wheels or the joints of its legs. After moving forward for a while, it stops to ask how far it has come. It can calculate the distance if it knows the diameter of its wheels and can measure how far the wheels have turned.

• Odometrics is commonly used in train operations as well. Odometry is often used to determine a train's relative location when it is between two stationary sensors, whereas crossing over stationary sensors in the tracks typically gives a train its absolute position.

Uses

• To provide better estimations, odometry can be used with a position estimator.

• Robots are employed for mapping in constrained areas because they are sufficiently stable to be able to recognize landmarks.

• When no other sources are accessible, there are several situations There is no other navigational data available than odometry.

Limitations

Odometry has certain drawbacks despite being a position sensor used by mobile robots often. Any sensing inaccuracy will get worse with time because it's a cumulative measurement. To avoid a significant buildup of errors, robots may occasionally need to employ additional sensors to identify their precise location.

One of the key problems in mobile robot applications is the precise localization of a vehicle. For autonomous navigation to be possible, a robot must continuously be aware of its location. As a result, researchers and engineers have created a number several methods, and systems for mobile robot location, including wheel odometry, laser/ultrasonic odometry, the GPS, GNSS, inertial navigation system (INS), and visual odometry (VO).

Odometry inaccuracy can be caused by ?

• Faulty measuring of wheel diameter
• Multiple-wheel drive systems with various wheel diameters
• Problems in pulse counting while using driving shaft encoders
• Odometry processing is slow (considering only cumulative counts, not the dynamic count differences).

Working of Odometry

A differential drive robot's two possible motion routes are depicted in the diagram above (each wheel is actuated by a separate motor). The encoder values are located next to the wheels. In the left diagram, the robot moves in a straight path, ensuring that the encoder values are always identical. The robot on the right, however, follows a serpentine course. The robot was making a right turn when the green wheel encoder read 80 and the red wheel encoder read 40 at the halfway point.

The encoders at the robot's final position, however, read the same value even though they are bigger than the readings on the left since the robot had gone a greater distance by following an S-shaped curve. This is because the robot subsequently turned back to the left. Your control program would think that the robot moved in a straight path if it just considered the encoders at the finish position. This is why mistakes can result from sluggish encoder processing. You should process the odometry information as quickly as you can unless your robot is mechanically guaranteed to move in a straight path.

How Odometry errors are reduced

• Orientation mistakes are more likely to occur in narrow-wheelbase vehicles.
• More accurate odometry is provided by the synchro-drive architecture.
• The wheels that are utilized for odometry should be razor-thin and incompressible.
• Castor wheels that bear a sizable percentage of the weight are liable to slide.

Conclusion

To summarize, odometry is the process of tracking a robot's absolute location. Motion algorithms can utilize this information to drive to locations or turn to absolute angles.