Odometry is the use of motion sensors to determine the robot’s change in position relative to some known position. The idea behind that is the incremental change in position over time.
The change in position that we called linear displacement relative to the floor, can be measured on the basis of revolutions of the wheel. This method provides good accuracy for the short-term measurement but it leads to a lot of error when done on significant displacement.
Uses of Odometry
- Can be used with a position estimator to provide better estimates.
- In some cases where no other references are available Odometry is the only navigation information available.
- Robots can have enough stability so that they are able to detect the landmarks and are used for mapping in the limited region.
Errors in Odometry:
There are two types of errors in odometry:
- Systematic Error
- Non-Systematic Error
Systematic error is caused by inherent deficiencies or inaccuracies in the system. Here, our system is a Robot on which we performing odometry. This includes:
- Inaccuracies in wheel diameter.
- Inaccuracies in wheelbase.
- Misalignment of wheels.
- Finite encoder resolution.
- Finite encoder sampling rate.
Measurement of systematic error:
Bornstein and Forg proposed a simple model for measurement of systematic error. They consider the two dominant causes of systematic error and proposed the following method:
- Unequal Wheel Diameters:
- Uncertainty about wheelbase:
Non-systematic errors are errors that are mostly caused by the environment or calculation/ estimation. Following are some causes of non-systematic error:
- Uneven floor
- Presence of unexpected objects on the floor
- Wheel Slipperage, Overaccelaration, Fast turning, etc.
Measurement & Reduction of Non-Systematic Error
Some information can be derived from the spread of return positive error. This can be thought the estimated standard deviation. However, it also depends upon the surface environment and robot. Hence, it is impossible to design a test procedure for the non-systematic error.
If the bumps are concentrated at the beginning of the first leg return position error will be small. Conversely, it will be larger for the end. Hence, return position error will not be a good measure, instead return orientation error will be a better choice.
Reduction of Non-Systematic Error
- Robots with small wheelbases are more prone to orientation error.
- The wheels used for the odometry should be knife-edge thin and non-compressible.
Auxiliary wheels and Basic Encoder Tailers
- Auxiliary wheels: The auxiliary wheels are used with the weight-bearing wheels which are made of steel specially for encoding, which makes it feasible for a differential drive, Ackerman vehicles, etc.
- Basic Encoder trails: A separate trailer is used mostly used with tracked vehicles because of the large amount of slippage during turning. It can be used when the ground has certain characteristics. While passing through obstacles or turning the trailer will be raised.
Attention reader! Don’t stop learning now. Get hold of all the important Machine Learning Concepts with the Machine Learning Foundation Course at a student-friendly price and become industry ready.