In many technologies today, sensors are a significant part of the set of components. Our project deals with distance or location sensors. As with all other sensors, there is also an error here in them. This error – can be split into two – a fixed offset error – which can be found by a set of measurements and then calibrated by rebooting with the offset error balance. And another error due to variable noise. Noise may be due to many factors, including sensor ambient temperature, distance measured, component thermal noise, quantization, etc. The nature of the noise may have different behaviors such as salt and pepper and Poisson noise. In this project it is been assumed to be a Gaussian noise (but under small changes the project can also be adapted to other types of noise).
The purpose of the project is to perform revaluation and forecasting for purposes that change the nature of traffic in the presence of noise, thereby reducing the revaluation error during the change and reducing the prediction error. The possibility of noise
filtering from a set of measurements obtained from the sensor was tested, as well as the prediction of a few steps ahead of the motion, with the possible motions being wired, circular and combined motion.
To do this, we will consider using two tools, the Kalman filter and the RNN network, which will be detailed in the above report.