Error modeling of various sensors for robotics application using Allan Variance technique

Present day mobile robots are meant for very precise applications. For very precise applications of mobile robots, accurate estimation of inertial parameters depends upon the accuracy of mathematical model & as well as accuracy (error characteristics) of the individual sensor measurements. Multiple sensors add redundancies to the system as well as it helps to estimate the system states accurately through judicious fusion. As sensor measurements itself are prone to various types of noises, the detail error modeling is very essential for estimation of appropriate signals from the noisy sensor data. The essence of the error modeling is to understand & characterize the different types of noises present in the measured. This paper illustrates the characterization and identification of noises present in the Encoder (Model: Maxon HEDL-5540) & Inertial navigation system (Model: Crossbow NAV440) measurements using Allan Variance technique.