Moving Object Tracking with a Kalman Filter as a Virtual Sensor in Robotino

This paper addresses the implementation of Kalman Filters (KFs) for estimating the velocity, acceleration, and jerk of object movements. The KF is a crucial algorithm in the field of optimal estimation and is employed as a virtual sensor to minimize the number of sensors required. This topic is particularly relevant in the field of mobile robotics, where advanced control requires sensors for position, velocity, acceleration, and jerk. Reducing the number of sensors is often suitable for minimizing faults and costs. KFs provide optimal estimation in the presence of noise and uncertainties. The results are validated using movements of the Robotino. Various tests confirm the performance of the KF and its robustness with respect to parameter variations and noise in the measurements. The proposed approach is quite general and can be applied to any system where object tracking for navigation is required, offering a significant reduction in the number of sensors needed. This not only makes the system more cost-effective but also simplifies the overall design and improves its scalability, making it suitable for a wide range of applications.