Sensor fusion offers significant advantages over analyzing individual data sources separately, leveraging the complementary strengths of multiple sensors in various applications. The use of multiple sensors enhances the accuracy and reliability of tracking systems. This paper proposes a comparative study of multi-sensor fusion algorithms, focusing on the integration of radar and lidar data for tracking dynamic objects in the environment. The sensor data used by the fusion algorithms mainly consists of dynamic states (position, velocity, and additional parameters) of surrounding objects. This work specifically evaluates the accuracy and consistency of the Extended Kalman Filter (EKF) and the Unscented Kalman Filter (UKF). The study uses different models, incorporating angular velocity data from radar. Implemented in C++ and visualized using Python, the results demonstrate that the UKF outperforms the EKF across various state variables. The UKF demonstrates lower Root Mean Square Error (RMSE) values due to its superior handling of non-linearity. The consistency of both filters is confirmed using the Normalized Innovation Squared (NIS) statistic, which also addresses how to deal with faulty data. These results highlight the potential of the UKF to enhance the precision of object tracking in autonomous driving applications.