This article covers implementation and testing of sensor data fusion for robot localization. System is capable of finding robot in known environment using LIDAR, odometry and RFID tags. The localization algorithm is based on the particle filter implemented in NVIDIA CUDA. It fuses 3D LIDAR, odometry and RFID data. The experimental setup was created to verify potential use in the indoor environments. Quantitative benchmark was performed using popular robotics simulation framework. The benchmark on low-energy CUDA enabled device is performed. Current study shows the advantage of data fusion for robot localization and demonstrates that current approach can efficiently solve the problem of mobile robot deployment in known environments.