Accelerating Graph-Based SLAM through Data Parallelism and Mixed Precision on FPGAs | IEEE Conference Publication | IEEE Xplore