Skip to content

Latest commit

 

History

History
20 lines (15 loc) · 803 Bytes

README.md

File metadata and controls

20 lines (15 loc) · 803 Bytes
Raw

Approximate Computing for an Extended Kalman Filtering Tracker

An implementation of online approximator for an extended Kalman filter tracker implemented over FPGA is presented.

Getting started

Please run Demo_detailed_visualisation.m and Demo.m to simulate Figure 4 of the paper, which contains dynamic approximation via four different Kullback–Leibler (KL) thresholds.

Also, please run Demo_Exact_Computation.m for exact computation, i.e target tracking with an extended Kalman filter without any approximation.

Paper

@inproceedings{Emambakhsh:2017,
  title={Learning to approximate computing at run-time},
  author={P. Garcia, M. Emambakhsh and A. Wallace},
  booktitle={IET 3rd International Conference on Intelligent Signal Processing (ISP)},
  pages={1--8},
  year={2017}
}