Machine Learning-based Compressive Sensing Vehicle Location Tracking ASIC Design
Tracking vehicle location can be a challenge due to many factors. First, the physical environment heavily affects the received signal, for example mobility in dense urban cities degrades the signal quality. Furthermore, location data accuracy worsens when vehicle speeds increase, making simple tracking methods unreliable.
Since human trajectories, whether vehicular or pedestrian, are complex and can be difficult to model. Classical signal processing reconstruction techniques fail to deliver satisfying results. To that end, novel algorithms based on compressive sensing combined with machine learning have been developed to accurately reconstruct trajectories with very low number of samples.
The objective is to optimize such algorithms, utilize their implicit parallel structures and implement them in hardware to speed up processing time. Such ASIC can then be integrated in the future with the already existing receivers.
- Interest in digital design
- Background in signal processing or digital communications
- VLSI II
- 20% Theory & Simulation
- 60% ASIC/FPGA Design
- 20% EDA tools