Geometric Tracking of Vehicular mmWave Channels to Enable Machine Learning of Onboard Sensors

conference paper

English

Estimating time-selective millimeter wave wireless channels with directive antennas poses a challenging task. A feasible way of relaxing this channel estimation problem is to focus on the tracking of a few multipath components (MPCs). Aligning antenna beams to the tracked MPCs increases the channel coherence time by several orders of magnitude. We propose to track the MPCs geometrically. Our geometric trackers are based on algorithms known as Doppler-bearing tracking. We reformulate recent work on geometric-polar tracking into an efficient recursive version. If the relative position of the MPCs are known, other sensors on board of a vehicle, for example, lidar, radar, camera, will be capable of performing supervised learning based on their own observed data. Learning the relationship between sensor data and MPCs allows onboard sensors to participate in the channel tracking. Joint tracking from many onboard sensors possibly increases the reliability of the MPC tracking.

MILLIMETER; COMMUNICATION; PROPAGATION; ESTIMATOR

ZÖCHMANN, E.; VUTHA, V.; RUPP, M.; HEATH, R.

9. 12. 2018

Institute of Electrical and Electronics Engineers Inc.

978-1-5386-4920-6

Proceedings of IEEE Global Telecommunications Conference (GC Wkshps)

1

6

6

https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8644440