Radar sensors detect, locate, track, and identify objects using radio waves. They provide several

abilities to detect objects in various environmental conditions and at different ranges, and enabling

advanced driver assistance systems (ADAS) in vehicle integrated sensor networks. In the ongoing

race to aim at higher leves of automation in automotive transportation it is required that an comple-

mentary network of sensors provide a comprehensive environmental perception and are interoperable

in terms of mutual interference avoidance and data processing.

Especially Frequency Modulated Contentious Wave (FMCW) radars operating in the 76 to 81GHz

range are sensitive to mutual electromagnetic interference. The rising number of radars on public

road makes receiving each others transmit signal more likely. Even radar sensors operating in local

sensor networks are at risk to interfere with each other, even with no line of sight due to multi-

path reflections. Both cases can lead to false detections of non excising objects having catastrophic

consequences on decision making for a fast moving autonomous vehicle. An effective interference

avoidance strategy in the field of automotive radar is needed.

In this work, a simple but effective methodology is introduced and investigated for interference

avoidance in FMCW radar sensor networks with real world hardware and measurements. The strat-

egy entails interleaving of chirp sequences in time or frequency domain and keeping the increase of

overall frame time to a minimum to ensure that synchronized sensor fusion with other sensor types

at high frame rates up to 40 fps remains possible. System time of two FMCW radar modules is

synchronized by implementing the Precision Time Protocol (PTP) in an local area sensor network al-

lowing for precise synchronous sequence triggering at the scale of nano seconds. Metric are defined

to systematically evaluate the performance of the proposed interference avoidance strategy.

The second key point of this work is an analytical investigation on bistatic FMCW signals. Based

on parametric descriptions the influence of reference oscillator and phased locked loop on radar data

is derived. The model is used in the context of Range-Doppler processing for qualitative comparison

with measured radar data. In summary, a comprehensive investigation of interference is performed

to facilitate an empirical approach.