LS telcom AG · XETR:LSX
LS telcom — LS Summit 2014
24 slides from LS telcom's conference presentation, Jul 2, 2014
Slide 1: Employing the very same electromagnetic wave simultaneously for communications and radar: It works!
Employing the very same electromagnetic wave simultaneously for communications and radar: It works! Martin Braun, Friedrich Jondral LS Summit, Lichtenau, July 2, 2014 Communications Engineering Lab Univ.-Prof. Dr.rer.nat. Friedrich K. Jondral KIT – University of the State of Baden-Wuerttemberg and N
Slide 2: Motivation
Motivation Frequency as a resource is scarce an should be used economically Similar techniques are employed by radio communications and radar Communalities in hardware and software radar & 11001011 . . . radar + communications communications communications communications radar radar 2 03.06.2014 Com
Slide 3: Radar Networks
Radar Networks Benefits: Frequencies are used economically. A combination of radar and radio communications may lead to enhanced road safety. Challenge: Networked radar and radio communi- cations systems are to be coordinated (differing requirements w.r.t. medium access). 3 03.06.2014 Communications
Slide 4: Topics
Topics OFDM Radar Estimation Theory OFDM Radar Networks Further Investigations Summary 4 03.06.2014 Communications Engineering Lab Univ.-Prof. Dr.rer.nat. Friedrich K. Jondral
Slide 5: OFDM Radar: Review
OFDM Radar: Review OFDM: Well established technology (LTE, WLAN, DVB, …), convenient for wireless communications 2000: First application of OFDM for radar 2008 - 2010: Research Project RadCom (JO 258/15-1, ZW 1801/1-1), simultaneous usage of an OFDM signal in radar and radio communications, partiall
Slide 6: OFDM Basics
OFDM Basics OFDM: Orthogonal Frequency Division Multiplex cyclic prefix OFDM signals are matrices carrying information: sub carriers, OFDM symbols, modulation alphabet ∩ C 6 03.06.2014 Communications Engineering Lab Univ.-Prof. Dr.rer.nat. Friedrich K. Jondral
Slide 7: Signal Processing by the Radar Receiver
Signal Processing by the Radar Receiver (Initially Single Target) Relative velocity (Doppler) Distance (delay) Complex attenuation Noise (AWGN assumption) The elements of the received matrix are divided by the elements of the transmission matrix: 7 03.06.2014 Communications Engineering Lab Univ.-Pro
Slide 8: Maximum Likelihood Estimator
Maximum Likelihood Estimator Derivation of the ML estimator for ODFM radar: Parameter vector: Estimation: Function to be maximized (noise is assumed to be AWGN): 8 03.06.2014 Communications Engineering Lab Univ.-Prof. Dr.rer.nat. Friedrich K. Jondral
Slide 9: ML Estimation at Discrete Arguments
ML Estimation at Discrete Arguments The log likelihood function is maximized by finding the maximum of a two-dimensional peridogram. Discretization narrows down the search domain. Einfache The indicesUmrechnung von conver- m and n are easily ted into distance Indizes andund in Distanz relative veloc
Slide 10: Outage Probability of a Radar Network
Outage Probability of a Radar Network Definition: The outage probability Pout of a radar network is given by the probability by which a reference object cannot be detected caused to network internal interferences. G(Φ) Φ dref The outage probability is complementary to the reference object‘s detectio
Slide 11: Quantization Error
Quantization Error Discretization allows for a numerical search for a ML estimator but causes errors (quantization loss, quantization noise). frequency index The maximum of the discretized peridogram is always found within half a quantizing step around the true maximum. 11 03.06.2014 Communications
Slide 12: Estimator Performance
Estimator Performance Numerical evaluation (by simulation) Signal parameters: center frequency 24 GHz bandwidth 93,1 MHz radar cross section 10 m² noise figure 10 dB 125 m temperature 290 K distance / m An object with fixed radar cross section is moved away from the radar. If the distance from the r
Slide 13: Simulation: Mean Estimation Error
Simulation: Mean Estimation Error quantized quantized mean estimation error / m quadratic Interpolation quadratic Interpolation optimization optimization half quantization step half quantization step distance / m The quantization error is dominant below the threshold There is no measurable differenc
Slide 14: Simulation: Variance
Simulation: Variance quantized quadratic interpolation optimization averaged CRB variance / m2s-2 quantization noise distance / m Velocity and distance are uncorrelated. Accurate predictions are based on quadratic interpolation or on the averaged Cramér-Rao Bound (CRB). 14 03.06.2014 Communications
Slide 15: Multi Target Detection
Multi Target Detection Multiple targets superimpose linearly (in ): Discrimination between targets and noise employs a threshold −20 −24 −25 80 −32 depends of the false alarm rate −30 −40 power / dBm −35 distance / m 60 −48 −40 should be chosen such that 40 −50 −45 −56 −50 the estimator works in a d
Slide 16: OFDM-Radar-Networks
OFDM-Radar-Networks Radar systems are transmitting periodically How to deal with collisions? Solution: subcarrier index Clocked access - different sub- carrier sets For high subscriber densities collisions (interferences within the network) cannot be avoided! Wanted: Theoretical background for evalu
Slide 17: Detection in a Networked Radar System
Detection in a Networked Radar System Medium access is globally clocked access probability Power received from the th interferer: Interferers are distributed over the two-dimensional plane according to a Poisson point process At the receiver the sum of all interferences form an AWGN (Central Limit T
Slide 18: A Lower Bound for the Outage Probability
A Lower Bound for the Outage Probability By which probability is the presence of one interferer sufficient in order to provoke radar outage? Dominant Interferer Phenomenon By which probability is no interferer too close to the reference system? φ The bound depends on the choice of the radar system p
Slide 19: Empirical Verification
Empirical Verification Directional radiation pattern: Cone Approximation of the outage probability: The approximation works well upper bound for small subscriber densities. simulation lower bound Simulations are costly and time-consuming. The approximation facilitates the verification of parameters.
Slide 20: Further Investigations
Further Investigations Signal Parameters Comprehensive discussion of all signal parameters with respect to the communication as well as to the radar subsystem. Investigation of WLAN as a radar waveform. х target position Alternative Estimators + estimation distance / m distance / m Multi Target Dete
Slide 21: Demonstrator
Demonstrator Signal parameters: Accordng to IEEE 802.11p Frame repetition rate: 5 Hz right lane left lane distance / m relative velocity / ms-1 21 03.06.2014 Communications Engineering Lab Univ.-Prof. Dr.rer.nat. Friedrich K. Jondral
Slide 22: Summary
Summary OFDM is suitable for communications and radar systems using simultaneously the very same electromagnetic wave Periodogram based method provides statistically optimal (maximum likelihood) estimators Networking Aspects Definition of outage probability in radar networks Analytic approximation o
Slide 23: Detailed Discussion
Detailed Discussion 23 03.06.2014 Communications Engineering Lab Univ.-Prof. Dr.rer.nat. Friedrich K. Jondral
Slide 24: Thanks for your attention!
Thanks for your attention! 24 03.06.2014 Communications Engineering Lab Univ.-Prof. Dr.rer.nat. Friedrich K. Jondral