Spectrum Sensing for Radar Communications with
Unknown Noise Variance and Time-variant Channel
Mengwei Sun
Beijing University of Posts and Telecommunications
Beijing, China
Email: mwsun@bupt.edu.cn
Mingjun Shi, Chenglin Zhao, Bin Li
Beijing University of Posts and Telecommunications
Beijing, China
Abstract—In this paper, a new spectrum sensing framework is
proposed for radar communications, which recovers other
informative states associated with realistic radar environments,
e.g. fading channel gains and unknown noise variance, when
detecting the occupancy of primary-band. We firstly formulate a
dynamic state-space model by full considering the unknown noise
variance and time-variant flat fading channel. On this basis, a
novel spectrum sensing scheme, relying on maximum a posteriori
probability criterion and marginal particle filtering technology, is
designed to estimate the state of primary user, time-variant
fading channel gain and noise variance jointly. Experimental
simulations show that the proposed method improves the sensing
performance significantly and estimate the fading channel gain
and noise variance accurately.
Keywords—spectrum sensing; radar communications; time-
variant flat fading channel; unknown noise variance; joint
estimation
I.
I
NTRODUCTION
The problem of scarce spectrum resource arises because the
conventional static spectrum management can not
accommodate the development of wireless technologies [1].
This problem also occurs in the field of radar communications
[2]. Future radar systems should be aware of their operational
electromagnetic environment and dynamically adjust its radio
operating parameters. Spectrum sensing is an important
technology to detect surrounding electromagnetic environment
and various sensing methods have been proposed such as
energy detection (ED) [3, 4], matched filtering (MFD) [5, 6],
etc. However, most existing sensing methods seem to become
ineffective when applied to the radar communication system in
which noise uncertainty and time-variant fading channel are
common.
There have been several kinds of approaches proposed to
reduce the detrimental effects caused by noise uncertainty,
including multi-antenna based spectrum sensing scheme [7],
cooperative spectrum sensing method [8] and the spectrum-
sensing algorithm based on sample covariance matrix of
received signal [9]. These methods set forth stricter
requirements for receiving equipment or correlation of received
signal. Thus, the practical application of these methods is
limited. On the other hand, there are two common methods to
address the spectrum sensing issue caused by time-variant
fading channel. One is to utilize cooperative technique as well
[10]. Another method is to consider the statistical properties of
time-variant fading channel but fails to model the evolution of
time-variant channel gain. Furthermore, there is no spectrum
sensing method has been designed to overcome the challenge
induced by unknown noise variance and time-variant fading
channel simultaneously.
In this investigation, based on maximum a posteriori
probability (MAP) criterion and marginalized particle filtering
(MPF) technology [11], a new spectrum sensing method is
proposed for single antenna system under the circumstance that
the noise variance is unknown and channel gain is dynamic. In
sharp contrast to existing schemes, the noise variance and time-
variant channel gains will be estimated blindly based on
observed signals, and at the same time, the primary user (PU)
states will be detected. To sum up, the main contributions of
this paper are two-folds.
Firstly, a novel dynamic state-space model (DSM) of
spectrum sensing is proposed. A major innovation of this
framework is that the time-variant flat-fading (TVFF) channel
gains, unknown noise variance and PU states are treated as
hidden states to be estimated.
Secondly, a sequential estimation scheme is proposed to
estimate the PU state, time-variant fading channel gain and the
unknown noise variance jointly. Based on Bayesian framework,
the joint probability distribution of the unknown parameters
and PU state is approximated by utilizing particle filtering (PF)
and marginalization concept. Simulations show that compared
with traditional methods, the sensing performance may be
improved significantly in realistic radar communication system
with TVFF channels and unknown noise variance.
The rest of this paper is organized as follows. In Section II,
we provide the dynamic state-space model of spectrum sensing.
On this foundation, the joint sensing algorithm is introduced in
detail in Section III. Numerical simulations and performance
analysis are provided in Section IV. Finally, conclusion is
generalized in Section V.
This work was supported by the National Natural Science Foundation of China (61379016, 61471061)
978-1-4673-6305-1/15/$31.00 ©2015 IEEE
IEEE ICC 2015 - Workshop on Radar and Sonar Networks
2513