COL 10(5), 052801(2012) CHINESE OPTICS LETTERS May 10, 2012
Improvement of signal-to-noise ratio in chaotic laser radar
based on algorithm implementation
Bingjie Wang (XXX''')
1,2∗
, Tong Zhao (ëëë ÕÕÕ)
1
, and Huakui Wang (uuu¿¿¿)
2
1
College of Physics and Optoelectronic Engineering, Taiyuan University of Technology, Taiyuan 030024, China
2
College of Information Engineering, Taiyuan University of Technology, Taiyuan 030024, China
∗
Corresp onding author: wangbingjie@tyut.edu.cn
Received October 18, 2011; accepted December 28, 2011; posted online February 20, 2012
Chaotic laser radar based on correlation detection is a high-resolution measurement tool for remotely mon-
itoring targets or objects. However, its effective range is often limited by the side-lobe noise of correlation
trace, which is always increased by the randomness of the chaotic signal itself and other transmission
channel noises or interferences. The experimental result indicates that the wavelet denoising method can
recover the real chaotic lidar signal in strong period noise disturbance, and a signal-to-noise ratio of about
8 dB is increased. Moreover, the correlation average discrete-component elimination algorithm significantly
suppresses the side-lobe noise of the correlation trace when 20 dB of chaotic noise is embedded into the
chaotic probe signal. Both methods have advantages and disadvantages.
OCIS codes: 280.3640, 140.1540.
doi: 10.3788/COL201210.052801.
Optical chaotic signals generated from lasers have at-
tracted extensive attention in recent years and have a
wide range of applications
[1−5]
. The chaotic laser radar
(chaotic lidar) was proposed
[6,7]
based on the wideband
and δ-function-like correlation properties of a chaotic
laser. In a chaotic lidar, the target is identified by cor-
relating the reference chaotic signal with the delayed
probe signal reflected or backscattered from the tar-
get. Although the correlation technique has been used
in the random-modulation continuous wave (RM-CW)
lidar
[8,9]
, the range resolution is still limited on the or-
der of meters by the modulation speed and pseudoran-
dom code rate. Compared with pseudorandom-code-
modulated CWs, noise-like optical chaotic waveforms
generated by nonlinear dynamical semiconductor lasers
have bandwidths larger than 10 GHz
[10]
, which ensure
high-range resolution and unambiguity. Furthermore,
the chaotic laser with multi-gigahertz bandwidth can be
easily obtained without an electro-optical modulator.
However, unlike the RM-CW lidar, in which an ideal
pseudorandom code with perfect autocorrelation charac-
teristics is used
[9]
, the side-lobe noise of the correlation
trace induced by the randomness of the chaotic signal
is unavoidable. The high side-lobe noise of the corre-
lation trace may bury the weak object correlation peak
and degrade the signal-to-noise ratio (SNR). In real ap-
plications, the chaotic probe signal in the transmission
channel may be disturbed by other signals and various
noises, such as the random noise produced by the lidar
system itself, the interferences or noises from the sun, the
active jamming from other pulsed or CW lidars, etc. The
noisy chaotic signal has more waveform distortions than
a clean chaotic signal. These distortions further increase
the side-lobe noise of the correlation trace and degrade
the SNR. Thus, the suppression of the side-lobe noise
of the correlation trace is a fundamental and important
problem in chaotic lidar applications.
The side-lobe noise of the correlation trace is effectively
suppressed by recovering the chaotic probe signal pol-
luted by channel noises. The chaotic laser signal is a
wideband and nonstationary signal. Conventional noise
reduction methods such as linear low-pass filtering do
not work well since the signal and noise often have over-
lapping bandwidths. Moreover, many noise reduction
methods are proposed for nonlinear time series
[11,12]
,
which are complex and impractical in specific situations.
The wavelet theory and its method provide a new tool
that reduces the noise from a chaotic sequence
[13]
. This
method is considered nonparametric and is applicable to
nonlinear noisy data even without prior information of
their dynamics.
In this letter, a chaotic ranging system is established
and the effects of different types of channel noises on the
SNRs are investigated. Wavelet denoising is employed
to detect the chaotic signal in noisy environments. After
the effect of noise reduction is limited using the wavelet
method, a correlation average discrete-component elim-
ination algorithm is developed to suppress the side-lobe
noise of the correlation trace. The experimental results
of the two algorithms for SNR improvement are also pre-
sented.
The diagram of the laboratory experimental arrange-
ment of the chaotic ranging lidar system is shown in
Fig. 1. The chaotic source is an 808-nm, 500-mW, single-
mode laser diode with optical feedback from a mirror.
The laser diode is biased at 200 mA (1.4I
th
) with 70-mW
output power, and a chaotic state with 10.24% optical
feedback is obtained. The chaotic laser generated by the
chaotic source is split into the reference and the probe
light using a beam splitter. The reference light is focused
using a converging lens and detected by a photodetec-
tor (PD). Meanwhile, the probe light is transmitted to
the target through the transmitter, and the reflected or
backscattered light is collected by the receiver and de-
tected by another identical PD. The transmitter consists
of a collimating lens and a beam expander, and the re-
ceiver is a Maksutov-Cassegrain telescope with a 9-cm
receive aperture and 1.2-m focal length. The waveforms
of the reference signal and the echo probe signal detected
by the two PDs are displayed and recorded on a digital
1671-7694/2012/052801(4) 052801-1
c
° 2012 Chinese Optics Letters