Introduction
A system for the detection and tracking
of vehicles approaching from the rear, such that an adaquate warning envelop
is created around the vehicle-to-be-warned (VTBW). That is to say,
if I choose to change lanes, such a system would warn [1] me of any hazard
or potential hazard in the intended lane.
Using an assortment of microwave transmitters,
receivers and transceivers (operating at frequencies of 10 GHz, 25 GHz,
and sweepable 10 GHz to 30 GHz) the system will basically work as follows:
There will be one emitter
located at the rear center--near the CHMSL--and
it will have a rather broad beamwidth in the horizontal
plane (covers three lanes).
There will be four receivers (having narrower
beamwidth), two mounted on each side of the vehicle--one high, one low--both
aimed straight back. (I am using only one on a side for now.)
Three modes:
1) Mode 1, Doppler:
The Doppler mode will sense the speed of
the vehicle-to-be-warned, and the speeds of the various rear-approaching
vehicles. To enable mode-1, the emitter is made to (switched) "leak" some
of its carrier RF--in equal amounts--to all receivers, thus producing Doppler
outputs (unique to each) from each receiver. It should be noted: these
receivers will be able to differentiate between approaching and receding
traffic.
2) Mode 2, Differential:
The relative reflected intensities detected
by the four receivers should help in the discrimination of the various
vehicles' locations. This second mode, will not have the leaky carrier,
but will detect relative CW signal strengths from each receiver. By analyzing
the relative differences between what each antenna "sees," there should
be an enhancement of the positional information. This coupled with range
and velocity data, should be synergistic.
3) Mode 3, Range:
The Range mode is used to derive distance
information of the target vehicles. This mode will also not use leaky carrier.
This method of range or distance measurement uses a relatively low bandwith
phase nulling approach, not fast, wide bandwith pulse techniques. It will
use swept AM modulation, i.e., the emitter (Gunn oscillator & antenna)
will be AM modulated with a linearly swept sinusoid, having a frequency
ranging from ~ 1 MHz to ~ 10 MHz.
[TOP]
Principle:
At a distance = 1/4 wavelength of the modulating
sinusoid, the reflected carrier at the receiver--after demodulation--will
be 180 degrees out of phase (null) from the originating modulating wave.
Knowing the exact frequency of the modulating wave, the approximate distance
to the target can be derived. (remember, 1/4 wavelength is the round-trip
distance for 180 degrees or1/2 wavelength) Note, there will be some error
in this approach if the rate of sweep--of the modulating wave, verses the
range isn't accounted for.
As the modulating frequency continues to
increase, the 180 degree "null" will again occur when the modulating frequency
is 1.5 times the first null frequency (fm); this will continue
to occur every additional half wavelength, i.e., 1.5 fm, 3.5
fm, 5.5 fm, 7.5 fm, etc.--each time resolving
the distance ever finer. That is: 1/2 WL at 10 MHz is a finer unit of measurement
than 1/2 WL at 1 MHz.
All three modes will be used in combination
in a very rapid fashion. (limited by PRF, Doppler frequencies at n x MPH
@ m x GHz, as well as, the sweep rate of the modulating frequencies, fm)
Summary:
If this approach proves to have merit, then
I propose that the next version use some sort of pulsed, coded spread
spectrum, i.e., TDMA, CDMA, FDMA, etc. Each has different advantages
and disadvantages,i.e., TDMA (Time Division Multiple Access) might be safer
(from a health standpoint) because it emits RF energy less than full time;
CDMA offers better security/reliability; FDMA is better surviving local
interference, etc.
This is a daunting task: because of the high
reliability needed, the operating environment, the kinds of information
needed, the finite time allotted for its detection and interpretation,
and the cost target as an automotive OEM device
It is a given, that the success of this approach
will depend, in great measure, on good signal processing (DSP) techniques.
Learning Curve
Playing with a Doppler RADAR device made
from a modified "fuzz-Buster" with the intentions
of better understanding the characteristics of ~ 25 GHz Doppler.
With an assortment of microwave transmitters,
receivers and transceivers--10 GHz, 25 GHz, and a sweepable 10 GHz to 30
GHz--to test an idea I have. |
