A SIMPLE EXPLANATION OF THE THEORY
Radar (radio detection and ranging) functions by exploiting a very simple concept: An electromagnetic wave traversing a medium will be partially or completely reflected when it encounters another medium of vastly different dielectric constant. What does this mean in everyday english? Think of a typical flashlight and its use in a dark, empty room. If you are inside a room with clear air, you do not see the beam of light until it intersects one of the walls. Why do you suddenly see the light when it hits the wall but not as it crosses the air? Because visible light lies in a particular frequency band to which air is relatively transparent but the wall is not (purists may object by stating the beam cannot be seen because clear air has few scatters in that frequency band but this is not the only reason). If this seems strange to you, then consider the fact that X-rays are used to take pictures of your bones because skin, organs and fluids are transparent to X-rays but not bone which can scatter or reflect the X-rays.
In a similar fashion, radiation of the frequencies typically used by radar systems will traverse air with little or no reflection but will be reflected strongly by metals, the ground and, albeit weakly, even water droplets present in the atmosphere. These reflections propagate back to the radar system which then interprets them to form the images we are typically familiar with. Many radar systems are in fact very similar to a flashlight that operates at a different frequency. A relatively narrow beam of light (i.e. typically near the microwave band) is emitted from the radar antenna and reflections are used to illuminate the target just as your eye can see the wall because of visible reflection from a normal flashlight. The typical wavelenth of visible light (i.e. what you see with your own eyes) is approximately 500 nanometers while the frequencies typically used in radar systems are in the 10 centimeter range. This means a radar image will never be as sharp as a convention image (e.g. a photograph) because it cannot resolve two points closer than 10 centimeters. In the absolute grossest sense, this is the theory behind radar systems.
INTRODUCTION TO A GENERIC PULSED RADAR
In this first article, we will discuss the one of the simplest and earliest radar systems: a generic 'pulsed' radar. A radar system is typically composed of a few major components. The first is a device designed to convert electrical energy into electromagnetic radiation of the desired frequency. The earliest radars used a magnetron to convert DC current into high energy microwave radiation. This is the same component found in microwave ovens. The high energy radiation is then guided to the radiating element of the system (e.g. the antenna) by a waveguide that confines the radiation to a set path through the system. Once the radiation arrives at the antenna, it is emitted by a central feed onto a parabolic dish that forms a tight radar (i.e. radiation) beam. It is not required that a radar system use the same antenna both for transmission and reception but this is the typical situation and what I will assume in most my articles (when this is not true I will explicitly say so).
Since the same element is typically utilized both for transmission (TX) and reception (RX) so that care must be taken to prevent energy from leaking into the RX components during transmission as this would damage the delicate RX components. This can be accomplished by alternating between transmission and reception whereby the high energy radiation is blocked from entering the RX components during TX but the returning reflected radiation is allowed to enter the RX components (and obviously blocked from leaking into the TX system). The radar, therefore, has its own operating frequency separate from the transmission frequency (i.e. the TX/RX switching rate). The rate of TX/RX switching is called the pulse repetition frequency (PRF) of the radar system. When in TX mode, the radar system emits a set of pulses of a given frequency whose echoes are subsequently received when in RX mode. The echoes are then processed in order to form an image.
In the simplest sense, a radar image is formed by calculating the range (i.e. the distance) the received energy came from. Since the speed of light 'c' in any given medium is a constant, this can easily be done by multiplying the speed of light by half the round-trip time of the pulses. This is called 'pulse delay ranging.' The azimuth of the target (i.e. the angle the target lies at relative to the radar system) is determined by the angle the antenna is facing. A typical radar beam is approximately 3-4 degrees wide and this give you an angular range where the target may lie (e.g. a target that appears to be at 21 degrees actually lies somewhere between 20 and 23 degrees). At large ranges, even a small angle error of 3 degrees can translate to hundreds or more meters of cross-range error (i.e. cross-range is defined perpendicular to the range). This can be corrected by a number of techniques that will be discussed in a future article.
A PREVIEW OF DIFFICULTIES TO DISCUSS
A number of difficulties have been glossed over in my very brief introduction to radar. One important difficulty arises from the competing demands of range resolution and sensitivity. For a radar system to increase how far it can see (i.e. its range) and how small an object it can see at a given range, it must increase its average emitted power. For a given frequency, the average emitted power can be increased by transmitting for a longer period (i.e. increasing the length of the TX pulse) but this will make it impossible to resolve closely separated targets because their echoes will overlap. Therefore, you are generally limited to pulse lengths of no more than 1 microseconds for adequate resolution. A number of techniques are used to solve this dilemma.
Typically, the received pulses are compressed to increase the received energy. One method used to compress the echoes is to encode the pulses by modulating the phase of the TX pulses. An example of this would be to flip every third pulse "upside-down" and observing the phase of the received echoes. Another method of compression called 'chirp' linearly increases the frequency of transmission as a function of time. Since the energy radiated is proportional to the frequency of emssion, this increases the total power emitted for any given length of time. The echoes are then passed though a filter that introduces a delay which decreases as a function of frequency (i.e. the lowest frequencies are delayed the most and the highest frequencies the least). This compresses the echoes into a narrow pulse and improves the resolution of the radar system. For radar systems that operate at too high a PRF, pulse delay ranging will not work well. Such CW (continuous wave) radars must rely on alternate techniques. One typical method is frequency modulation. The emitted energy has its frequency modulated and range is determined by timing the delay in the frequency modulation of the receieved echoes.
More advanced radars can determine the relative velocity of a target by observing the tiny shifts in frequency of the echo due to doppler shifting. These radars are typically known as 'pulsed-doppler' radars and they are not simple to build as they impose a number of additional constraints on the radar system. In the next article, I will introduce a generic pulsed-doppler radar and illustrate the main differences between it and its simpler cousin. Future articles will return to discuss how angular resolution is improved (i.e. getting better than 3-4 degrees), automatic radar tracking is done, synthetic aperture radar (SAR) and discuss the special requirements of steath radar systems (i.e. passive and active ESA).