signal have been demonstrated [16], A variant of this approach would be to use a laser pilot beam which is modulated at the microwave frequency. The system architecture is simpler if the pilot beam and master clock can be eliminated. This alternative is possible if the adjustment of the phase of the output oscillators is done using information from the receiver antenna. Without a pilot beam, the local phase must also be continually adjusted to compensate for the fact that the individual elements relocate in position due to flexing and rotation of the array. If the elements move at a velocity v, the oscillators must be reset on a time scale short compared to the time it takes the elements to move the distance of a wavelength, or a time scale of (v/c) f. The oscillator drift time must be long compared to the round-trip delay between the transmitter and the receiver. In this case the receiver antenna on the ground can generate the information required to reset the local oscillators. There are many possible approaches for doing this. In the following discussion, we assume: (1) low- drift local oscillators, (2) a master-computer with the capability of addressing each individual oscillator, (3) simple computation capacity at each oscillator. The simplest information for the receiver antenna to generate is the spot size and center location of the beam. Using this information, the master computer could then continuously tune each individual oscillator to minimize the spot size and keep it centered on the receiver, by sending each individual oscillator a message “advance/ delay your phase by At”. Clearly, with a number of oscillators on the scale of -108, adjusting the phase of any individual oscillator will not contribute significantly to changes in the spot, i.e. the signal to noise ratio would be too low. A technique must be used to adjust the oscillators globally. For example, linearly delaying the phase of the oscillators across the aperture would have the effect of slewing the beam. Focussing the spot could be done by adjusting the oscillator phase in fourier harmonics; e.g., commanding each oscillator to delay its phase by an amount At cos(nx/d), where x is the x- coordinate of the oscillator, d the diameter of the satellite, and n the order of the fourier harmonic. If this decreases the spot size, the change would be kept; if it increases the spot size a change the other direction would be tested. Many techniques exist for finding the global optimum. Terms in sin(x), sin(y) and cos(y) must likewise be optimized. Use of distributed intelligence at each element will allow considerable simplification of the design. Since the system must adjust a number of degrees of freedom equal to the number of oscillators, N-108, the minimum time required is N/f, e.g., for a 10 GHz beam, t >108/(109 sec-1) =.10 seconds. In practice it would be impossible to calculate the spot diameter as fast as once per cycle, and a minimum update time could require more like 103 cycles; thus, an update time -100 seconds. For the 10 GHz system this means that the oscillators must drift no more than Af/f<10'12.
RkJQdWJsaXNoZXIy MTU5NjU0Mg==