947,378. Radar tracking system. SOC. NOUVELLE D'ELECTRONIQUE. July 28, 1960 [July 28, 1959; Jan. 20, 1960; Feb. 8, 1960], No. 26384/60. Heading H4D. Relates to a two-dimensional automatic tracking pulse radar comprising a directional aerial rotating continuously in azimuth, a window gate for selecting echoes from a target within a window area of predetermined shape and adjustable position, means for generating during any one aerial rotation period a pair of position error signals representing the difference between the actual co-ordinates of the target as determined by the radar and the co-ordinates of the centre P of the window area (i.e. the predicted target co-ordinates), means for applying a predetermined fraction α of said position error signals to produce a step displacement of the window in such a direction as to reduce said errors, means for summing a predetermined fraction # of the position error signal with corresponding signals obtained during the previous aerial rotation periods to give a pair of velocity signals and means for applying said velocity signals to cause an additional progressive displacement of the window so that the window tracks the target, a separate tracking loop being provided for each target to be tracked. According to the present invention the predicted coordinates are expressed as binary numbers and all the operations are effected by digital processes except the velocity correction which may be effected by a combined digital-analogue process. The digital computation circuits preferably employ transistors. To initiate tracking a new target, the initial position of the window is set manually. Cartesian x, y or polar d, 0 co-ordinates may be employed, the corresponding shape of the window area being square, Fig. 1, or an annular segment, Fig. 11. Due to the finite width of the radar beam, a single target will produce a number p of echoes as the beam sweeps through the target and the actual co-ordinates of the target are taken to be the arithmetic mean of the co-ordinate of the corresponding echoes. An electronic marker indicating the position of the window is also produced on a P.P.I, display. Cartesian co-ordinate system, Figs. 1, 3, 5, and Figs. 2, 4, 6-10 (not shown).-Fig. 3 shows the main blocks of the system which are shown in more detail in Fig. 5. Since the x and y co-ordinates are handled in the same manner the x channel is shown in full and only those parts of the y channel are shown which cooperate with the x channel. Radar transmitter sync. pulses and the instantaneous bearing 0 of the aerial are applied over leads 14 and 13 to a known form of co-ordinate converter 100 which produces x and y pulse time-bases such that the x co-ordinate of a target is given by the number of pulses Nx that have been generated when the echo is received, the y co-ordinate being given by a corresponding number Ny. The x time-base is applied via a lead 20 to a counter 200 to which is also applied via a lead 22 a parallel binary number representing the complement #X of the predicted co-ordinate X (i.e. the x co-ordinate of the centre point P, Fig. 1, of the window). The counter 200 produces at an output 48 a binary number e equal to the difference between Nx and X and the complement e of e is produced at an output 49, the signals e and e being applied to gates 401 and 402 respectively. In order to facilitate determining the sense of the error e, the origin of the predicted X co-ordinate is offset so that X is always positive, Fig. 4 (not shown). When the magnitude of the error e is less than A, Fig. 1, an X gate pulse is produced by a matrix selecting circuit 300 such that the duration of the X gate pulse corresponds to the x dimension 2A of the window. The X gate pulse, a corresponding Y gate pulse and the video echoes (lead 38) are applied to a window gate circuit 310 to select the echoes within the window. The first of such echoes is selected by a discriminator 403, Fig. 6 (not shown), actuated by a pulse derived in a circuit 311 from the leading edge of the X gate pulse and the selected first echo opens the gate 401 to apply the corresponding binary error e 1 to a counter 412. The succeeding echoes open the gate 402, thereby applying the corresponding error complements #e 2 -#e p to two interconnected counters 413, 414 controlled by a logic circuit 420 to which the gated echoes are applied so that when the window gate 310 closes only the error complement #ep corresponding to the last gated echo remains stored in one of the counters 413, 414. When the window gate 310 closes a pulse derived in the circuit 311 from the trailing edge of the X gate pulse triggers on a bi-stable flipflop 429 to trigger on a pulse generator 430 whose output is applied to the counters 412 413 and 414 and when the number of pulses generated is equal to the binary member e p whose complement is stored in the counter 413 or 414 a matrix circuit 426-428 generates a transfer signal which triggers off the bi-stable flip-flop 429 -to block the generator 430. The outputs 52, 53, 54 from the counter 412 then constitute respectively (1) a binary number e m representing the mean error e1+ep/2 , (2) the complement #e m of e m , and (3) the sign of e m . The sign signal is applied through an OR circuit 419 to trigger a bi-stable flip-flop 437 controlling a pair of AND circuits 438, 439 which control gates 440, 441 to which the signals em and e m are applied such that when the transfer signal from the matrix circuit 426-428 is applied through a delay 431 to the AND circuits 438, 439 either the mean error e m or its complement e m is stored in the counter 436 according to the sign of the error. This stored number is converted in a digital-toanalogue converter 452 into a corresponding analogue voltage which is applied to a storage capacitor 454 through a gate 453 which is opened by a delayed transfer signal after a further delay in a delay circuit 432 so that after a number of aerial rotations the resultant voltage across the capacitor 454 represents the speed of the window and this voltage is applied through an amplifier 455 to control the recurrence frequency of a pulse generator 456. The delayed transfer signal from the delay circuit 432, after an additional delay in a delay circuit 433, also triggers on a bi-stable flip-flop 434 which triggers on a pulse generator 435 whose output is applied to the counter 436 and when the number of pulses generated is equal to the number stored in the counter 436 a matrix circuit 442 generates a control signal which triggers off the flip-flop 434 to block the generator 435. The outputs from the pulse generators 435 and 456 are combined and applied to the count-up or count-down inputs of a counter 460 in a prediction circuit 450, Fig. 3, through gates 457, 458 controlled by the sign flip-flop 437 so that the number of input pulses is added algebraically to the contents of the counter. The resultant number thus represents the predicted co-ordinate X and its complement X is applied via the lead 22 to the counter 200. A voltage of amplitude proportional to X is produced by a digital-to-analogue converter 459 and applied together with a corresponding Y voltage to indicate the position of the window on a P.P.I. display. The counter 200 is reset before the commencement of each radar pulse recurrence period and the counters 412 and 436 are reset after the mean error has been computed by pulses from a timing device 101 actuated by an advanced sync. pulse applied over lead 15. To initiate tracking a new target the predicted co-ordinate counter 460 and the speed storage capacitor 454 are reset by signals from a circuit 461 actuated by a " pick-up " signal applied over a lead 64 and an initial approximate co-ordinate X 0 is set in the counter 460 by generating in a manual controller 462 two signals representing the sign and magnitude of the co-ordinate X 0 . The sign signal is applied to the sign flip-flop 437 via the OR circuit 419 and the magnitude signal is converted in a time modulator 448 into a pulse pair whose spacing represents the magnitude, the pulses being applied to trigger on and off the flip-flop 434 controlling the pulse generator 435. The speed generator 456 may also be controlled manually. Since the co-ordinate error obtained in the first aerial rotation period does not represent the target velocity, the gate 453 applying the coordinate error analogue voltage to the speed storage capacitor 454 is blocked during said period by a bi-stable flip-flop 463 which is triggered on by the " pick-up " signal 64 and triggered off by a delayed control signal from the matrix circuit 442. Digital control of speed generator 456.-In a modification, Fig. 7 (not shown), of the Fig. 5 system, the target speed is represented by a binary number formed in an additional speed counter (480), Fig. 8 (not shown), by summing therein the successive pulse trains from the co-ordinate error pulse generator 435 and the speed pulse generator 456 comprises an R.C. oscillator, the magnitude of its frequency control resistance being determined by a plurality of relays controlled by the flip-flops in the speed counter, Figs. 9a and 9b (not shown). In a further modification, Fig. 10 (not shown), the outputs from the co-ordinate error pulse generator 435 are applied to the predicted co-ordinate counter 460 and the additional speed counter (480) through separate pulse multipliers or dividers which are adjustable to vary the factors α and (3 which determine the characteristics of the position and speed control. Polar co-ordinate system, Figs. 11, 12 and Figs. 13, 14, 14a-14c (not shown).-In this case the window area comprises an annular segment C1-C4, Fig. 11, the predicted coordinates of its centre being do, #o, and echoes within such area are selected by applying a radial range gate pulse and an azimuth gate pulse to the window gate 310, Fig. 12. The range c