Classifications

• • G—PHYSICS
  • G01—MEASURING; TESTING
  • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
  • G01S3/00—Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received
  • G01S3/80—Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received using ultrasonic, sonic or infrasonic waves
  • G01S3/802—Systems for determining direction or deviation from predetermined direction
  • G01S3/805—Systems for determining direction or deviation from predetermined direction using adjustment of real or effective orientation of directivity characteristics of a transducer or transducer system to give a desired condition of signal derived from that transducer or transducer system, e.g. to give a maximum or minimum signal
  • G01S3/807—Systems for determining direction or deviation from predetermined direction using adjustment of real or effective orientation of directivity characteristics of a transducer or transducer system to give a desired condition of signal derived from that transducer or transducer system, e.g. to give a maximum or minimum signal the desired condition being maintained automatically

Definitions

• the present invention relates to measuring and tracking systems and, more particularly, to a system wherein the manner in which a camera tracks a moving subject is dependent upon the context and history of the subject's movements.
• a typical automated tracking system utilizes a two step process to eliminate any errors between the position of a tracked subject and the camera direction.
• the tracking system measures the angular error between the tracked subject and the orientation of the tracking device/camera.
• the error information is used to generate a correction signal to control the motion of the tracking device/camera.
• the new camera position is then utilized in calculating future angular errors.
• the resulting tracking is mechanical with a tendency to be just as annoying to the viewer as if the tracking was very inaccurate. For example, constant readjustment of the camera position to every motion of the subject can cause rapid eye fatigue in the viewer.
• a human cameraman directs a camera to a subject with less than complete accuracy.
• the cameraman further adjusts the tracking style according to the context in which the subject moves and the immediate past history of the subject's motions.
• the resulting video contains minor tracking inaccuracies which are visually more pleasing to a human viewer than the exacting and mechanical tracking of an automated system.
• Another object of the present invention is to provide a system which allows a camera to track a moving subject in a specific style which is appropriate to the context in which the subject is moving.
• a further object of the present invention is to provide a system which allows a camera to track a moving subject in a style which is determined by the system, based on the past history of the subject's motions.
• a system for permitting a device to point at and track the movements of a subject comprises a signal source and a control means which receives signals from the source and generates therefrom control signals.
• a memory means stores plural sets of data values, each of the sets representing the operating parameters of said system in a specific operational mode. Means are provided for selecting one of said plural sets of data values.
• a processor processes the control signals in accordance with the selected data values and produces processed control signals.
• a drive unit moves the device about an axis in response to the processed control signals to maintain the device pointed at the subject.
• the means for selecting one of the plural sets of data values comprises a user-defined signal for indicating the desired data value set and its corresponding operational parameters.
• the means for selecting a data value set is a function of the processor which, when processing the control signals, determines whether the selected data value set is appropriate for the moving subject. If not, the processor selects a new data value set and thereby changes the operational mode of the system.
• FIG. 1 is a pictorial illustration of the drive unit of a tracking system suitable for use with the present invention
• FIG. 2 is a diagrammatic illustration of the subject, such as an athlete, wearing the transmitter unit of a tracking system suitable for use with the present invention
• FIG. 3 is a functional block diagram of a drive unit according to the invention.
• FIG. 4 is a flow chart of the error correction process of prior art automatic tracking systems
• FIG. 5 is a flow chart of the error correction process executed by the drive unit of FIG. 3 according to a first embodiment of the present invention
• FIG. 6A is a flow chart of the algorithmic steps for calculating the dead zone of the process of FIG. 5;
• FIG. 6B is a flow chart of the algorithmic steps for shifting the dead zone of the process of FIG. 5;
• FIG. 6C is a flow chart of the algorithmic steps for continuity control of the process of FIG. 5;
• FIG. 6D is a flow chart of the algorithmic steps for motion control of the process of FIG. 5;
• FIG. 7 is a flow chart of the error correction process executed by the drive unit of FIG. 3 according to a second embodiment of the present invention.
• FIG. 8 is a flow chart of the algorithmic steps for automatic context determination of the process of FIG. 7.
• the present invention provides a smart tracking system which in an exemplary embodiment, is an active tracking system comprising two units.
• the first unit is a small, battery-operated, omni-directional ultrasonic transmitter which is carried by the subject to be tracked.
• the transmitter produces brief, bursts of sound at frequencies above the range of human hearing.
• the second unit comprises a motorized camera drive capable of tilting and panning an attached camera. It should be understood that the present invention may also be implemented with a passive tracking system, not requiring a transmitter worn by the subject.
• a battery-operated ultrasonic receiver is attached to the camera drive.
• the receiver is tuned to the same frequency as the transmitter carried by the subject.
• the receiver employs three ultrasonic microphones, together with electronic circuitry to detect and amplify the ultrasonic signals generated by the transmitter and intercepted by the microphones.
• the microphones are located at three of the four corners of an imaginary square centered on and perpendicular to the axis of the camera lens. The signals emanating from the subject will arrive at all three microphones simultaneously only if the camera is pointed directly at the subject.
• the receiver measures the relative arrival times of signals received by the microphone, determines the direction and degree to which the camera is not pointed directly at the subject, and causes a servo-system to drive motors to re-orient the camera (and receiver) to point toward the subject matter. This re-orientation or tracking is done in a manner that is visually pleasing to human viewers, and is adaptable to a variety of contexts.
• a television or motion-picture camera or other device 10 (hereinafter referred to as a camera, without loss of generality) is attached to a platform or mount which is part of a motorized drive unit, generally designated 12, that is in turn mounted to a tripod 14 or other support.
• Drive unit 12 can tilt camera 10 up and down about horizontal axis 16 from about 30° below the horizontal to about 30° above the horizontal.
• the drive unit can also pan the camera left and right about vertical axis 18 through repeated 360° circles.
• the drive unit responds to ultrasonic signals received via microphones 20A, 20B and 20C to control direct-current motors (not shown) to move the camer mount.
• a subject 22 as shown in FIG. 2, who is to be recorded is provided with an ultrasonic transmitter 24 to be worn or carried on his or her person.
• the transmitter unit may include, for example, an elastic belt 26 with buckle for wearing around the waist of the subject. Attached to the belt are three or more transducer units (only one transducer unit 28, being shown in the drawing) connected by flexible cables embedded in belt 26.
• Drive unit 12 includes two direct-current motors 42 and 44, respectively. Through conventional mechanisms, motor 42 tilts the camera mount up and down and the motor 44 pans the camera mount left and right. Since the camera is securely fastened to the mount, it undergoes the same motions.
• the motors are driven by electronic circuitry located on the circuit boards (not shown in FIG. 2) inside the drive unit housing. The circuitry is powered by batteries (also not shown).
• Three detector circuits 30A, 30B and 30C are shown in the block diagram of FIG. 3, one for each of the three ultrasonic transducers 20A, 20B and 20C, respectively. Also shown are two motor driver circuits 32 and 34, one for each of the two DC motors 42 and 44, respectively. A controller circuit 36 and digital processor 46 operatively connect the detectors to the motor drivers.
• Each of the transducers 20A, 20B and 20C receive the ultrasonic bursts emitted by the transmitter unit and supplies a corresponding electrical signal to its associated detector circuit.
• the detector circuit detects the presence of an ultrasonic burst and produces a rectangular pulse whose leading edge coincides with the beginning of the detected burst.
• pulses 50A, 50B and 50C are routed to controller 36, which measures the delay between the pulses from the three detectors and loads digital error registers (not shown) with binary signals, VERR and HERR. These 4-bit signals represent the angular error along the vertical and horizontal axes, respectively, between the camera direction and the subject. VERR and HERR are updated each time an ultrasonic burst is received.
• transmitter unit 24, and drive unit 12 including direct-current motors 42 and 44, and the manner in which the drive unit 12 tracks the location of transmitter unit 24 to keep the camera pointed at the subject, is discussed in detail in U.S. Patent 4,980,871, which is incorporated herein by reference, and will not be repeated for brevity.
• Digital processor 46 receives the VERR and HERR signals over busses 52 and 54, respectively.
• Processor 46 may be a microprocessor.
• One such microprocessor suitable for use with the present invention is the model MC 68705C8 available from Motorola, Inc., Phoenix, Arizona.
• Processor 46 manipulates the values of VERR and HERR, in a manner to be described more fully below, according to the particular tracking style desired.
• the processed error signals, PVERR and PHERR are converted into DC motor drive voltages by identical horizontal (i.e., pan) and vertical (i.e., tilt) motor driver circuits 32 and 34, respectively which drive the tilt and pan motors in the proper direction and at the proper speed to keep the camera pointed at the subject.
• Smart Tracking Processes are identical horizontal (i.e., pan) and vertical (i.e., tilt) motor driver circuits 32 and 34, respectively which drive the tilt and pan motors in the proper direction and at the proper speed to keep the camera pointed at the subject.
• a typical automated tracking system utilizes the elementary two step process illustrated in FIG. 4, to eliminate any errors between position of a tracked subject and the camera direction.
• the tracking system measures the angular error between the tracked subject and the tracking device/camera, as indicated by block 70 in FIG. 4.
• the error information is used to generate a correction signal to control the motor speed and motion of the tracking device/camera, as indicated by block 72.
• the new position of the tracking device/camera is then utilized in calculating future angular errors.
• error signals VERR and HERR are manipulated by digital processor 46 in accordance with a 2-bit, user-supplied, STYLE signal, provided via buss 60, which indicates the desired style of tracking.
• the STYLE signal selects a set of parameters stored in a digital processor 46 or in a separate semiconductor memory.
• the parameters comprise a number of data variables which represent acceptable system limitations for the selected style or mode of tracking, as well as variables representing real time data about the position of the tracked subject.
• the set of system parameters are used by processor 46 to perform a series of calculations to adjust the values of error signals VERR and HERR.
• the processed signals PVERR and PHERR are supplied over busses 62 and 64, respectively, to control motor drives 32 and 34, respectively, which, in turn, cause camera 10 to point to the subject as closely as possible within the constraints that the resulting motion be pleasing to the eye of the viewer.
• the user-supplied STYLE signal describing the desired tracking style is replaced with a signal generated by digital processor 46, based on the past history of the tracked subject's motions as described by the most recent values of some of the variables.
• FIGS. 5 and 7 illustrate the processes executed, according to the first and second embodiments, respectively, by digital processor 46 to process signals VERR and HERR.
• the processes adjust the response characteristics of the tracking system to give it attributes that result in videos that are pleasing to a human viewer.
• Each of these algorithmic processes represented by blocks 78, 80, 82, 84, 86, 88, 90 and 92, are described in detail below.
• the user/operator of the tracking system selects the desired style or mode in which to track a subject. This selection is done upon powering the system, as indicated by block 78 of FIG. 5.
• the user communicates the selected mode to digital processor 46 with the two-bit STYLE signal, supplied over line 60.
• processor 46 may be programmed to select a default set of parameters representing the initial tracking mode or style upon power up.
• processor 46 sets all relevant variables to the appropriate values associated with the particular style of tracking selected.
• a first process as executed by digital processor 46 to produce a tracking style which is more pleasing to human viewers includes a deadzone controller, represented by process block 82 in Fig. 5 and implemented with the algorithmic steps illustrated in Fig. 6A.
• a deadzone controller represented by process block 82 in Fig. 5 and implemented with the algorithmic steps illustrated in Fig. 6A.
• the present invention allows the subject to move anywhere within a central portion of the camera frame referred to as the "deadzone", without adjusting the camera position.
• the deadzone controller is implemented with a. number of variables, stored in digital processor 46, which represent various system parameters. The value of these parameters varies according to the desired tracking style specified by the user. Processor 46 manipulates the variables to achieve the desired controller function.
• the deadzone is described by a pair of angular values (DEAD LEFT, DEAD RIGHT) that correspond to the range of angular error that is tolerated in each direction before the camera position is readjusted.
• the deadzone is symmetric about the center of the subject and is set to (- THETA, + THETA), where the variable THETA is an angular value and is a parameter of the tracking context.
• the total angular range of the deadzone is equal to twice the value of THETA.
• the ERROR variable represents the value of the angular ERROR along one axis between the tracked subject and the direction of the tracking device/camera.
• the deadzone controller takes the value of the ERROR variable and remaps the variable into a new ERROR variable which accounts for the deadzone.
• the ERROR variable will represent the angular error on the horizontal axis HERR, however it could equivalently represent the angular error along the vertical axis, VERR.
• the deadzone is represented by the values (DEAD LEFT, DEAD RIGHT).
• Digital processor 46 determines whether the value of ERROR is still within the deadzone, as indicated by decisional step 96. If the value of the ERROR variable is within the deadzone, the ERROR variable is set to zero, as indicated by processing step 94. When the value of ERROR variable is not within the deadzone, the subject has moved out of either side of the deadzone.
• the direction of the ERROR is determined, as indicated by decisional step 98. If the value of ERROR is to the left of the deadzone, the value of DEAD LEFT is subtracted from the value of ERROR, as indicated by processing step 100. Conversely, if the value of ERROR is to the right of the deadzone, the value of DEAD RIGHT is subtracted from the value of ERROR, as indicated by processing step 102. In this manner, as the subject moves out of either side of the deadzone the angular error is transformed by subtracting the edge of the deadzone so that it describes how far in degrees from the edge of the deadzone the subject is rather than how far from the center of the camera frame.
• the deadzone may be implemented as a zone of reduced response.
• the response to the system i.e. the relation between ERROR and MOTOR SPEED
• the reduced response zone can be set to a very low gain rather than a gain of zero.
• This implementation allows the system to slowly recenter the camera on the tracked subject instead of not moving at all as in the case with the deadzone implementation.
• such an implementation requires very close control over the motors 42 and 44 at low speeds, making the implementation practical only with precision stepping motors and not typical analog motor controllers.
• such an implementation complicates the processing performed by the directional continuity controller and motion controller, as explained hereinafter.
• a second process executed by digital processor 46 comprises a deadzone shift controller, represented by block 84 in Fig. 5 and implemented by the algorithmic steps illustrated in Fig. 6B.
• the purpose of the deadzone shift is to allow the earner-, to catch-up with the moving subject rather than always lagging by THETA.
• a number of limits are initialized.
• the va r iable MAXDEAD represents the value of THETA plus tie maximum number of degrees that one edge of the deadzone may be shifted toward the center of the camera frame.
• the MAXDEAD value is a function of motor speed.
• the motor speed for which MAXDEAD is greater than THETA, its minimum value, is the point at which the deadzone is allowed to shift.
• the motor speed for which MAXDEAD is equal to (2 • THETA) is the point at which the deadzone has shifted sufficiently for the tracking system to begin leading the subject.
• variable DEADSHIFT represents the maximum rate in degrees per second that the deadzone may be shifted, and is likewise a function of the motor speed.
• the value of DEADSHIFT is typically some small fraction of the motor speed.
• the values of the variables of MAXDEAD and DEADSHIFT are set to a function of the value of MOTOR SPEED, which represents the current value of the speed of the motor 44, as indicated by processing step 104.
• the shift of the deadzone occurs in the following manner.
• the value of ERROR is compared to the right edge of the deadzone, DEAD RIGHT, as indicated by decisional step 106. If the value of ERROR is beyond the right edge of the deadzone, the deadzone is shifted left, as indicated by processing step 108. This leftward shifting is accomplished by subtracting the value of DEADSHIFT from both the value of DEAD LEFT and the value of DEAD RIGHT. If the value of ERROR is not beyond the right edge of the deadzone, a comparison is made to determine whether the ERROR value is beyond the left edge of the deadzone, as indicated by decisional step 110.
• the deadzone is shifted to the right, as indicated by processing step 112. This rightward shifting of the deadzone is accomplished by adding the value of DEADSHIFT to both the value of DEAD RIGHT and the value of DEAD LEFT.
• the new values of DEAD LEFT and DEAD RIGHT now represent the shifted deadzone which is then compared to the value MAXDEAD to determine if the deadzone has been shifted beyond the allowable limit of MAXDEAD.
• the value of DEAD LEFT is compared to the value of -MAXDEAD, as indicated by decisional step 114. If the left edge of the deadzone is beyond the allowable limit, i.e. DEAD LEFT ⁇ -MAXDEAD, the deadzone is shifted to the left limit. This is accomplished by setting DEAD LEFT to -MAXDEAD and setting DEAD RIGHT equal to (2 • THETA - MAXDEAD), as indicated by processing step 116.
• the right edge of the deadzone is compared to the value of MAXDEAD, as indicated by the decisional step 118. If the right edge of the deadzone is beyond the allowable limit, the deadzone is shifted to the right limit by setting DEAD RIGHT equal to MAXDEAD and setting DEAD LEFT equal to (MAXDEAD - 2 • THETA) , as indicated by processing step 120.
• the limit of the deadzone, MAXDEAD will eventually shift to greater than 10° and the deadzone will begin to shift to ( ⁇ -10, ⁇ 0), with the effect of leading the tracked subject.
• the maximum deadzone, MAXDEAD will decrease again to 5°, and the deadzone will begin to shift back towards its normal rest value of (-5, +5) which corresponds to its normal centered position, allowing the subject to move within the symmetric deadzone in the camera frame.
• the size of the deadzone, (2 • THETA); the function controlling the rate at which the deadzone is shifted, DEADSHIFT; and the function controlling the amount the moving subject is led by the camera, MAXDEAD, are all parameters that may be adjusted depending upon the tracking strategy, as specified by the use .
• the choice of the DEADSHIFT function effects the overall stability of the control servo-loop of the system and should be chosen to avoid instability, as can be appreciated by one skilled in the art.
• a third process executed by digital processor 46 to produce a tracking style which is more pleasing to human viewers includes a directional continuity controller, represented by process block 86 in Fig. 5 and implemented by the algorithmic steps illustrated in Fig. 6C.
• Human viewers are very sensitive to changes in the direction of the motion of an image. Accordingly, it is important to minimize the number of times the direction of camera motion changes, viewers are particularly sensitive to directional changes that occur either at low speeds or when the tracking device/camera moves from high speed motion to a stop and reverses direction.
• the directional continuity controller guarantees that there is not a series of directional changes that would be annoying to the viewer.
• the directional continuity controller is implemented with a number of variables, stored in digital processor 46 and representing various system parameters. The values of these parameters vary according to the desired tracking modes specified by the user. Processor 46 manipulates the variables to achieve the desired controller function.
• the variables relevant to the directional continuity controller are as follows.
• the LAST DIRECT variable represents, the sign of the direction in which the camera was last moving.
• the ERROR DIRECT represents, the sign of the angular error between the camera direction and the subjects position, i.e., the direction the camera must move to close-in on the subject.
• the STABLE TIME variable represents, in seconds, the elapsed time since the last motion of the camera.
• the REVERSE DELAY variable represents the minimum time that the subject is required to move in a new direction before the camera will begin to move. Until the time specified by the REVERSE DELAY variable is met the camera is not allowed to change direction. Note that some portion of the REVERSE DELAY time will be spent as the subject crosses the dead zone of the camera frame.
• the algorithmic steps executed by the directional continuity controller are illustrated in Fig. 6C.
• the LAST DIRECT and STABLE TIME variables represent the last direction of camera motion and the elapsed time since such motion, respectively. If other control processes try to move the camera in the same direction, such motion is allowed and the elapsed time is reset.
• processor 46 compares the value of the LAST DIRECT variable to zero, as indicated by decisional step 122. If the value of LAST DIRECT equals zero, indicating that the camera has not recently been in motion, the camera is allowed to move in the direction of the error, as indicated by processing step 126.
• the value of the STABLE TIME variable is set to zero and the value of the LAST DIRECT variable is set to the value of the ERROR DIRECT variable.
• the LAST DIRECT variable is not equal to zero, indicating that the camera has recently been in motion
• the value of the ERROR DIRECT variable is compared to the LAST DIRECT variable, as indicated by decisional step 126. If the ERROR DIRECT variable equals the LAST DIRECT variable, indicating that the error direction and the last camera direction are the same, processing step 126 is executed as previously described.
• the elapsed time since the last motion is compared to the REVERSE DELAY constant, as indicated by procedural step 128. If the value of the STABLE TIME variable is greater than or equal to the REVERSE DELAY value, then processing step 126 is executed as previously described. If however, the elapsed time since the last motion is less than the time specified in the REVERSE DELAY variable the camera is not allowed to reverse directions, as indicated by processing step 130. In processing step 130 the value of the STABLE TIME is incremented and the ERROR variable is set to zero. The value of the LAST DIRECT variable remains unchanged.
• the directional continuity controller allows the camera to move in the direction of its most recent motion.
• a minimum threshhold of time is required before the camera is allowed to reverse directions.
• the value of the time delay specified in the REVERSE DELAY variable is a parameter of the tracking context, as indicated by the user-specified STYLE signal. Typical ranges for the value of the REVERSE DELAY variable can be from 1/6 of a second to greater than 5 seconds, without causing instability in the servo loop of the tracking system.
• a fourth process executed by digital processor 46 includes a motion controller, represented by process block 88 in Fig. 5 and implemented by the algorithmic steps illustrated in Fig. 6D.
• the background typically represents a greater portion of the picture than does the subject. If the motion of the scene against the background is jerky, the picture will be disconserting to the viewer, even if this results from tracking the inconsistent motions of the subject.
• the motion controller performs a pseudo filtering function in the response of the system by selecting acceleration time constants that are different from the deceleration time constants for the motors driving the camera. As a result, the camera motion starts smoothly while avoiding any tendency to overshoot the subject when the motion is halted. In addition to the filtering function, the motion controller places limits on the maximum velocity and acceleration of the camera depending on the context of the subject's motion.
• the motion controller is implemented with a number of variables, stored in digital processor 46 and representing various system parameters. The value of these parameters varies according to the desired tracking modes selected by the user.
• the variables relevant to the motion controller are as follows.
• the variable ERROR is defined as similar to that used by the dead zone controller.
• the variable LASTERROR represents the previous value of ERROR.
• the variable MAXACC represents the maximum allowable acceleration, measured in degrees per second squared, for the selected tracking mode.
• the variable MAXVEL represents the maximum allowable velocity, measure in degree per second, for the selected tracking mode.
• K ACC represents the exponential smoothing time constant that controls the rate of which the motor 44 is accelerated.
• the value (1 - K ACC) is the fraction of the exponentially averaged history of the signal that is left after one time step.
• the variable K DEACC represents the exponential smoothing time constant that controls the speed at which motor 44 is decelerated.
• the value (1 - K DEACC) is the fraction of the exponentially averaged history of the signal that is left after one time step.
• the variable K DEACC is greater than or equal to the variable K ACC to make the acceleration as smooth as possible while avoiding overshooting of the subject when the camera motion stops.
• the values of both K ACC and K DEACC are in the range between zero and one.
• the motion controller operates as follows. If the value of ERROR would lead to a camera velocity that is greater than the maximum allowable velocity prescribed by the tracking mode, the velocity is limited. This is accomplished by comparing the value of ERROR to the value of MAXVEL, as indicated by decisional step 132. If the value of ERROR is greater than value of MAXVEL, the velocity is limited by setting ERROR equal to the value of MAXVEL, as indicated by procedural step 134. If the value of ERROR is not greater than MAXVEL in step 132 it is compared to the value of -MAXVEL, as indicated by decisional step 136. If the value of error is less than -MAXVEL, the velocity is again limited by setting ERROR equal to -MAXVEL, as indicated by process step 138.
• LAST ERROR which is the acceleration
• the camera velocity is limited to the current camera velocity plus the maximum acceleration.
• the value of ERROR is compared to the sum of the LAST ERROR and MAXACC, as indicated by decisional step 140. If the value of ERROR is greater in step 140, the acceleration is limited by setting the value of ERROR to the sum of LAST ERROR and MAXACC, as indicated by process step 142. Otherwise, the value of ERROR is compared to the current velocity minus the maximum allowable acceleration, as indicated by decisional step 144. If the value of ERROR is less than the value of LAST ERROR minus MAXACC, the acceleration is again limited by setting the value of ERROR to LAST ERROR minus MAXACC, as indicated by processing step 146.
• digital processor 46 determines whether the camera is accelerating, as indicated by decisional step 148. If the camera is accelerating, a smoother acceleration is achieved as indicated by process step 150. A smooth acceleration is achieved by computing a smooth error value, NSERROR using the value of ERROR and the value of the last smoothed error, LSERROR, with Eq. (1) below.
• NSERROR (1-K ACC) • LSERROR + K ACC • ERROR Eq. (1)
• Process step 150 is the equivalent of low pass filtering.
• the new smooth error, NSERROR, of Eq. (1) is then used to control the speed of motor 44.
• the above pseudo low pass filtering technique uses different time constants when accelerating and decelarating.
• the sensitivity of viewers to the smoothness of acceleration is greater than the sensitivity to the smoothness of decelaration.
• the ratio between K ACC and K DEACC is typically 1:2.
• the user of the tracking system specifies a tracking context so that the system performance is specifically tuned for the context of the subject's motion.
• a typical selection of modes or tracking styles may include slow, general and fast action modes.
• a slow action mode good for lectures, walking sports, playing with young children or any activity without running, is be characterized by slow, smooth motion with a tendency to lag the subject if the subject engages in a quick motion.
• a general action mode good for general play is appropriate for smooth motion with a tendency to lag the subject only when the subject is at a full run.
• a fast action mode good for tennis, basketball, soccer or any activity with full speed running and quick changes of motion would be appropriate for quickly tracking even the fastest moving subject and has a tendency to react more quickly than is visually pleasing if the action isn't fast enough to mask it.
• the user-supplied STYLE signal describing the desired tracking style
• a signal generated by digital processor 46 is replaced with a signal generated by digital processor 46, based on the past history of the tracked subject's motion as described by the most recent values for some of the system variables, as explained hereinafter.
• Figure 7 illustrates processes executed by digital processor 46 to process signals VERR and HERR.
• processor 46 automatically adjusts the previously described processes so that the response characteristics of the tracking system give it attributes that result in videos that are pleasing to the human viewer.
• the algorithmic process represented by blocks 78, 80, 82, 84, 86, 88, and 90 are similar to those described with reference to Figure 5 and the first embodiment of the present invention.
• a fifth process executed by digital processor 46 includes an automatic context controller, represented by process block 92 in Figure 7 and implemented by the algorithmic steps illustrated in Figure 8. Since a part of the function of the tracking system is to determine the position of the subject being tracked, there is a significant amount of information available that can be used to predict the style of future motions. Typically, the real activities that the subject engages in are quite stable in their range of motions.
• the parameters of the subject's motion are stored and then compared with the parameters expected to be associated with such a tracking context.
• the extracted parameters include: the spectral characteristics of the motion, the percentage of time the current context tracks properly, the maximum angular acceleration of the subject, and the minimum angular speed of the subject. These parameters are chosen more for ease of calculation than as being the best parameter .
• the variables relevant to the automatic context determination controller are as follows.
• the CONTEXT variable is the equivalent to the STYLE signal and designates the currently selected tracking context or mode.
• the MISSTRACKING variable represents the length of time the subject is outside of the dead zone.
• the MAXMISSTRACK variable represents the maximum allowable percentage of time t " ' e subject may be allowed outside the dead zone for the particular tracking style selected by the CONTEXT variable.
• the AVG MOVE variable represents the average time the subject moved in a single direction.
• the MOTOR SPEED variable is the same as previously defined regarding the first embodiment.
• the DELTA ERROR variable represents the change in angular error, in degrees, since the angular error of the subject was last measured.
• DELTA ERROR is added to the MOTOR SPEED to give the speed of the subject, CURRENT SPEED.
• the DELTA SPEED variable represents the change in CURRENT SPEED, in degrees per second, since the angular error of the subject was last measured.
• the MAX SPEED variable represents the maximum angular speed of the subject.
• the MAX ACC variable represents the maximum angular acceleration of the subject.
• the automatic context determination controller 92 determines whether the current tracking context is appropriate for the motion of the subject. Referring to Figure 8, with each new value of ERROR, the amount of time in which the current system mode mistracked the subject is computed, as indicated in processing step 160. The absolute value of the ERROR variable is used to determine whether the subject is currently being mistracked. If the absolute value of the ERROR variable is less than or equal to a maximum allowable error, the value of MISSTRACKING is not incremented, indicating that the subject is still moving within the acceptable zone. However, if the absolute value of the ERROR variable is greater than the maximum allowable angular error, the value of MISSTRACKING is incremented, indicating that the subject has left the acceptab.e zone.
• the percentage of the time the system is mistracking is calculated by dividing MISSTRACKING by the total time, and compared to the maximum allowable percentage for the current mode, MAXMISSTRACK, as indicated in decisional step 162. If the percent of time the system is mistracking the value of MISSTRACK is greater than the value of MAXMISSTRACK, the value of the CONTEXT variable is set to the next faster contes;r or tracking mode, as indicated by processing step 164. Next the maximum angular speed of the subject, MAX SPEED, is calculated as indicated by processing step 166. This is done by setting the value of MAX SPEED to the maximum of the last value of MAX SPEED and the absolute value of the current speed of the subject, CURRENT SPEED, which is computed as in Eq. (3) below.
• the absolute value of the speed of the subject is used so the value of MAX SPEED will be direction independent.
• MAX ACC the maximum angular acceleration of the subject
• the absolute value of the difference in speed of the subject is used so the value of MAX ACC will be direction independent.
• the average time of the subject's movement is computed as indicated by processing step 170.
• the average time is calculated by simply computing the average time between directional changes and setting the AVG MOVE variable to such average.
• each available context or mode is a range of values that correspond to appropriate ranges of motion of the subject.
• a determination is made whether the current values characterizing the subject's motion more closely match those of a slower context, as indicated by decisional step 172. If the characteristics of the subject's motion will closely match a slower mode or context, the value of CONTEXT is set to the next slower context, as indicated by processing step 174. Otherwise, the tracking context remains the same.
• a GAIN variable may be necessary for controlling the overall gain of the servo-loop in the tracking system.
• a MOTOR OFFSET value would allow the system to account for the minimum start up voltages required when utilizing analog motors. By adding a MOTOR OFFSET to the ERROR value during the motor control process the zero-crossing response of such a motor can be evened out. Setting the MOTOR OFFSET variable to zero has the effect of adding to the dead zone in a way which is difficult to compensate for with other variables in the system.
• the overall gain of the tracking system i.e., the response of MOTOR SPEED to ERROR, may range over a factor of 4:1.
• the ratio of K ACC to K DEACC may range over a factor of 1:3.
• the values of K ACC and K DEACC may each range over a factor of 3:1.
• the REVERSE DELAY time may range from 1/6 second to 6 seconds.
• the value of THETA may range over a factor of 5:1.
• Such systems may be active, as in the exemplary embodiments, or may be passive tracking systems, such as infrared tracking systems which do not require a transmitter to track a subject.
• Other such tracking systems include systems which use frame to frame motion comparison and pattern recognition techniques to identify portions of an image that are to be maintained within the camera frame.
• Still other systems may be sampled, as in the exemplary embodiments, or may be continuous, and may use infrared, radio frequency, or sonic transmission to communicate between the transmitter/receiver.
• Other systems may use precision stepping motors to control the camera or tracking device instead of analog motor controllers.
• the algorithmic processes represented by blocks 78, 80, 82, 84, 86, 88, 90 and 92 are implemented with a plurality of system variables and a programmed digital processer. It will be obvious to those reasonably skilled in the art that functionally equivalent processes may be achieved with an implementation using small, medium or large scale logic components as well as analog circuit elements. Such hardware implementations are well within the scope of those reasonably skilled in the art and are within the scope and spirit of the present invention.
• HERR horizontal angular error signal
• VERR vertical angular error signal
• digital processor 46 may perform the control functions for both the horizontal and the vertical angular error signals, as shown in FIG. 3, or that two such digital processors may be used, one for each axis of motion.
• contemplated tracking systems may allow different tracking modes to be assigned to each axis.
• a contemplated tracking system may provide for automatically or user-selected tracking modes which fix the horizontal and vertical axis tracking styles to each other, the parameters of such modes being defined for a specific two-axis activity.

Landscapes

• Physics & Mathematics (AREA)
• Engineering & Computer Science (AREA)
• General Physics & Mathematics (AREA)
• Radar, Positioning & Navigation (AREA)
• Remote Sensing (AREA)
• Studio Devices (AREA)
• Closed-Circuit Television Systems (AREA)

Applications Claiming Priority (2)

Publications (1)

Family

ID=24310120

Family Applications (1)

Country Status (4)

Families Citing this family (4)

Family Cites Families (7)

• 1991
  • 1991-09-04 AU AU88533/91A patent/AU8853391A/en not_active Abandoned
  • 1991-09-04 JP JP3517834A patent/JPH06509449A/ja active Pending
  • 1991-09-04 WO PCT/US1991/006316 patent/WO1992004641A1/en not_active Application Discontinuation
  • 1991-09-04 EP EP19910918522 patent/EP0549716A1/de not_active Withdrawn

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events