WO1994011070A1

WO1994011070A1 - Electronic tennis ball detection system

Info

Publication number: WO1994011070A1
Application number: PCT/AU1993/000586
Authority: WO
Inventors: George Seymour Gray
Original assignee: George Seymour Gray
Priority date: 1992-11-17
Filing date: 1993-11-16
Publication date: 1994-05-26

Abstract

A method of identifying the type and impact location of an object striking a surface, the method comprising the steps of providing at least one sensor means under the surface, the sensor means producing a signal in response to an object striking the surface in the vicinity of the sensor means, processing the signal to accentuate one or more parameters of the signal, and comparing the signal or parameters with predetermined signals or parameters whereby the type and impact location of the object is identified.

Description

Title : ELECTRONIC TENNIS BALL DETECTION SYSTEM

Technical Field

This invention relates to an electronic line fault detection system. Preferred embodiments of the invention distinguish between different types of objects impacting on or near the line.

The invention will herein be described with particular reference to a tennis ball detection system, but it will be understood that the invention is applicable to other fields of use, in particular, other ball games.

Background Art

As is well known, tennis is played on a court marked with lines. When a ball bounces on or close to certain of the lines an umpire must rule on whether the ball is within the designated area bounded by the lines. Because the tennis ball may be travelling at high speeds, it is often difficult to judge by eye whether the ball is "in" or "out". When tennis is being played professionally, the umpire's rulings may have considerable importance for players and other involved parties such as sponsors. A system which is consistent, reliable, and repeatable is desirable as it will minimise arguments, and reduce the need for trained officials.

A number of systems have been proposed for automatically detecting whether a tennis ball is "in" or "out" . Most such systems utilise a tennis ball which is provided with a conductive outer surface.

For example, one system involves a plurality of closely spaced parallel exposed electrical conductors which extend along or are adjacent the lines. Contact of the conductive outer surface of the ball with adjacent conductors completes an electrical circuit. If the conductors of the circuit completed by the ball are within an "in" area of the court the apparatus signals the ball is "in". Such systems are exemplified in patents US 3 883 860 and US 1 370 333.

Systems dependent on conductive connection between exposed conductors on the court surface are susceptible to failure as a result of resistive corrosion either of the conductors on the court or of the ball, or covering of the conductors by insulators such as dirt and to failure because of short circuits for example by moisture .

Moreover, electrically conductive balls behave differently to normal tennis balls, cause undue wear of rackets, and the conductive surface of the ball tends to fail as a result of wear.

In US 3 774 194 there is described a system which does not require exposed conductors. Instead a receiving antenna wire extends longitudinally of the court line and is buried beneath the line. There is provided a radio transmitter and a ball containing three coils at right angles acting as a resonant circuit tuned to the radio frequency of the transmitter. The ball, when in the vicinity of the court antenna, acts as a coupler causing vertically polarised radio waves from the transmitter to be sensed in the horizontal court antenna. In another embodiment, a ball having a ferro-magnetic metal or metal oxide included in the rubber composition thereof or having a thin layer of metal deposited on the outer surface of the rubber ball beneath the felt is used to unbalance a balanced bridge circuit outer layer. That system is subject to interference by external signals and the balls required for use in the system do not have the properties of normal tennis balls and are expensive to manufacture.

Other known systems include laser beam systems as are currently in use at Wimbledon. This system has the advantage of being able to use normal tennis balls but has obvious disadvantages in that the beam can be broken by objects other than the ball and accordingly the system covers service only and cannot be used in doubles matches or on boundaries likely to be stood on by a player.

Another system which has been used employs capacitive transducers. Capacitive transducers are likely to be swamped from signals from feet, racket strikes on the court, or the like. Also, variations in capacitance with ground conditions are likely to cause problems.

Another system is as proposed in the present Applicant's patent US 4 664 376. In this system a plurality of coils are buried adjacent a boundary line. When a metallic ball interrupts the magnetic field created by one of the coils, there is a corresponding reduction in amplitude in the coil and if this reduction exceeds a predetermined level then the ball is judged to have struck the relevant area.

None of the systems so far proposed has won wide acceptance and there is a continuing need for a satisfactory system of ball detection.

In summary, the systems proposed to date require special purpose balls and/or are overly sensitive to environmental factors and/or are unable to satisfactorily discriminate between various kinds of relevant events (such as ball strike, footfall, racquet strike, etc . ) .

Disclosure of the Invention

An object of the present invention is to provide a system which avoids at least some of the previously discussed disadvantages.

According to a first aspect of the invention there is provided a method of identifying the type and impact location of an object striking a surface, said method comprising the steps of: providing at least one sensor means under said surface, said sensor means producing a signal in response to an object striking the surface in the vicinity of the sensor means; processing said signal to accentuate one or more parameters of said signal; and comparing said signal or parameters with predetermined signals or parameters whereby the type and impact location of the object is identified.

Preferably the sensor means are located either side of a boundary line marked on the surface or comprise pairs of sub-sensors located either side of the boundary line to facilitate determination of the side of the boundary line on which the object has struck. As used herein, the term "under" in reference to the positioning of the sensor means under the surface is intended to encompass the situation where the sensor is placed on the surface or forms part of the surface. Preferably also, the parameters include amplitude change, phase change, and time related parameters (such as duration) with particular characteristics of one or more of the parameters being indicative of the type and location of the object which has struck the surface.

Preferably also, said pairs of sub-sensors are provided by a transformer including a primary coil energised with alternating current and a pair of secondary coils coupled to said primary coil. Most preferably, the coils are flat spirals which are located on a double-sided printed circuit board with the primary coil on one side and the secondary coils on the other. A particularly preferred embodiment is where the secondary coils are connected in series opposite and fitted with a variable magnetic coupling, this construction being known as a linear variable differential transformer or LVDT.

Preferably also, said processing includes an amplifier-amplitude limiter connected to each of the secondary coils. The two signals coming from the amplifier-amplitude limiters are then subtracted in a dynamic amplitude difference amplifier and the output from the dynamic amplitude difference amplifier is passed to a phase sensitive rectifier with the output from the phase sensitive rectifier being fed to a low-pass filter. The filtered signal is then digitised in an analog to digital converter and the digitised signal is input to a computer for comparison of the signal with predetermined signal characteristics to identify the type and location of the object which has struck the surface. Optionally, the signal may undergo further digital signal processing prior to being compared to discriminate the type and location of the object.

According to a second aspect of the invention there is provided a method of detecting the type and proximity of an object containing conductive material, said method comprising the steps of: providing at least one sensor means, said sensor means producing a signal in response to the proximity of the object to the sensor means; processing said signal to accentuate one or more parameters of said signal; and comparing said signal or parameters with predetermined signals or parameters whereby the type and proximity of the object is identified.

As used herein, the term "conductive material" is used in relation to materials which create a measurable disturbance when in the proximity of the sensor means. Such materials include magnetic and metallic materials.

Brief Description of the Figures

A preferred embodiment of the invention will now be described, by way of example only, with reference to the accompanying drawings in which: Fig. la and lb respectively schematically illustrate the coil arrangement of an LVDT with a pair of twelve turn secondary windings on one side and a twelve turn primary windings on the other.

Fig. 2 is a block diagram of the sensor and analog signal processing system.

Fig. 3a and 3b collectively illustrate a detailed circuit diagram of the self-compensating analog signal processing system used in a preferred embodiment of the invention. A list of components is given in Appendix 1.

Fig. 4 is a circuit diagram of the self adjusting amplifier-limiter extracted from Figure 3a.

Fig. 5 shows a voltage regulated power supply which provides ± 7.5V, 1A for all circuits.

Fig. 6 shows a crystal controlled 50kHz master frequency reference generator.

Fig. 7 shows a signal generated by an "out" ball

Fig. 8 shows a signal generated by an "in" ball.

Fig. 9 shows a signal generated by a metal racquet passing above the surface.

Fig. 10 shows a signal generated by a ball containing a conductive material passing over the surface.

Fig. 11 shows a signal generated by the impact of an aluminium racquet. Fig. 12 shows a signal generated by the impact of a wooden racquet.

Fig. 13 shows a signal generated by the impact of a metal racquet.

Fig. 14 shows a signal generated by the impact of a foot.

Best Mode

Referring to Figures la and lb, the sensor means is a double-sided printed circuit epoxy/glass-fibre board 1. On one side 2 of board 1 is the primary winding 3 and on the other side 4, two opposing secondary windings 5 and 6. In this embodiment all the windings are easily produced by the photo-etching of copper in respective flat rectangular spirals each having twelve turns. This particular arrangement has been found to be suitably sensitive, however, it will be appreciated that many other arrangements could also be used. Printed circuit board 1 may also be of a sandwich construction incorporating a reflective layer to further improve the performance of the transformer.

A transformer with secondaries connected in series opposite and fitted with a variable magnetic couple is known as a linear variable differential transformer or LVDT. Under balanced conditions the output is zero, while any changes affecting the balance of coupling between the windings causes an output. In the LVDT the energy transfer is from electrical to magnetic to electrical energy. For low frequency the operating frequency has little effect upon the effectiveness of the couple. As frequency is increased, the capacitance between primary and secondary becomes dominant. This puts an upper limit on the frequency which can be used. The lower limit of frequency is governed by the inductance of the primary and the current required to produce a certain value of magnetic field.

The LVDT has several inherent properties of practical importance.

(1) The frequency of oscillation is not critical if kept to a low value compared to the frequency corresponding to Q max, where Q is the quality factor of a winding and is a measure of the amount of energy stored in the winding for a given excitation frequency.

(2) The output from the LVDT does not need compensation for temperature because it is differential.

(3) The output contains information about the position of the ball.

(4) The LVDT can be made using printed circuit technology.

(5) The detector unit can be made in modules accurately when manufactured in quantity. Primary winding 3 of the sensor means, tuned with a capacitor to 50kHz, is excited with a low-level 50kHz signal. A weak, balanced magnetic field above secondary windings 5 and 6 is produced by this excitation signal. The sensor means is buried under the surface of the tennis court (or possibly on the surface) and objects striking the surface are sensed by measuring the minute dynamic differential disturbance of the sensor mean's balanced magnetic circuit. This disturbance is partially due to an interruption in the flux of the magnetic field above one of the secondary windings by the object and related capacitance change in the case where the object contains conductive material but is also due to slight mechanical deformation of the sensor means due to the impact of the object regardless of the conductive nature of material in the object. Hence, when an object strikes the court surface a signature signal shape is produced. Similarly, an object containing conductive material will produce a signature signal upon approach and departure in addition to its impact signature. Larger objects produce larger signals of longer duration as will be clear with reference to Figures 7-14. Additionally the signatures of simultaneously striking objects can be separated by post processing of the combined signature to isolate individual sources.

Objects containing conductive material produce a distinctive "header" signal prior to impacting. This signal is indicative of the approach of the object into the field and will also be produced when the object doesn't impact but rather passes in close proximity to the sensor. This gives an additional distinguishing feature to the signature signal of an object containing conductive material. Accordingly, a ball containing metal is particularly easy to identify. A ball containing a tuned metal loop housed within a mylar casing has also been found to be a particularly identifiable construction.

Each sensor means is coupled to a sensitive analog signal processing circuit, shown as a block diagram in Figure 2, which extracts information about the object's impact location and nature. The signal produced by the ball is extremely small, and special techniques are necessary to remove possible interference and noise whilst ensuring the very high gain required for a useful signal.

As previously mentioned the primary winding 3 of each sensor means is energised by a source of very accurate and stable frequency and amplitude. The source for synchronising the sensor means excitation is a 50kHz crystal master oscillator 7, as illustrated in Figure 6, located in the power supply case.

Operating all the sensor means at exactly the same frequency ensures reduced interference between adjacent sensors.

Oscillator 7 includes a 1 MHz crystal oscillator for providing a signal to buffer 8 which forms a digital signal which is sequentially divided by ten and two by intergrated circuits 9 and 10 to produce a 50 kHz square wave signal in conductor 11. This division by 20 can be performed by a single integrated circuit, if required. This square wave is then passed through a 50 kHz band-pass filter 12 to remove any unwanted harmonics and other spurious frequency components. The resultant signal is a 2V frequency reference input which is supplied to conductor 13 in Figure 2.

More particularly, the 2V, 50kHz frequency-reference signal is first applied to a local (per sensor) amplitude control circuit 14. This circuit stabilises the excitation amplitude across each sensor, accommodating for individual variations; the signal is then applied to a tuned power amplifier 15 which produces 0.2V across the primary winding 3.

With reference to Figures 2 and 4, the two secondary windings are connected to two separate amplifiers-amplitude limiters 16 and 17 respectively, which control very precisely the long-time average output signal amplitude. This guards against long term change such as caused by metal reinforcing, and changes caused by expansion, etc of such reinforcing. Only fast changes of secondary winding signal amplitude and phase, as produced by impacting objects or proximal mobile conductive objects, affect the output short-term changes of signal amplitude and phase of the amplifiers-limiters. These fast changes of amplitude and phase are transmitted with a voltage gain of about twelve. Long term variations are cancelled out by the compensating circuitry incorporated in the amplifier. The very effective signal-level of the amplifier-limiters accommodates for a range of average input signals from 0.14V to 0.18V RMS and produces an output voltage of 2.5V peak without any adjustment. This is the maximum range of average signal variation expected from external factors, e.g. spreading characteristics of individual installed sensors, iron bars embedded in the ground etc.

The two 50kHz signals coming from the amplifier-limiters are subtracted in the next circuitry block, the dynamic amplitude difference amplifier 18. Because the average output signals of the amplifiers-limiters are closely matched, only dynamic signal amplitude and phase differences, e.g. due to impacting objects, will produce a significant signal amplitude and phase change at the output of this amplifier, which provides an additional gain of about twelve.

Further processing of information about the impact location and object type or proximity is done in the next block, the phase sensitive rectifier 19. This part of the circuit converts the 50kHz signal, amplitude and phase modulated with the object presence information, into a base-band pulse signal, which represents the extracted signature. At this stage the base-band signal is of only about l-5mV peak, and contains residual 50kHz and other noise. An output amplifier low-pass filter 20 cleans the rectified signal of extraneous noises, limits its band width to about 330Hz and amplifies it by another factor of about 1200, bringing it to about 1-5V peak.

At this high voltage level and defined band width, the signal is sent to the A/D converter at the input of the computer, for further digital signal processing and recognition.

As can by understood from the above, the total voltage gain for.the dynamic object signal, from sensor to computer is above 160,000. This very large but stable voltage gain achieved for the relevant signals explains the very high sensitivity of the system.

To virtually cancel any residual output base-band signal due to static amplitude or phase differences between the outputs of the two amplifier-limiters 16 and 17, an additional signal-processing channel is used.

The filtered output of the phase sensitive rectifier 19 is applied to an absolute DC value circuit 21 and then to an automatic phase reference correction circuit 22.

The system is configured to maintain a 90° phase relation between the phase reference and the average input signals of phase sensitive rectifier 19. Thus the DC output of the phase sensitive rectifier 19 is practically nulled. Saturation of phase sensitive rectifier 19 is therefore avoided and maximum sensitivity for dynamic amplitude and phase changes is ensured. As a result, even large steady state imbalances between the two channels, of whatever origin (component spread or temperature variations) are corrected.

The secondary windings 5 and 6 are excited out of phase by single primary winding 3 and the difference is implemented by adding the two out of phase signals in a summing amplifier.

It has been found that the presence of metal, particularly reinforcing iron in the concrete floor underlying the court, tends to distort the field pattern. A sheet of aluminum placed under the sensor means controls the effect of the iron with the effect of the iron being reduced with increased thickness of the sheet of aluminum. This effect is known as the skin effect whereby the field entering material is attenuated exponentially with the thickness of the material. In this case the thicker the aluminum, the less field there is to couple with the iron.

It has also been found that digging a trench under the lines of the court thereby removing reinforcing steel (in the case of a concrete surface) or non-homogenous earth, filling the trench with a synthetic waterproof material such as hot-mix bitumen, and then mounting the sensors on the material assists in the performance of the system. The trench filled prevents rising damp and also provides a homogenous base underneath the sensors. Mounting the sensors in this manner assists in isolating the sensors from ground effects including ground-borne vibration and also provides homogenous characteristics in terms of thermal expansion and the like. The trench may be of any depth but 1.5 metres has been found to be particularly effective.

It is desirable that the sensor and associated signal processing circuit be permanently buried in the court and never require adjustment after initial assembly. To this end the circuitry, as illustrated in Figure 3, has been developed to be self-compensating. This circuitry is capable of compensating for any steady state disturbances or variations including, for example, the presence of reinforcing iron, metal net posts, etc. The componentary used is listed for convenience in Appendix I.

Signals generated by the impact of an object on the sensor means have been captured using a digital storage oscilloscope. Generally speaking there has been found to be a large swing away from zero, followed by a horizontal (constant) section, followed by a swing back to the zero position. This is the primary effect of the impact of the ball (or of any other object) on the transducer. With the LVDT the direction (positive or negative) of the signal swing indicates which of the secondary windings has been impacted upon.

There then follows a large swing away from zero in the opposite direction, followed by a horizontal section, followed by a swing back to the zero position. This is the effect of the bouncing of the transducer after the object which hit it departs. There is then usually a tail of damped oscillation of relatively small amplitude.

The discrimination between a ball, metallic racquet, foot or other object can be made in many ways. For example the duration of the initial pulse is much shorter for a ball than for a racket and the duration of the pulse is even longer for a foot as will be appreciated with reference to Figures 7-14. Referring in particular to Figures 11, 12 and 14, each include respective inserts having an arrow for illustrating the direction of travel of the object in question. This object is an aluminium racquet, a wooden racquet and a foot respectively. The point of impact of the racquet or foot, as the case may be, on the sensor means is illustrated as a respective imprint in a particular quadrant of the circuit board, these quadrants being designated 1A, IB, 2A and 2B respectively. Further signal processing and analysis in the time domain and/or the frequency domain can be used to extract significant features and discriminate ball impact against any other impact. It has been found that phase change analysis is particularly useful. Similarly another way of extracting information is to analyse the frequency spectrum. Since the signal is basically a pulse, it naturally contains a large number of frequencies. Some frequencies, apart from the large DC term, dominate and these plots are useful in discriminating between different objects. An auto correlation of the frequency spectrum is particularly useful in discriminating between objects.

Once the nature and location of the object causing the signal is identified the information can be conveniently presented to a monitor, usually viewed by the umpire, in visual and/or audible form. Such a monitor could also incorporate scoring facilities, player details and the like.

An additional problem is encountered in monitoring the central service line. In this case the critical edge of the line changes depending on whether the serve is being directed towards the forehand or backhand court. One way of addressing this problem is to position two sets of sensors one above the other with the sets being laterally offset such that they monitor different sides of the line dependant on which side of the line is critical for a particular serve. As the thickness of the sensor PCB is only about 2mm, the lower sensor is only minimally less sensitive than the upper sensor. The side of the service line which is being monitored can be switched depending on the court being served to and this can be done at any moment by a simple command entered, for example, by the umpire. Sensor accuracies are unaffected by switching of service lines. Although the invention has been described with reference to specific examples, it will be appreciated by those skilled in the art that the invention may be embodied in many other forms.

DESCRIPTION COMPONENT NAMES

Claims

CLAIMS : -

1. A method of identifying the type and impact location of an object striking a surface, said method comprising the steps of: providing at least one sensor means under said surface, said sensor means producing a signal in response to an object striking the surface in the vicinity of the sensor means; processing said signal to accentuate one or more parameters of said signal; and comparing said signal or parameters with predetermined signals or parameters whereby the type and impact location of the object is identified.

2. A method as defined in claim 1 wherein the object contains conductive material and the sensor means detects the proximity of the object.

3. A method as defined in claim 1 or 2, wherein the sensor means comprises a pair of sub-sensors located either side of a boundary line.

4. A method as defined in claim 3, wherein said pair of sub-sensors is provided by a transformer including a primary coil energized with alternating current and a pair of secondary coils coupled to said primary coil.

5. A method as defined in claim 4, wherein the coils are flat spirals which are located on a double-sided printed circuit board with the primary coil on one side and the secondary coils on the other.

6. A method as defined in claim 5, wherein the secondary coils are connected in series opposite and are fitted with a variable magnetic coupling.

7. A method as defined in claim 1, wherein the parameters include any one or more of amplitude change, phase change, or time related parameters.

8. A method as defined in claim 4, 5, 6 or 7, wherein said processing includes an amplifier-amplitude limiters connected to each of the secondary coils.

9. A method as defined in claim 8, wherein the two signals coming from the amplifier-amplitude limiters are then subtracted in a dynamic amplitude difference amplifier.

10. A method as defined in claim 9, wherein the output from the dynamic amplitude difference amplifier is passed to a phase sensitive rectifier.

11. A method as defined in claim 10, wherein the output from the phase sensitive rectifier is fed to a low-pass filter.

12. A method as defined in claim 11, wherein the filtered signal is digitised in an analog to digital converter.

13. A method as defined in claim 12, wherein the digitised signal is input to a computer for comparison of the signal with pre-determined signal characteristics to identify the type and location of the object which has struck the surface.

14. A^' method as defined in claim 12, wherein the digitised signal undergoes further processing prior to being input to a computer for comparison of the signal with pre-determined signal characteristics to identify the type and location of the object which has struck the surface.

15. A method as defined in any one of claims 1 to 6 or claims 8 to 13, wherein a homogenous base is provided under the sensor means.

16. A method as defined in claim 7, wherein a homogenous base is provided under the sensor means.

17. A method of detecting the type and proximity of an object containing conductive material, said method comprising the steps of: providing at least one sensor means, said sensor means producing a signal in response to the proximity of the object to the sensor means; processing said signal to accentuate one or more parameters of said signal; and comparing said signal or parameters with predetermined signals or parameters whereby the type and proximity of the object is identified.