This invention relates to golf-driving ranges.

According to the present invention there is provided a golf- driving range in which a target is set at a distance from a plurality of launch positions from which golf balls are driven individually to land on an impact surface of the target, and a catcher extends peripherally of a section of the impact surface for receiving entry therein of balls that after landing on the impact surface roll on the impact surface into the catcher, wherein the catcher comprises an energy absorber that is operative for dissipating kinetic energy of each ball entering the catcher to bring the ball to a halt before it passes to a collection-hub of the catcher, and the golf-driving range includes

instrumentation for determining data related to parameters of travel of each ball in at least part of its passage between its launch position and the collection-hub of the catcher.

The energy absorber of the golf-driving range of the invention may comprise flexible material weighed down to form a flap suspended within the catcher for deflection on impact by individual golf balls entering the catcher, and may be located above pathways onto which balls that have been brought to a halt by the energy absorber fall to travel by rolling into the collection-hub.

The instrumentation may include sensors for deriving data related to the entry of individual balls into the catcher and their entry into the collection-hub, and may include piezoelectric cables in the target or microphones for deriving data related to the time and location of landing of individual golf balls on the impact surface. Sensors for deriving data related to the times of launch and/or velocity vectors at launch of individual balls from the launch positions, may also be included. In these circumstances, at least some of the golf balls launched may be tagged for identification purposes, and the sensors for deriving data related to the times of launch and/or velocity vectors at launch of individual balls may also derive from tagged balls data related to their respective identifications. Data related to identification may also be derived from tagged balls received at the collection- hub of the catcher. The tag of tagged balls may be provided as an optical barcode or matrix code, but more especially may be provided as a radio frequency identification device (RFID) embedded inside the ball, and is commonly associated with a known launch position and/or a known player. The code of each ball used in a golf-driving range is normally unique, but non-unique tags such as 1-bit tags can be used for example to identify a ball hit by one specific player or hit from one launch position only. A RFID reader deployed at the collection-hub of a target reads the embedded code of each ball as it arrives at the reader and may transmit the code to a data analysis and user communication system of the range. This provides a means of identifying a ball that has arrived at the target and from this the position from which it was launched.

A second means of identifying the launch position is by 'shot identification' . This can be provided by monitoring the pattern of play of players who successfully land their shots on various targets of the range. This establishes who is playing which target and the likely timing of their next shot at a given target. When a tag fails during impact but the ball carrying that tag lands on a target and arrives at a RFID reader, the launch position of that ball can often be established from the history of previously struck balls. This process is improved if the time of impact of each ball at its unique launch position is recorded. It is then possible to match the times of arrival of balls at the relevant target with the times of recently-struck golf balls. Here, the only features of the golf shot that are used for identification are the time of impact at the launch position and the time of arrival at the target. Other launch parameters can be measured to enhance the shot-identification process, including the launch speed, the launch elevation angle and the launch direction in azimuth of each ball as it is hit from one of several launch positions. An extra benefit of recording launch impact times is that a player's user interface (such as a computer display) can be updated when a ball fails to land on a target shortly after a time-out confirms that the ball did not reach a target. This also provides asset tracking to ensure that balls do not go missing, especially when some of the balls are not tagged.

A further enhancement of the shot-identification process may be provided by measuring the positions and times of landing of balls on the impact surface and the positions and times of landed balls where and when they intercept a ball-stop mechanism (primarily the energy absorber) of the catcher at the perimeter of the target.

This, combined with measurement of the impact parameters of balls at the launch positions, provides strong correlation between launch position and landing behaviour and significantly improves overall identification reliability.

Several seconds can elapse between a ball landing on a target, rolling off the impact surface and eventually arriving at a RFID reader or other sensor in the catcher. One objective of the present invention is to provide fast and predictable transfer of balls from their landing positions on the target to an appropriate RFID reader or other sensor. In this respect the operation of the ball-stop mechanism (the energy absorber, the pathways from it to the collection-hub, and the collection-hub) of the catcher is crucial. Balls skid or roll off the edge of the landing surface at high speed so the ball-stop mechanism within the catcher must be sufficiently robust to survive such impacts and also must be designed to ensure that balls do not ricochet out of the ball- catcher. The main purpose of the ball-stop mechanisms is to remove nearly all the kinetic energy of balls as they enter a ball-catcher such that they are momentarily halted irrespective of their arrival speed. This ensures that the speed of passage through the catcher is independent of the speed of arrival and ensures that balls do not bounce around violently, which increases the time occupied in the catcher as well as risks bouncing out.

There is the possibility of an error in ball-identification occurring if only one RFID reader is deployed to decode all the balls that enter a ball-catcher. This occurs in the circumstances in which two balls with different accuracy scores arrive at the RFID reader almost simultaneously and the RFID reader has no means of deciding which ball was the more accurate. In this event, correct identification can be accomplished by arranging

predictable delays between the arrival times of different balls at the central hub where the time delays are dependent on the positions of entry into the ball-catcher of the different balls. One purpose of shot-identification is to increase system

reliability. Golf balls are subjected to exceptionally high shock when they are hit, with peak accelerations frequently exceeding 40,000 g. This reduces the useful service life of any embedded RFID tag, which is most likely to fail on impact, so reading the code of each RFID ball before it is struck does not prevent identification failure. In the event of any form of RFID- identif ication failure, the shot-identification process replaces RFID identification and the integrity of the overall

identification process is maintained.

The shot-identification process can allow a substantial reduction in the number of RFID readers that need to be deployed, and a significant reduction in the complexity of RFID-hardware

integration. This may be achieved by providing a small percentage of RFID balls such as not more than 10%, but preferably not more than 1% of all the balls used in the driving range, and using shot -identification as the default means of identification. The RFID tagged balls provide secure and accurate identification, whereas shot identification relies on the probability of correctly matching the launch and landing parameters of different shots, and this process is generally less reliable than RFID identification. Thus, the RFID-tagged balls are used as 'prize balls' where a valuable prize or the outcome of a competition is dependent on their unique identification. RFID-tagged balls are significantly more expensive to manufacture than standard-range balls, so limiting their number to a small fraction of the total ball stock reduces running cost of the driving range. Prize balls can be dispensed at a service desk where players pay extra to obtain a small quantity of prize balls or they can be provided free of charge as a loyalty incentive or the like. In these situations, only one standard RFID reader is required for ball dispensing purposes (such as an off-the-shelf hand-held reader) and the huge cost of integrating RFID readers in all the driving bays is avoided. The sorting of prize balls from the bulk of balls retrieved from the range outfield, may be carried out at low cost, preferably, by distinguishing prize balls by colour or visual distinctiveness from standard balls, or other attributes such as buoyancy. These distinctive qualities are useful for detecting and sorting the two different types of ball once they have been played onto the outfield and retrieved in random order, and can also be used to make it apparent to users which type of ball they are playing. Separating the two types of ball can also be achieved by detecting (but not necessarily decoding) the embedded RFID tag.

Players hit balls from playing bays towards targets positioned at various distances D metres downrange and the landing surface properties ensure that the motion of any ball that lands on a target is rapidly converted to a rolling motion (i.e. such that the ball's peripheral motion equals its translational motion). Thus, the majority of balls that land on a target impact-surface do not bounce off but instead roll off the edge of the target, where they are intercepted by a ball -catcher . The targets may be circular, oval, polygonal or irregularly shaped. Preferably, the width and length dimensions (or the circle diameter) of the targets should be at least 10% of D to give players a good chance of landing on the target. For short range targets, for example up to 100 metres or so, it is

preferable to increase the target size to be at least 20% of D as this increases the chance of achieving a 'hole-in-one' shot and gives players of lesser ability a chance to at least land on a target. Targets at 100 metres distance or more can all be a standard 25 metres diameter.

Targets may be provided within a short-range indoor facility, in which case the impact surface may be horizontal and players can hit balls towards the target from different directions. More usually, targets are provided on an outdoor driving range where players hit balls from one end of the range and the impact surfaces are inclined upwardly in the general direction of ball travel. This ensures that the target surfaces present an improved visual perspective and that rain water drains off into suitable soak-aways near the lower edges. In this configuration the aiming point is preferably centred on the distal (and thus uppermost) end of the target so at least one ball-catcher is provided having its inlet extending around the uppermost part of the periphery of the impact surface. Preferably, the gradient for the impact surface has a value greater than the coefficient of rolling friction of a standard golf ball on the impact surface. This ensures that balls do not accumulate on the impact surface as this is unsightly, prevents ball identification and interferes with other balls.

The trajectory of a ball as it rolls across the impact surface may be measured using Doppler radar, machine vision or other well- known technologies. The impact surface is preferably

substantially flat and smooth with essentially-uniform bounce properties. A flat impact surface enables measurement of roll trajectories to be carried out simply and reliably by measuring the time of landing and the position of landing of a ball on the impact surface, and the time of entry and the position of entry of the ball at the inlet of the ball-catcher. It is preferable that the impact surface is not significantly convex or concave; a convex shape such as a spherical or conical dome will cause a fast-rolling ball to fly off the surface before being collected in the ball-catcher, whereas a concave spherical or conical dish can trap landed balls spiralling round the target for a long time before eventually reaching the bottom of the dish. Convex or concave geometries moreover result in three-dimensional roll tra ectories, which require considerably more measurement data and computation to determine initial roll speed and direction.

Immediately after landing on a target, a ball may bounce slightly and thereafter skid to a slower rolling speed. The bounce and skid contact removes the ball's backspin, which is always present in a lofted golf shot, and further skidding continues until the ball eventually starts to roll. It is well established that the ratio of initial roll speed to initial skidding speed (on

horizontal firm surfaces) is determined almost solely by the spin of the golf ball as it first starts to skid. High backspin creates more skid and greater speed reduction compared with low backspin. The speed-reduction ratio is independent of the quality of the skid surface. A rough surface will create high

deceleration for a short period and vice versa for a smooth surface.

With the golf-driving range of the present invention, impact surface of the target is preferably only slightly inclined so that it does not significantly alter the direction of skid and roll from the azimuth direction of the ball's trajectory just prior to landing. Moreover, the gradient and rolling friction of the landing surface are desirably substantially constant and known so the slight deceleration due to these factors can be determined. Knowledge of a ball's roll speed and roll deceleration (due to gradient and friction) thus provides good estimates of the horizontal velocity and backspin of a ball at the instant that it lands on a target and starts to bounce and skid. This in turn provides a method of estimating the ball's flight duration. The direction of the initial skid and roll trajectory along the target impact surface is virtually identical to the ball-trajectory azimuth-angle as it touches down on the impact surface of the target. A useful estimate of the initial roll speed can be found from the average speed of a ball measured from the time of landing on the impact surface to the time of rolling into the ball-catcher and from a measure of the distance travelled during this time interval .

An advantage of measuring a bail's landing parameters on a target is that it allows reliable prediction of the likely bounce and run of the ball on other representative surfaces such as a golf course fairway or green. For example, a ball that carries a given distance with long flight duration will have a higher trajectory and less forward velocity compared to a ball that carries the same distance in a shorter time. The ball with the shorter flight duration will bounce and roll on a fairway much further than the other ball and this information is important to golfers who seek to improve their drive length. On the other hand, a high

trajectory with high backspin is desirable for approach shots where golfers want their shot to 'hold' on a green and travel only a short distance beyond the landing spot. Yet another advantage of measuring landing parameters is the ability to measure and record the draw or fade (flight deviation caused by sidespin) and estimate the imparted backspin of shots. This facility is of interest to advanced players and is normally a feature that is only possible with expensive radar tracking or high technology launch analyser systems.

A golf-driving range in accordance with the present invention will now be described, by way of example, with reference to the accompanying drawings, in which:

Figure 1 is a schematic plan view of the golf-driving range according to the invention;

Figure 2 is a sectional side-view of a short-range target of the golf -driving range of Figure 1, the section being taken on the line A-A of Figure 1 ; Figure 3 is a schematic view from the rear of the inside assembly of a ball-catcher of the target of Figure 2;

Figure 4 is a part-sectional view taken on the line B-B of the ball-catcher of Figure 3;

Figure 5 is a detailed view of a ball-funnel assembly that forms part of the ball -catcher of Figures 3 and 4 ;

Figure 6 is a schematic plan view of a long-range target of the golf-driving range of Figures 1 and 2 ;

Figure 7 is a sectional side-view taken on the line C-C of Figure 6; Figures 8(a) and 8(b) are, respectively, representational side and plan views illustrative of a form of ball-catcher that may be used as an alternative to that of Figures 3 and 4 in the long-range target of Figures 6 and 7. Figure 9 is a front view of a second embodiment of a long-range target of the golf-driving range of Figures 1 and 2 ;

Figure 10 is a sectional side-view of the long-range target of Figure 3 , the section being taken on the line D-D of Figure 9; and

Figure 11 is a detailed view of the inside rear-view of a ball- catcher of the long-range target of Figure 9. The golf-driving range to be described includes a data analysis and user communication system (not shown) that typically comprises a central computing system, data links to sensors and to user interface devices such as touch screen displays, audio annunciator devices, mobile phones, on-line printer and the like. A possible data link between various distributed parts of the system may be by means of the Internet, with components such as the targets, the tee-off bays, and user interfaces of the range each assigned a unique Internet Protocol address. Referring now to Figures 1 and 2, the golf-driving range includes a short range trapezoidal target 1 that is positioned at range D from launch positions at the tee-off or playing bays 2. The target 1 lies on a substantially flat base 3 and is one of several targets disposed at different distances on the outfield of a driving range. Typical dimensions for this short range target are; 3.7 metres rear width (12 feet), 7.4 metres front width (24 feet) and 5.6 metres front to back length (18 feet) . The base 3 for this shape and size of target is readily constructed using standard sizes of panel material with supporting timbers, decking materials or the like. Different shapes and sizes can be adopted; for example, very small targets with surface area of about 1 square metre or so can be positioned at various distances

downrange and used as special prize-winning targets (owing to their small size, an element of luck as well as skill is required to land a ball on such targets) . All the balls used in the range-facility may be RFID-enabled but to reduce costs it can be preferable to provide only a small number of tagged 'prize balls' . These may be dispensed from a service desk and so avoid the need to install RFID readers at each tee-off bay 2. An essential requirement is that the data analysis and user communication system 'knows' the player and/or the launch position of each ball. Although it is possible to allocate a known set of RFID balls to a particular player who can then choose to play at any available driving bay 2, it is preferable to ensure that the data analysis and user communication system always knows which set of RFID balls is being played from a given driving bay; the order in which the balls from a set are played is not

important. This allows the data analysis and user communication system to implement shot identification and provide detailed shot analysis that is of benefit for training and practice. A quantity of RFID balls can be dispensed into a basket or suitable

receptacle and the basket/receptacle itself can be electronically tracked .

Each playing bay 2 is provided with a driving mat 4. Players hit shots off teeing devices or directly off a mat in the playing bays and aim to play their shot such that the ball rolls near the target-aiming flag 5. The aiming flag 5 is positioned at the centre of a ball-catcher 6, which collects balls that roll up to the end of the target. Sensors within the ball-catcher 6 measure the position of ball-entry points relative to the aiming flag 5 and also the time of entry. One RFID reader is deployed within the ball-catcher 6, which reads the identifying codes of balls as they arrive in succession on the target 1, and sends the acquired codes to a central computer (not shown) of the data analysis and user communication system. The central computer provides data links to all the peripheral devices such as sensors and interfaces fitted on the target 1 and user interfaces. The user interfaces may be mobile devices or fixed units such as audio annunciator devices or touch screen displays. Data from each target on the outfield is analysed and the current activity of balls landing on the targets is shown on a graphical display. In Figure 1, a ball 7 is hit from tee-off position 8, lands on the target 1 and rolls across the landing surface 9 into the ball- catcher 6 several feet to the left of the flag 5. Since in this case range D is only twenty five metres or so, the azimuth flight trajectory 10 is substantially straight even if sidespin is imparted to the ball 7.

Piezoelectric cables 11, arranged in a grid configuration and attached to the target base, provide means of detecting the time and position of landing by measuring the relative times and amplitudes of acoustic vibration caused by landing impact. The times and positions of a ball first landing on the target 1 and then entering the ball-catcher 6 are used to measure the direction and the average speed of the bounce and roll trajectory 12. The direction of the roll trajectory 12 is closely aligned with the final azimuth direction of the ball flight trajectory 10. The very small change in azimuth direction resulting from the landing surface gradient can be calculated and applied as a correction to find the true direction of the ball as it approaches the target. In the illustration of Figure 1, the aforementioned change in azimuth direction is known in golf as 'a break to the left' .

Limiting the gradient of the landing surface and providing very low rolling friction ensure that left or right breaks in direction are small so corrections are small and accuracy assured. A reconstruction of the ball speed throughout the roll trajectory 12 can be derived from measurements of the start and end positions of the roll trajectory 12 and the time taken to travel this path, combined with knowledge of the gradient and the coefficient of rolling friction of the landing surface 9. Thus, the arrangement of Figures 1 and 2 provides measurements that can be used to estimate the flight duration of balls that land on the target and also the direction of the final flight trajectory. This allows the data analysis and user communication system to match landing balls with launched balls on a probability basis so as identify the likely launch position of successful shots.

Shortly after the ball 7 is hit from tee-off position 8, a second ball 13 is hit from tee-off position 14 and enters the ball- catcher 6 at a position close to the flag 5 and to the right of the flag 5 with roll trajectory 15. In an illustrative example, ball 7 is a prize ball containing a RFID tag but its entry position is too far from the flag 5 to win a prize, while ball 13 is a standard ball that lands close to the flag 5 in a prize winning position. Normally it is possible for the identity of the launch positions to be determined correctly just by knowing the unique RFID codes of all the balls hit from tee-off position 8, but this is only the case if the two ball landing times are sufficiently separated. However, if the balls arrive at the RFID reader at nearly the same time, it is not possible to determine by RFID identification which of the two balls landed to the right of the flag 5 in a prize winning position and which landed too far to the left to win a prize. Under these circumstances, shot

identification analysis allows the data analysis and user

communication system to determine that ball 7 landed to the left and ball 13 to the right.

A first estimate of shot identification is derived from the directions of the two roll trajectories 12 and 15. Roll

trajectory 12 points in the direction of tee-off position 8, whereas roll trajectory 15 points in the direction of tee-off position 14 and these differences provide a first means of identification. Shot-direction identification is not dependent on any launch parameter measurements at the tee-off positions 2, but it is preferred to make measurements of at least the time of impacts at the different tee-off positions 2 to improve shot- identification accuracy. In the illustration of Figure 1, the times taken for ball 7 to travel from its landing spot to the RFID reader inside the ball-catcher 6 is considerably longer than the corresponding time of ball 13. Thus, if both balls arrive at the RFID reader at nearly the same time, the launch time of ball 7 is significantly earlier than that of ball 13 and measurement of the appropriate launch times are used to identify reliably the correct landing positions of each ball. Preferably, the travel duration of balls through the ball-catcher 6 is accurately predictable as this improves the discrimination of shot identification that uses timing information as described above.

To reduce costs, 1-bit tagged balls can be used instead of uniquely coded RFID balls. For example, a simple magnetic material insert such as a small magnet or other magnetic material can be used. A solid ferrite core or ferrite slurry filler can be sealed inside a golf ball after being inserted via a small borehole. The ferrite is electronically detectible by standard search-coil techniques. When several such 1-bit tagged balls are used, it is preferable to print additional identification symbols or numbers as a supplementary means of identification in parallel with shot identification. Typically, the 1-bit tagged balls are used as prize balls and comprise a very small percentage of the total ball stock available for play. They are randomly mixed in with the bulk of standard un-tagged balls and are preferably distinctively coloured as well as printed with a unique number or the like. For example, 100 prize balls may be used, randomly mixed and un-sorted from the bulk stock, and each printed with a two digit unique identifier between 00 _. and 9_9. When a prize ball turns up, the player records the number printed on the ball either by memory, by pen and paper or more preferably by electronic recording means linked to the data analysis and user communication system; the ball is then hit towards a target. The frequency of prize winning chances is very low and the chances of successfully achieving a prize winning shot is also low so a successful prize winning shot is normally identifiable by the shot identification process. In this respect it is advantageous to know the time and launch position of prize balls when they are hit as this gives very secure shot identification. When two prize balls become available almost simultaneously and contention about the shot results occurs, then the true outcome can be found by checking the recorded unique numbers .

To assist the shot identification process, the order of play amongst numerous players can be regulated by introducing subtle changes in ball delivery rate and display information tempo and, where possible, directing players in neighbouring tee-off bays to aim at different targets. This can be achieved by creating game scenarios where players progress from one target to another but in differing order, etc. Ball delivery times and the tempo of user displays can both be under computer control. Trends in playing patterns at different bays are analysed and used to optimise the timings of result displays and ball delivery. In this respect, automatic teeing mechanisms 16 disposed underneath the driving mats 4 provide the best method of ball delivery. The aim is for the play regulation process to be covert and not noticeably interfering. With reliable landing data and launch timing data available, small differences in impact times will provide

consistently correct identification.

It is important that the impact surface 9 has consistent roll and bounce characteristics. Preferably, the coefficient of rolling friction of a golf ball on the landing surface 9 is less than 0.1 but more preferably is less than 0.05. In order to ensure that the rolling friction is low and does not vary significantly in wet or freezing conditions it is preferable that the landing surface is covered with a water- impermeable membrane with suitable sealing at joints and edges. The cover membrane ensures that impact absorbing material on the landing surface is kept dry so that its rebound characteristics do not significantly vary with weather conditions. Suitable impact-absorbing materials include low- rebound foam or soft sand or the like. It is desirable to ensure that golf balls landing on the target have a low probability of bouncing off the target and that instead they roll at speed into a ball-catcher 6 on the periphery of the target. In this respect, it is preferable that the rebound coefficient of the landing surface for short range targets is less than 0.24, measured as the ratio of a golf ball's rebound velocity to its impact velocity normal to the landing surface. In a short-range target, the vertical descent velocity of a steeply descending ball is rarely more than 10 metres per second and is much less for low

trajectories. The bounce height resulting from an impact speed of 10 metres per second and a rebound coefficient of 0.24 is 29.2 centimetres or just less than 12 inches (305 millimetres) . This provides a practical limit for the entry height of a ball-catcher. Thus, any golf ball landing on a target with a descent velocity of up to 10 metres per second will always enter a ball-catcher

(rather than bounce over), provided the ball-catcher is able to catch balls that bounce in at a height of up to 29.2 centimetres.

Referring now to Figures 3, and 4, the main structure of the ball- catcher assembly 6 in the target of Figures 1 and 2 comprises a right-hand gutter 30, a left-hand gutter 31, a top cover 32, a rear cover 33, an energy-absorbing stop-flap 34, a funnel 35, a ball collector 36 and a 'hole-in-one' receptacle 37 (the right- and left-handedness of gutters 30 and 31 are reversed in the back view of Figure 3) . Balls roll off the impact surface 9 and enter the ball-catcher 6. The majority of balls roll at sufficient speed to jump over the gap between the edge of the landing surface 9 and the flap 34. The flap 34 dissipates essentially all the kinetic energy of balls that collide with it, so the balls are brought momentarily to a halt and drop almost vertically down onto one or other of the two inclined flat-bottomed gutters 30 and 31. The floor of both gutters are lined with low rolling friction, impact absorbing material 38 to limit rebound and ensure that balls start rolling very shortly after dropping inside the gutter. Any low-velocity ball that just rolls off the edge of the impact surface 9 also drops into one of the two gutters. The flap 34 depicted in Figure 4 comprises a fold of tarpaulin running along the length of the ball -catcher 6, suspended from the top cover 32 and weighed down within the bottom inside of the fold with metal chain 39. The flap 34 is flexible and heavy, and deflects slightly on impact with a golf ball but only slowly restores to its original position and thus halts the ball with minimal rebound. Ideally, the flap 34 is designed to stop the ball with negligible rebound and sideways slipping such that it comes to rest momentarily and at a consistent height from the edge of the landing surface, irrespective of its speed and direction of entry into the ball-catcher 6. To ensure that lateral slipping is suppressed, a bristle surface or other suitable mechanical arrangement can be provided on the flap 34 to engage and hold the ball momentarily. Other ball-stop means such as cushioning or other low rebound material can be used.

An illustrative path of a ball 20 as it leaves the landing surface 9 at position 20a and eventually drops into the funnel 35 at 20d is shown by arrowheads and dotted lines/curves 40. After the ball rolls off the landing surface 9 it hits the flap 34 at position 20b, drops vertically and lands on the gutter 30 at position 20c and then rolls down and enters the funnel 35 at 20d. The drop distance from the ball height at position 20b to its height at position 20c is linearly proportional to MD and the length of the ball roll path down the gutter is also linearly proportional to MD. MD is the 'miss distance' distance of the ball from the aiming flag 5 and is a measure of the shortest distance from a ball's roll trajectory to the aiming flag 5. As the ball drops and then rolls it accelerates and the resultant increases in drop and roll velocities are precisely related to the drop and roll magnitudes. Thus, the value of MD can be extracted from any one of several possible measurements involving elapsed time or velocity, including at least one of: the drop duration of a ball as it drops from the ball-catcher entry point to a gutter, the velocity of a ball as it drops onto a gutter, the roll duration of a ball as it rolls down a gutter, and the exit velocity of a ball as it rolls off the bottom end of a gutter. For each gutter 30 and 31, an array of light emitting devices (LEDs) and an array of co-acting detectors are arranged between the lower part of the hole-in-one receptacle 37 and a sensor assembly 41. Light- interrupt detection beams 42, 43, and 44 are generated by an upper and lower line array of LEDs at one end, and matching arrays of light detectors at the other end. The two LED arrays in a pair are operated in fast multiplex mode, with one array switched off while the other is switched on. The beam-forming devices (not shown) are arranged to ensure that balls are detected accurately over a detection space that allows for variations in drop position across the width of the gutter. Beams 42 and 44 are parallel and thus can be used to measure ball-drop velocities just before entry into a gutter. Cross beams 43 are available by default and can provide a crude but direct measure of the ball position along the length of the gutter. This arrangement provides good rejection of spurious signals caused by raindrops, insects and windborne debris, etc. Optionally, additional sensor can be installed to detect the time that a ball impacts the gutters 30, 31 and the funnel 35.

The length and gradient of the gutter is chosen to ensure that the time delay td between a ball passing off the edge of the impact surface and its time of arrival at the said collection-hub is within a desired maximum response time. Preferably, td should be less than three seconds to ensure that players see the result of their shots in a timely manner. The rigid plastic gutter lined with low friction impact absorbing material 38 ensures that td is consistent and predictable. This is necessary to ensure that the data analysis and user communication system knows the order and timing of balls being read by the target RFID reader. Otherwise, balls arriving at nearly the same time, but with significantly different MD values could be confused. Measurements of MD and td for all balls that enter a ball-catcher can be accumulated and used to update an algorithm or look-up table for predicting the times of arrival of balls at the RFID reader. Random timing errors will occur and care is required to minimise randomness. Preferably, the errors in predicted td should have a one-sigma distribution of not more than 0.2 second. As well as ensuring that balls with different MD values are correctly identified, the process of analysing the MD and td data can show up trends and diagnose problems in the target mechanisms.

Sometimes a RFID ball may drop into the funnel 35 from gutter 30 at almost the same time that a second RFID ball drops in from gutter 31 so an 'identity collision' occurs. These balls will have different MD errors right and left of the flag 5 but a 50% chance of the correct MD value being attributed to the correct RFID ball. In this event, the correct identity of the balls can be found by shot identification. Measurements of the times of launch and launch velocity parameters are correlated with the landing parameters as measured by the various sensors installed in the target. In the very rare event that a reliable identification cannot be achieved then the data analysis and user communication system can award both shots the higher of the two scores.

However, a hole-in-one shot is an exception and preferably absolute means are provided to ensure that a hole-in-one shot is correctly attributed to the correct RFID ball. The hole-in-one receptacle 37 has an opening diameter equal to a standard golf hole (4.25 inches) so balls in line with the flag drop into the receptacle 37 rather than into one of the gutters. The ball is held inside the receptacle by gate 45 until a local processor predicts that no other balls are going to drop into the funnel 35. Referring now to Figure 5, a golf ball 50 containing an embedded RFID tag 51 is shown in three positions 50a _; 50b, and 50c as it travels through the ball funnel 35 of Figures 3 and 4. The ball funnel 35 comprises a gutter hopper box 52, a 92.5 degree gutter bend 53 (round pipe version) , a short length of 65 millimetre square gutter downpipe 54 and attached sensors. The round section gutter bend 53 is adapted slightly so that the attached square downpipe 54 can rotate about the pipe axes. The various gutter items 52, 53, 54 and also 30, 31 are standard rigid plastic building parts, but a custom built assembly can be used if preferred .

After rolling off the gutter 30 the ball 50 flies onto the sidewall of the funnel (position 50a) 35 at speeds of up to 3 metres per second for typical ball-catcher designs. Impact absorbing lining 55 prevents excessive bounce so the ball quickly rolls into the gutter bend 53 (position 50b) and from there it bounces and rolls through the gutter pipe 54 and out the open end (position 50c) . The gutter pipe 54 tilts downwards by nominally 2.5 degrees to allow any rainwater or hosed wash-water to drain away. Speed limiting baffles (not shown) may be provided to slow down the ball and ensure that balls are separated as they roll through the gutter pipe 54. A RFID reader antenna 56 is attached to a lower side of the square gutter pipe 54. The antenna sensitive axis is thus at 45 degrees to the ball-roll axis.

Analysis of the ball rotation shows that the sensitive axis of the RFID tag 51 will always be within 60 degrees to the antenna axis at some point through any quarter-turn excursion. A golf ball quarter-turn roll down the 90 degree V-groove formed by the square pipe corresponds to a 24 millimetre roll distance. Provided that the antenna 56 has operating sensitivity over 24 millimetres down the pipe axis, the reader-to-tag coupling will be at least -6dB relative to maximum coupling, where maximum coupling corresponds to zero degrees alignment. Referring now to Figures 6 and 7, a long range oval target 60 has its minor axis aligned along the general direction of ball flight. A landing surface 61 has three ball-catchers 62, 63 and 64 positioned at the rear perimeter proximate to an aiming flag placed at a target hole 65, with ball-catcher 63 positioned at the centre rear and closest to the target hole. A buried conduit 86 allows balls that drop into the target hole 65 to roll into the ball-catcher 83 where its time of entry and RFID identity code (if available) is read. In an exemplary and non-limiting

configuration, the width of landing surface 61 (along its major axis) is 23 metres and the depth (along the minor axis) is 15 metres .

For long-range targets such as that depicted in Figures 6 and 7, the vertical descent velocity of a landing ball can reach 24 metres per second or more and for such targets it is preferable to have rebound coefficient of not more than 0.1 in order to limit bounce height to below 12 inches (305 millimetres) for the great majority of golf shots. Rebound coefficient can be non-linear and increase significantly at high impact energies due to 'bottoming' when the top layer of a landing surface compresses and a harder substrate provides most of the reactive rebound force. It is thus important that the stated rebound coefficients are valid at the appropriate impact velocities. Balls are hit from driving bays (not shown) at distances of typically 50 to 250 metres to the left of the target as depicted in Figures 6 and 7. Most of the balls that land on the landing surface bounce very slightly and then roll at speed towards one or other of the three ball-catchers where they are collected and their entry times and positions at the ball-catcher inlets are measured and recorded by sensor arrangements within the ball-catcher assemblies. The landing surface is slightly inclined upwards from front to rear so that balls that bounce weakly onto the surface roll back towards the front edge or towards the sides. A raised lip 67 is provided along the extent of the perimeter edge excluding the edge occupied by the ball-catchers. The purpose of the raised lip 67 is to deflect balls that land on the target but do not roll into a ball- catcher. The deflected balls roll down the incline towards the front edge of the target and are collected in a lower retrieval sump 68, which keeps collected balls out of view from the driving bays. Gutters 69 are provided at the lowest part of the front perimeter to ensure that balls roll into the sump 68 and do not slow to a stop against the inside of the raised lip 67.

Optionally the sump 68 can be provided with a RFID reader

arrangement to identify RFID balls that land on the target but do not achieve entry into one of the ball-catchers. The raised lip 67 and sump 68 are useful without RFID capability as this

arrangement minimises ball clutter in the vicinity of the target. In one exemplary construction, each of the ball-catchers 62, 63 and 64 is about eight metres long and gutters used within each ball-catcher are made up from standard 4-metre lengths. The ball- catchers are similar to those described earlier with reference to Figures 1 to 5 , but are much longer and the two halves of the assembly are rotated about a central vertical axis to follow the curved side of the oval target. In order to conform to the curved edge the rigid gutters can be replaced with curved ramps

constructed from semi-rigid plastic net material or the like. Two exemplary ball shots are illustrated in Figure 6. Ball 70 has flight trajectory 71 that carries to the right of centre on the landing surface 61 and then rolls towards ball-catcher 62 along roll trajectory 72. The landing time and position can be measured by analysing the time of arrival of the landing impact sound by microphones 73 that are housed in recesses inside the raised lip 67 and the ball-catchers 62 to 64. The ball's time and position of entry into ball-catcher 62 are also measured and metal sensors detect the presence or absence of a RFID tag within the ball 70, and a reader decodes any readable RFID device embedded therein. These data are transmitted via cable or wireless means to the data analysis and user communication system.

Ball 74 illustrates a shot that falls slightly short of the target 60 and then bounces onto the lower left corner of the landing surface 61. The bounce trajectory 95 and subsequent roll

trajectory 76 are characteristic of a much slower ball speed compared with the corresponding trajectories 71 and 72 of ball 70. The ball 74 rolls slowly up the landing surface 61 and then turns back, rolls into raised lip 67 and eventually rolls down into sump 68 via gutter 69.

Metal detectors are required only if a mixture of RFID balls and standard balls is being used. If all the balls in the facility are RFID balls and the entry position sensors detect a ball but the RFID reader fails to decode the tag in that ball, then that 'ball under test' is highly likely to contain a failed tag. A less likely event is that the reader itself has failed. Two successive RFID-read failures are indicative of reader failure and the data analysis and user communication system reports that maintenance or replacement is required on the faulty reader. In either event, it is important to still identify the player of the 'ball under test' so that he or she is not deprived of a score. The launch position of the 'ball under test' (which identifies the player) is first determined by matching an estimate of its time of impact with the times of impact of recently hit RFID balls. If only one shot is a close match to the estimated time of impact then that shot's launch position is known to be that of the 'ball under test' . If two or more candidate shots provide a match, then other tests can be applied. These test use measurements of ball launch parameters at each launch position, including launch speed, launch angle and azimuth direction. The launch conditions of different candidate shots are used to eliminate those shots that are not consistent with the measured carry-distance and landing- direction of the 'ball under test' . Figures 8a and 8b show details of part of the inside of a ball- catcher of Figures 6 and 7. A ball 80 travels at speed along roll trajectory 81 and impacts a buffer plate 82 positioned to

intercept balls that fly off the landing surface in a plane parallel and slightly above the impact surface 61. The buffer plate 82 absorbs substantially all the ball's kinetic energy such that the ball drops onto a slightly sloping base plate 83 and then rolls down trajectory 84, through light interrupt beams 85 that are emitted and received by electro-optical devices 86 mounted on divider fins 87. The divider fins 87 are positioned along the base plate 83 and sub-divide the entry portion of the ball-catcher into two or more sections to improve measurement accuracy and allow the ball-catcher assembly to closely follow the curvature of the target perimeter. The divider fins 87 are designed to allow balls to roll down the base plate with minimal deflection in the event that a ball hits the fin on entry. The slope of the upper surface ensures that the ball rolls steadily with moderate acceleration and no skid as it passes through the light-interrupt beams. The position of the ball roll trajectory along the length of the section and its time of entry are readily found from analysis of the light interrupt data. After rolling down the base plate 83, the ball drop into a gutter (not shown) which takes the ball quickly into a collection-hub as described in relation to Figures 3 to 5. The arrangement is such that balls start their downward roll trajectory 84 at low and consistent speed

irrespective of the speed and direction of their entry velocity along trajectory 81 and drop off the plate 83 with a known constant velocity. The time of arrival at the collection-hub relative to the time of entry into the ball-catcher is dependent on the position of ball entry as measured by the light interrupt beams 85. Thus, when two or more golf balls enter a ball-catcher at almost the same time, the times of arrival at the collection- hub allows the data analysis and user communication system to correctly identify the entry position of balls from a choice of possible positions as their RFID codes are acquired at the collection-hub .

Referring now to Figures 9, 10 and 11, a long-range target 90 comprises a flat landing surface 91, supported on a framework 92 assembled above ground level 93, a central ball-catcher 94 with four additional bal 1 -catchers disposed symmet ically about the target centre and a central aiming flag 95 attached to the upper part of the central ball-catcher. The impact surface is

relatively short and steeply inclined compared to target 60 of

Figures 6 and 7 but is of similar width. Typically, but without limitation, the landing surface 91 is 23 metres wide by 3.7 metres deep and thus is less than one third of the area of the impact surface of target 60. The design of target 90 is simple and suitable for modular construction. Each catcher assembly can be provided with rear-hinged panels that open back to allow access for maintenance and inspection; Figure 11 shows the central ball- catcher with rear panels swung back to reveal the inside assembly. The target instrumentation comprises a RFID reader system (not shown) with six antennae connected to a multiplexing reader unit and a data link to the data analysis and user communication system. One antenna is provided in a dedicated 'hole-in-one' detector assembly 96. Balls that roll into the detector assembly 96 are read by the RFID reader system and registered as a hole-in- one event. The time of arrival and identity code of such balls are transmitted to a data analysis and user communication system. The remaining five antennae are provided in the collection-hubs 97 positioned at the lower ends of ramps 98. Two such ramps 98 in each ball-catcher provide a conduit arrangement that transports balls from a ball-stop device 99 to the collection-hub 97.

By default, the RFID reader system measures the times of arrival of balls at each collection-hub 97. The ramps 98 are, in this exemplary configuration, 2.3 metres long and inclined at about 20 degrees to the vertical . A low-rebound and low friction cladding material on the ramps helps the balls to roll quickly with minimal bounce when they drop down from the energy-absorber ball-stop device 99. A ball that starts to roll from the top of a ramp will take slightly less than 1.5 seconds to roll from top to bottom. By contrast, a ball that drops down closer to the centre just missing a central deflecting member 100 will take over 0.5 second to roll down the remaining short length of ramp. Thus, a measure of the time of arrival of a ball at a RFID reader provides a measure of the time of arrival of the ball at the enery-absorber ball-stop device 99 to within one second.

Figure 10 shows two exemplary balls 101 and 102 landing on the landing surface 91. Ball 101 lands directly on the target with flight trajectory 103 and roll trajectory 104, whereas ball 102 bounces slightly short of the target with flight and bounce trajectories 105 and roll trajectory 106. The target impact- surface is sufficiently deep and inclined to make it reasonably easy for players to land their balls close to or on the target such that their shots end up entering one or other of the five ball-catchers. However, balls that fall a few metres short of the target will not have sufficient kinetic energy to enter a ball-catcher. For any given target distance, there is a confined range of launch speed and launch angle combinations that can reach the target and enter a ball-catcher. Shot identification can use time of launch and launch velocity parameters measured at the launch positions and match these with the time of arrival at the energy-absorber ball-stop device 99 to identify balls that fail to be decoded by the target RFID reader system. The probability of RFID tag failure in the landing golf balls is small but not negligible and is most likely to occur in balls that are driven long distances towards targets like long-range target 90. It is thus very beneficial to have shot=identif ication for recovering a ball identity even if the process is limited at the target to measuring time of arrival at the energy-absorber ball-stop device 99. If desired, additional sensing can be implemented. For example, simple vibration sensors attached to the ramps 98 and central deflecting members 100 can improve the accuracy of the time of arrival measurements. In particular, a vibration sensor 107 attached to the central deflecting member of the central ball- catcher 94 provides an improved degree of game-scoring accuracy by measuring balls that land near the aiming flag 95.