A challenge for you expert folks.....about tennis!!

[usermar]

Hi everyone, one of mine best friend acquired some years ago the product patent to build Plexiglas panels to use like a tennis wall. A normal tennis wall used to be a simple wall were the net line is painted and where tennis players may exercise by hitting the ball towards the wall and on and on to learn, train, warm up, etc.

What my friend what to do now is to add electronic score when particular areas of the wall will be hit, requiring therefore to "digitize" the wall, catch the signal coming from the sensors and elaborate on them to show up scores and make the tennis wall more like a game and certainly less boring particularly for children that wants to learn tennis play and get some fun out of it.

Now the point is.....what do we need to do to get this done? Which technologies we should use? Who can be able to design such a solution and integrate into plexiglas walls?

Interesting and challenging.....isn't it?

I'm excited to learn folks......thanks!!!

 

[Electrobrains]

Ok, I'm first out with some ideas:

1. Install a matrix of LDR's on the back side of the (transparent) Plexiglas. Each time the ball hits the board it will produce a short shadow spot on the matrix, which can be high pass or band pass filtered and detected as a ball hit. The information of course would have to be analyzed by a micro controller etc.

2. This might be more difficult, but would allow the Plexiglas to be pigmented: Install a matrix or pattern on the back side of the Plexiglas, made out of metal foil tape. The foil areas would have to act as capacitive proximity sensors (probably in an RC oscillator circuit with frequency detection through for instance PLL technology). This would be a more difficult approach, but the more nifty. The condition would be that the ball hit would be able to induce a capacitive or electro-statical change at the spot. If it wouldn't work, a metal film on the front side with a fixed potential could work. Through the hit a slight bending or squeezing together of the Plexiglas could be used for electric detection. The signal would sure need amplification and filtering etc.

 

[usermar]

Very interesting Electrobrains and thank you! I'm not an expert at all, and I apologize if I may ask stupid questions. Will both solution 1 and 2 give the possibility to identify in which area of the wall the ball hit so to give different scores for different hits? And second, question, will both solution be realiable and robust enough to physically resist to thousands of hit and stay outside in the rain, low and high temperatures and whatever weather conditions?

Appreciated!

 

[CocaCola]

For reliable detection, I would likely explore the same tech used in resistive touch screens... A laminated resistive touch screen on the plexiglass will allow for precise and reliable impact detection, as long as it can be designed to withstand the repeated force of impacts...

 

[usermar]

Thanks CocaCola! One important aspect to industrialize the solution will of course be to find the best price/performance ratio, since the idea is to sell/lease them to tennis club and has to be a reasonable price as well..........

 

[Electrobrains]

Very interesting Electrobrains and thank you! I'm not an expert at all, and I apologize if I may ask stupid questions. Will both solution 1 and 2 give the possibility to identify in which area of the wall the ball hit so to give different scores for different hits? And second, question, will both solution be realiable and robust enough to physically resist to thousands of hit and stay outside in the rain, low and high temperatures and whatever weather conditions?
Appreciated!

Both ideas mentioned would allow for detection of the desired areas where the ball might hit.
With the first idea, It would simply be at the places where the LDR's (Light Dependent Resistors) are placed and in the second case, it would be the places that are covered by the foil.

The "accuracy" of such wall, would depend on how the sensors are placed, how many they are etc. In the second case, it would matter how many stripes, patches or fields of the foil are used and where they are placed. Each differentiation of a separate area on the wall, would need an isolated piece of foil for that place and a separate electronic input to "read" that area.

From these two ideas, I would definitely prefer the second one - if it would work well physically. It would be an easy, flexible and very cheap solution for the wall itself. You could for instance use prefabricated adhesive copper tape (available from for instance 3M, intended for shielding). The electronic side would be more advanced and difficult to develop than the first idea.

The question about reliability and robustness is very important!
In this stage it's too early to say anything about the reliability (and even if it's possible to make it according to these ideas).

Basically, you can make almost anything with enough investment of resources.
Concerning robustness, I think the second idea would be in advantage again. It would be necessary to cover the metal foil with something like plastic foil or epoxy resin.
The connectors and enclosure for the electronic parts would also demand special attention.
According to my feeling, the road is quite long till the details of such project would need to be discussed.

The idea from Coca-Cola is also interesting. If an already established technique could be used, it would be a good help for the development.
My feeling concerning that idea is that the electronic part might get cheaper, but the wall more expensive.

 

[CocaCola]

Another option is to engineer a ball to do the impact sensing, with the location tracked by cameras... Much like they did for a few years when Fox televised hockey games and the puck had a glowing ring around it, or like they use in the current 'Hawk Eye' tracking system for sports...

This is a more costly systems but they might provide a better long term solution in a permanent installation...

And last but not least we zoom back to earlier (and still used) touch screen tech, where there was a vertical and horizontal grid of IR/visible or lasers mounted along the perimeter of the screen just above the surface of the screen... When the ball impacts it will break these beams in that location and you will have the x and y coordinates... This option is nice as the electronics are removed from impact zones, and also allows for positioning dynamics... Say you wanted to project a moving box for the players to target, it would be simple enough to pair the impact and box location in real time with a basic computer system over the entire area... This type of touch screen is still primarily used in high traffic unmonitored kiosk and the sorts as it requires little maintenance to work properly, and even if the screen (or touch area) is dirty (or even opaque) it still works fine

 

[(*steve*)]

One way is to have sensors at each corner and track the delay between each of them recording the (presumably) fast rising edge of the signal when the board gets hit by the ball.

This would then allow you to triangulate the position from and three of these sensors (and get some idea of accuracy by doing more than one triangulation)

 

[Rleo6965]

Ooops. I have to delete my suggestion. CocaCola already suggested the XY coordinate with beam sensor.
I think that's more reliable and simpler. No mechanical component to wear.

 

[usermar]

I like the beam idea too....it seems to have less mechanical stress therefore more reliable...but...then...do I need to build special tennis ball to be able to track them or the beams can easily track normal tennis balls as well?

Wow!!! You're really outstanding guys!

 

[Electrobrains]

The beam method is probably the cheapest and best idea till now. You wouldn't need any special balls, just a solid object interrupting rays of light.
If you put your hand in there you would also detect the place - maybe something on the small scale for boxers or shooters who want automatic reading on where they hit the targets...
Steve's idea with 3 or 4 sensors (vibration or sound) at the edges I think is a very interesting high tech solution! On long term, with a lot of units sold, it would probably be the cheapest thing. The disadvantage is that it would need advanced time measurement and signal processing, which would give higher cost there (for smaller amounts).

 

[Rleo6965]

No need for special tennis ball. Ordinary tennis ballonly need to obstruct the XY infrared beam ( Horizontal and vertical arranged as grid ) before it hit the wall. Distance of beams from wall would be 2".

 

[KrisBlueNZ]

I was going to suggest Steve's idea as well. I think you might want at least two bump sensors on each edge. I don't think timing measurement would be a problem since vibrations move relatively slowly compared to a microcontroller's input capture clock.

I assume there will be a clear delineation between 'above the net' and 'hit the net' in this application. You could have two separate sheets, one for the net area and one for above the net. That would make detection more accurate around that critical area. Or perhaps extra sensors near the borderline area.

I don't favour the shadow detection idea unless you have a very bright light perpendicular to the surface. The matrix of lasers sounds excellent, but tricky to set up and very expensive. The spacing would need to be less than the diameter of the tennis ball, and the grid could be placed say 1.5 tennis balls diameter away from the surface. Then by timing the exact interruptions of each of the four interrupted beams as the ball approaches then leaves the surface, you could calculate the speed, angle, impact point, etc. That would be a very interesting project!

 

[CocaCola]

The matrix of lasers sounds excellent, but tricky to set up and very expensive.

I beg to differ on both accounts, small laser pointers are dirt cheap now and focusing IR LEDs or what not isn't overly complicated either... Setup IMO isn't that hard at all, especially if you used machined rails to hold the sensors that would make alignment almost brainless... You could even use larger lenses on the receiving end to compensate for small misalignment, yes this would slightly diminish x/y accuracy but that is something that would have to be taken into account...

Is it the cheapest and easiest option, likely not but I wouldn't consider it very expensive, tricky or costly in itself... A lot depends on the final goal and budget of the project...

 

[Electrobrains]

If the very interesting features suggested by KrisBlueNZ (speed, angle etc.) not would be necessary, a low-cost version could be made by letting the laser beam reflect back and forth a few times through mirrors.
By doing that you could cover a whole surface with only one ray (one transmitter and one receiver). That would save cost, but would need good adjustment of the mechanical parts.

 

[KrisBlueNZ]

I beg to differ on both accounts, small laser pointers are dirt cheap now and focusing IR LEDs or what not isn't overly complicated either

I bow to your experience.

Setup IMO isn't that hard at all, especially if you used machined rails to hold the sensors that would make alignment almost brainless... You could even use larger lenses on the receiving end to compensate for small misalignment, yes this would slightly diminish x/y accuracy but that is something that would have to be taken into account... Is it the cheapest and easiest option, likely not but I wouldn't consider it very expensive, tricky or costly in itself... A lot depends on the final goal and budget of the project...

Agreed. I would say it would be the most fun option anyway

 

[KrisBlueNZ]

If the very interesting features suggested by KrisBlueNZ (speed, angle etc.) not would be necessary, a low-cost version could be made by letting the laser beam reflect back and forth a few times through mirrors.
By doing that you could cover a whole surface with only one ray (one transmitter and one receiver). That would save cost, but would need good adjustment of the mechanical parts.

I must be missing something. If you're using a single beam reflecting back and forth, and a single receiver, how can you tell which part of the beam got broken?

 

[Electrobrains]

I must be missing something. If you're using a single beam reflecting back and forth, and a single receiver, how can you tell which part of the beam got broken?

In such low-cost version, I don't tell which part of the beam got broken!
For instance, if the beam is bouncing back and forth horizontally from net level and 20cm up, it could be used to indicate an optimal ball hit in that area. If one simple vibration sensor is added to the board, it could be combined:
Area hit AND Sensor= green lamp
No area hit AND Sensor= red lamp
Area hit AND No sensor= nothing, it was a fly!

 

[CocaCola]

I bow to your experience.

Just to give an example

These laser modules are less than $1 delivered, if you purchase in bulk...

http://www.dealextreme.com/p/2-5mw-focusable-dot-laser-module-4-pack-3v-46390?item=18

And LDRs are pennies...

It's certainly not a supper cheap option, but I cringe at the cost of this project anyway... Plexiglass has got stupid expensive, I don't know how thick these walls are but a 1" thick 4' x 8' sheet retails for $500 - $600 on it's own... These walls are going to cost thousands in raw material alone, unless they are flimsy thin...

 

[KrisBlueNZ]

I think we should all reconsider the requirement for this setup to be OUTSIDE in the weather, with dirt, dust, falling leaves, polluted rain, etc. I don't know whether ANY of the suggestions made so far would be practical under those conditions.

I expect there are already solutions out there, for professional tennis players and/or elite gymnasiums (gymnasia?) but they would all be indoor. If you can afford that kind of system, you can afford a room to put it in.