Using multiple Sharp-IR sensor

[Darkenz]

Hi,

I am currently doing a project where I am required to do object detection between 0-3 m. And due to requirements, I have to use Sharp-IR sensors which I haven't had much experience with before. So I am thinking of tackling this problem by using 2-3 different sensors to cover different ranges. So for example, I would be using one sensor to detect the longer range (1.5m to 3m) and another sensor to detect the shorter range (0m to 1.5m). So my concern is that would there be any interference between these sensors as I would be putting them in parallel to each other. Bear in mind I would be driving these sensors using an Arduino uno.

Any help or direction is appreciated,
Thanks

Kevin

[robotmaniac]

You should read the Sharp IR tutorial... Then i you still have doubts come back
 

[Soeren]

Hi,

[...] So my concern is that would there be any interference between these sensors as I would be putting them in parallel to each other.

You cannot drive them in parallel. Each sensors signal line needs a separate A/D-C input to read them. Their outputs are analog voltage and if one is at eg. 0.2V while the other was at 3.0V... I'm sure you can imagine the issues

The ground (0V) lines could (and should) be connected in parallel to a low impedance ground point, but the positive lines should be switched, or they will interfere with each other.

Each should have a cap over the supply lines mounted right at the sensor and this means that you need a relatively long pause in-between deactivating one sensor and powering up the other (or you can just ignore the first nn ms after each switch.

You'd want to drive each for at least the length of a couple of samples, to rule out noise and spurious pulses (by averaging and dropping vastly deviating values).
Use the "long range" sensor alone until it catches something and when the measurement gets towards the "near" end, switch to the "short range" sensor (and back if the detected object(s) gets out of range again.
You'll want a generous overlap of their ranges for the smoothest operation (not a problem, as you cannot get them any other way).

[Darkenz]

Hi,

Thanks for the quick reply.

Robotmaniac,
Is this the tutorial you're referring to? http://www.societyofrobots.com/sensors_sharpirrange.shtml
if so, yes I have read this article previously, however I dont think it mentioned anything of employing two or more different sharp IR sensors looking at the same direction.

Soeren,
Forgive me if I have misunderstood your understanding, however it sounded like you were thinking that I would be driving these sensors through the same signal and voltage source pin. I know perhaps my question is quite ambiguous but the matter here is not about the wiring and etc.

Lets just say imagine a case where two different type of IR sensors are being driven by 2 prototyping board (arduino) and these two sensor are facing the same object in parallel. Would there be any interference in the readings between these two sensors. Probably a simple drawing would help.

Cheers,
Kevin

[robotmaniac]

yes I meant that one and it kinda those but doesn't go deep abot you help but i think it answers  (at least what I think is you doubt) here it is

An issue that these range finders have in common with sonar is cross interference. This means that the signal emitted by one sensor can potentially be read by another sensor and therefore give you bad readings. However, unlike sonar which have sound signals that can bounce off of multiple walls, you just need to make sure your IR beams do not cross in parallel (the wide parts of the football shaped beam not overlapping). This makes sense because you have over redundant sensors if the two beams cross.

[Soeren]

Hi,

Lets just say imagine a case where two different type of IR sensors are being driven by 2 prototyping board (arduino) and these two sensor are facing the same object in parallel. Would there be any interference in the readings between these two sensors. Probably a simple drawing would help.

I answered that already, in the second half of my previous post, but to be more specific...

The sensor emits a short flash of narrow beam IR, which is then triangulated in its position sensor (photo diode array) and hence, you need to have only one energized at a time, to get good readings. two sensors running out of synch would give false readings - only one must emit at a time and it should be read only from the sensor that emitted the beam.

If you switch between them and want at least 3 readings, let each stay on for ~150ms.
This is only needed in the range overlap and when an obstacle is either nearer or farther away, just keep the relevant sensor on.

[Admin]

Lets just say imagine a case where two different type of IR sensors are being driven by 2 prototyping board (arduino) and these two sensor are facing the same object in parallel. Would there be any interference in the readings between these two sensors. Probably a simple drawing would help.

In parallel, very likely. If both sharp IR sensors operate on the same frequency, and pulse at the same time, then definitely.

I've never seen any interference when putting them at slight angles to each other.

It's easy to determine - just mount them as you like and look at the data in Hyperterminal.

[codyspraker]

I am not entirely sure why you have two pointing at the same object and read different distances. Why didn't you choose to have only one but perform a long range check followed by a short range check?
Anyway the solution I would offer if this setup is necessary is to have only one operate at a time. After the first sensor signals and receives data, then have the second immediately follow and record at the different distance. This will avoid any interference from one another.

[Darkenz]

Thanks guys,

I'm almost definitely sure that they will intefere with each other. I can't do physical testing yet as I am still waiting for all my parts to arrive. However I am now thinking the possibility of using mutliple mosfet to cut off the power going into the individual sensor and switch each one of them on momentarily only when required. Any one has experience with this approach? mind sharing any good direction or reading?

@ codysparker

Imagine designing a device that could detect obstacles between 0 -3 m in 360 degree range. So lets say theres an object 50 cm at 45 degree and 250 cm at 90 degree, I need the device to be able to detect both objects and give readings. With the current limitations on Sharp IR sensor, theres only a sensor which could detect between 20cm to 150 cm and another 100 cm to 300 cm. So one solution to be able to detect both objects while my device is scanning the whole 360 area is to have these two sensors looking at the same direction while the device is rotating.


Regards,
Kevin

[Admin]

If you are scanning, just mount them at an angle to each other. You will then get two maps from the scan, one for each sensor. Combine those two maps in software, knowing the angles of each data point.

[Darkenz]

Well, the case I was describing to cody was just an example. My project itself is a bit different. for my case I need to scan from 0- 270 degree.
From 1 - 90 degree and 181-270 degree (horizontally) I need to be able to scan and detect object from 150cm to 300cm. As for 91-180 degree (horizontally) I need to be able to scan and detect object between 5cm to 300 cm. And all of these also need to be done with 0 - 5 degree vertically. And due to budget Im could only use two motors to do the panning and tilting.

I hope that clears the ambuigity a bit of what I am trying to achieve.

Cheers,
Kevin

*EDIT*

Found an easy to understand yet interesting article that can be used to for this problem. For anyone else who happen to be trying to do a similar approach, this could be helpful.

http://www.sqlskills.com/blogs/paulselec/post/Arduino-figuring-out-transistors-and-associated-resistors.aspx