Cheap (inverted?!?) half-assed triangulation.

Projects General Guidance
1

Hi. o/
I have an idea and would like to know if the reasoning is sound

What I want

I have a “arduino controlled 4wd car” and a signal being emitted by another arduino (tower from now on) somewhere in the house.

I want to make the car able to go toward the tower in a straight line (the car can detect obstacles, so it would eventually get to the signal).

To do that, the car must detect the signal and know what direction to go.
My idea (that I took from owls) to that.

Tower emitting signal. 3 receivers (Rx from now on) on the car (red triangles). The receivers are synced. Car is around 250mm*180mm

If I know that Rx 1 received the signal at 1.5 mili seconds, Rx 2 in 1.6 mili, and Rx 3 in 1.0 mili.
I also know that I need to turn the slightest bit to the left AND that I need to turn 180 degrees, since Rx3 should be the last to receive the signal.

One easy way to do that, would be measuring response time. But that is a no go. As self-imposed condition (at least for now), i don’t want communication between tower and Rxs. Tower is pure Tx and Rxs are pure Rx.

With no communication, my idea is the following.
Tower will emit a changing signal.
1, 2, 3, 4, 5, 6, etc…

At a fixed point in the timeline each Rx will receive a different signal

Ex: at instant 3 in the timeline,
Rx1 received 2
Rx2 received 3
Rx3 received 7

In this case, since i know the order and the time between different messages, with some math I can find out what maneuver the car has to make to align itself with the signal in a straight line.

The problem

From one room to another in my apartment, the signal takes around 1 milisecond from tower to Rx (information from the response time divided by two).

The bigger the code, the longer arduino takes to loop around.

Distance between Rxs is too small. Im pretty sure they will all receive the same message at each loop.

Before I start buying stuff (only have one pair Rx Tx with communication and one Rx Tx with no communication) to try this, is this idea possible?

What other problems can you guys think of?
Any idea to go around the problem?

thanks

fig1.jpg
fig1.jpg608×513 23.3 KB

2

What type of signal? If a radio signal, the scheme, as I understand it, won't work at all. Radio waves travel about 1 meter in 3 nanoseconds.

From one room to another in my apartment, the signal takes around 1 milisecond from tower to Rx

That sounds like radio, and you are confusing the time it takes for the message to be processed by the receiver, with the time it takes the signal to travel from one point to another (see above).

3

I don't think owls can hear RF

4

for response time, i used the nrf24. Why would radio not work (wavelength?)?

i would use the signal necessary to make it work

5

Why would radio not work (wavelength?)?

Speed. See reply #1 (edited).

To elaborate, radio can be made to work by using advanced technology, such as ultra wide band radio. Pozyx is such a system, and took years to develop. It is not cheap.

Pozyx fails in situations where reflections confuse the receiver, so Pozyx also includes motion processing sensors and MPUs to keep track of position in those circumstances.

In principle your idea could be made to work with sound, which travels at a speed that an Arduino can easily deal with. Lots of people have tried, with varying degrees of success.

6

One option is to have three fixed light beacons (like lighthouses for ships) and have a light detector on the car that can be rotated and which can identify the angle at which it sees each light.

...R

7

I am doing something like those lighthouses, but in another situation. In this case, lineofsight requirements would eff stuff up. Would wifiwork/bluetooth offer different response times should i use Tx/Rx communication?

9

phmarriel:
In this case, lineofsight requirements would eff stuff up.

Then your project is effed :slight_smile:

What about having more "light-houses"?

Maybe the "light-houses" could flash at different frequencies to allow the robot to identify which one it is looking at.

...R

10

if you are looking to create a directional signal, so that you know what direction the lighthouse is from the current location, then maybe, some directional antennas would get you to the room with the lighthouse.

11

Robin2:
Then your project is effed :slight_smile:

What about having more "light-houses"?

Maybe the "light-houses" could flash at different frequencies to allow the robot to identify which one it is looking at.

...R

(╯°□°)╯︵ ┻━┻ goddammit
any other idea?
how do i tell that little bastard of a robot which way he is pointing without the need of on-site lighthouse?

my biggest problem with them is the setup.
would work fine at home, but any other place would require the removal and reinstallation of them at the new place. dream scenario is not needing constant line of sight

dave-in-nj:
if you are looking to create a directional signal, so that you know what direction the lighthouse is from the current location, then maybe, some directional antennas would get you to the room with the lighthouse.

had not thought of directional antennas.
dont really know anything about them so ill research a little.
do you know how big/small of a angle/cone the see?
mounting the antenna (maybe 3???) on rotating plataforms might do the trick

edit
wait wait wait
maybe i dont even need direction antennas if a make a lighthouse in the robot
an actual lighthouse. BUT INVERTED

i put a rotating cover on the nrf. we a very small opening. this cover would me made of ???? in order to blind it. when i do get a signal, it means the cover is facing the tower. than i just ajust the robot accordingly

12

how do i tell that little bastard of a robot which way he is pointing without the need of on-site lighthouse?

Magnetometer/compass.

13

AWOL:
Magnetometer/compass.

would that work?
in my head, that would give me absolute directions, but in order to know if it is looking at the tower emitting the signal i would need relative.

14

I don't know what you mean by "work", but you asked how the robot would know which way it was pointing.

15

phmarriel:
my biggest problem with them is the setup.
would work fine at home, but any other place would require the removal and reinstallation of them at the new place.

I suspect you could build in an initialization routine in which you position the robot at some position equi-distant from all the beacons and then let it figure out the geometry. Should hardly take 5 minutes of your time.

...R

16

ohhhh
i was asking the wrong question
if i asked about my problem instead of about my solution it might have been easier
direction finding in this case

Robin2:
I suspect you could build in an initialization routine in which you position the robot at some position equi-distant from all the beacons and then let it figure out the geometry. Should hardly take 5 minutes of your time.

...R

i bet this will do for fixed locations. ill try it when i get to the lineofsight tracking.

[/quote]

AWOL:
I don't know what you mean by "work", but you asked how the robot would know which way it was pointing.

that is true. ill be more specific. I want the robot to find the direction a signal is coming from so he can go there and "rescue" the missing tower.

dave-in-nj:
if you are looking to create a directional signal, so that you know what direction the lighthouse is from the current location, then maybe, some directional antennas would get you to the room with the lighthouse.

you have my thanks, dude.

from directional antenas to cantennas to something like this

took no time at all

from here on out i can make something that will work. and hopefully prettier and smaller than the example above

direction finding solved
another question

you guys know what i could do (do this thing to a particular parameter or use i particular kind of signal) to improve my chances of the robot working in the middle of a university.

there will be all kinds of celphones, laptops, alien experiments, wifis, heavymachinery in the campus.
is there something i could focus on to increase reliability?

17

ps
maybe my original idea wasnt that impossible

Psuedo-doppler direction finding works by rapidly switching between several antennas. The difference in the time that the signal arrives at each antenna can be used to calculate the transmitter's direction.

https://www.rtl-sdr.com/building-an-rf-direction-finding-robot-with-an-rtl-sdr/

18

Old 27mhz rc transmitters and receivers have good penetration and not so much else on that frequency I think

19

Bringamosa:
Old 27mhz rc transmitters and receivers have good penetration and not so much else on that frequency I think

thats good to hear..
im pretty sure i have a age old remote control car lost somewhere.
ill try it out.
thanks

20

phmarriel:
ohhhh
i was asking the wrong question
if i asked about my problem instead of about my solution it might have been easier

That is always the case but you will not be surprised to learn that many Threads here take a very long time to get around to describing the basic requirement. :slight_smile:

...R

21

Guess it is ok to close the topic.
Main question has been answered