Echolocation?

Instinctz

I wanted to bring this up with and get some feedback from folks about using this type of object detection.

Bats use echolocation quite effectively with great accuracy as well. If I understand correctly, the ping sensor is a simple echo locator.

Does the PING sensor have any easy to access or implemented ways to detect doppler effects, such as items moving away, or towards you.

I am thinking that in order to achieve the high level of sound accuracy i'm after I will have to design my own Sonar setup, since the ping is a simply distance detection unit. But has anyone else done any work in this area?

Mike Green

The PING has no way to derive Doppler information from the returned sound pulse.· It's a simple tuned circuit with an on/off detector and it looks only for the leading edge of the returned pulse.

Sound processing is very complex whether done on audible or ultrasonic sound.· You need to analyze phase shifts and multiple returns among other things.

Instinctz

Thats kinda what I was understanding, I figure for an advanced echolocation sensor (actually multiple sensors) you would need to reconstruct a bat ear in order to achieve vertical detection and like a bat use the other ear to help detect horizontal differences. Which would require many small (hopefully) sensors tuned within each "ear".

Including the ability to detect a multitude of phase shifts per channel (reciever) and an incrediable amount of multiple returned frequencies, I think a person could create a very sophisticated object detection system.

I mainly bring this up because I am not satisfied at all with current technological object detection, or any other means of 'sensing the enviroment' that has been seen in other robotics. Take asimo, the supposily most advanced robot out there, can not climb stairs with out each and every action being exactly programmed. That isnt something I would expect after billions of research and development. Even then it still fails quite often.

I know this subject is complex, and most agreeably out of my scope, but... that is the realm I work best in.

I was just curious if any other fellow Parallax members have went into this field very much...

bjhamltn

The new parallax x-band sensors might be of interest to you.

kwinn

Re "billions of research and development".

It is very easy to underestimate the complexity and processing power of even a small animal's brain. It is not until you sit down and try to write software to have a robot perform the simple every day tasks any creature must be capable of doing to survive that you start to appreciate how complex it truly is. Of course nature has had billions of years in a global size laboratory to get to this point so I am not discouraged. We have only been at it for 50 years or so.

Instinctz

Oh don't misunderstand me.. i'm acutely aware of the great sophistication of even the simplest thing as a gnat.

Which is probably what started me on my project for a humanoid robot, and simply making the thing walk is going to take me past current abilities. What I perhaps am under-estimating is what can I think our current level of robotics is, per where I think it should be.

So, we'll do a little self enlightenment to understand the design process and just maybe I will agree asimo was worth the costs and time. Or maybe I'll make something worth while and ground breaking, unlikely but oh well.

This "echolocation" plan of mine so far seems to be the best way to detect quite precisely the distance of everything somewhat quickly.

tronsnavy

Very aggressive project to say the least.· Would it be practical??? Don't know.· If you are still interested in a complicated detection system such as this, using Doppler techniques might be harder to implement than say, a "monopulse" radar system.· Sensing Doppler frequencies requires analog or digital filters... or using FFT techniques (discussed in detail in the propeller forum).· FFT, in radar speak, simplifies the use of digital filters.· Look it up if you want a more detailed explanation.·Monopulse systems are of two general types... The first is Amplitude comparison monopulse.· The second is Phase comparison monopulse.· The later, I would think, would be more advantageous.· In it, the antenna is divided into halves.· Two lobes are produced by both halves of the divided antenna, point in the same direction; consequently, the return received through one lobe has the same amplitude as that received through the other regardless of the angle of the target relative to the antenna boresight line.· However, if an angular error exist, the phases of the returns will differ because the difference in mean distance from the target to each half.· There is much more to this but it's like having 4 bat ears in one antenna.· Once again, do some research and you will see what I am talking about.· Don't have time right now to explain further, but if you still have questions after your research, post a reply.· Of course, the radars that I just discussed are RF based (speed of light).· Even though you are tying to sense objects only inches away (vice miles), I am not quite sure if a serious bottleneck would result.· Have fun.
Regards, Bob