Anti-missile defense system

[EEngineer88]

Hello all,

To fill you in on my situation, I am working on an anti-missile defense system senior design project:

The project consists of a camera, micro controller and the physical interception system. Moving object detection will be done on the stable camera using image processing (image subtraction..etc). Information will then be sent through the micro controller to control either a stepper (in which I have to assemble to the 2 steppers to form the pan tilt module) or a ready-made pan tilt module such as:

http://www.servocity.com/html/spt50_sub-micro_pan___tilt.html

Here comes the challenge:

I need a pair of motors that are able to move extremely fast to keep up with the time budget of the image processing being done. The whole process (should be less than 10ms) which means I need to find motors that can rotate as fast as 0.001s/60degrees. However, the motors also need to also be as precise as 0.5 degrees/step as the ball that were throwing would be around 2-5m in front away from the camera.

I found some nice steppers from "Oriental Motors" as well as "JVL" and "Ezi-STEP ALL", However I couldn't not find a ready made pan/tilt servo module that is fast enough.

I will be mounting a laser or a small torch on the motors.

Your help would greatly be appreciated on what kind of motors are best for this situation of speed and precision.

Thank you

 

[PsiStar]

I couldn't resist this

A little more serious, your problem is a all about feedback control systems. Nichols charts & the like. If you haven't studied this yet, hopefully you will soon or at least talk with someone. An issue with fast steppers & response time is under, over, or optimal damping which are directly affected by the mass being moved.

I had a stepper driven positioning table once that was under damped. The thing bounced back & forth like crazy. A small weight attached to the system drive shaft was all it took to settle it down.

Compumotor may still have information on their web site for designing some of what you need. They are pricey although there is always ebay. Yet that mosquito tracking thing really is what you want isn't it?:awe:

 

[William Gaatjes]

These websites may help you :
http://www.evarobotics.com/_product_71617/FW-A001_EvoDrive_ST-17
The evarobotic systems gives stepping motor drives a servo feedback system, greatly enhancing control.


http://www.kelinginc.net/

According to my colleague DC brushless servo motors work best.
http://www.kelinginc.net/DCBrushlessMotor.html

But when not to much complexity is required the right stepping motor with a proper transmission /gear set(Do not know the proper word) will work fine.

 

[EEngineer88]

:awe:, that thing would be impossible to build as an undergrad! but very awesome indeed!

Well I forgot to mention that I am going to implement this in 2D which makes things much simpler!

Regarding the steppers, I finally found some high precision ones with a basic step angle of 0.36 degrees per step.

[FONT=&quot]Oriental Motors PK546PMA[/FONT]
&
1. NMB Technologies Corporation PG35S-D48-HHC2

Lets hope shipping doesnt take long!

My next task is to form a pan/tilt module out of the 2 steppers!

 

[William Gaatjes]

You will be very happy to do some background reading about the resonance and torque control of stepping motors. If you are going to program and design the entire stepping motor control circuit yourself, you will need this information to use the motor as good as possible.

This might help a bit :
http://www.solarbotics.net/library/pdflib/pdf/motorbas.pdf

http://www.cs.uiowa.edu/~jones/step/physics.html

 

[TecHNooB]

I couldn't resist this

A little more serious, your problem is a all about feedback control systems. Nichols charts & the like. If you haven't studied this yet, hopefully you will soon or at least talk with someone. An issue with fast steppers & response time is under, over, or optimal damping which are directly affected by the mass being moved.

I had a stepper driven positioning table once that was under damped. The thing bounced back & forth like crazy. A small weight attached to the system drive shaft was all it took to settle it down.

Compumotor may still have information on their web site for designing some of what you need. They are pricey although there is always ebay. Yet that mosquito tracking thing really is what you want isn't it?:awe:

Nice video

 

[PsiStar]

I have tried to resist this al-l-l day, but I am weak. I suspect that they are sold out or at least not available from woot. But of course there is a support forum & perhaps the "new" has finally worn off for someone & a used one is available for cheap? LOL

I don't know how you are really going to demonstrate that your project works, but this could be a cheap way ... and it is already USB. Add camera + programming. :awe:

 

[DrPizza]

Have you taken a look to see what they use for the laser mosquito zappers? Apparently they've been built with common, off the shelf components. Google it & you'll probably find something. I believe they had TED talk about it at one point. Not only can they track the mosquitos in real time, but they can use the lasers to determine from the frequency of the beats of the wings whether it's a mosquito at all, and whether it's a female (only the females were killed.)

 

[uclabachelor]

Hello all,

To fill you in on my situation, I am working on an anti-missile defense system senior design project:

The project consists of a camera, micro controller and the physical interception system. Moving object detection will be done on the stable camera using image processing (image subtraction..etc). Information will then be sent through the micro controller to control either a stepper (in which I have to assemble to the 2 steppers to form the pan tilt module) or a ready-made pan tilt module such as:

http://www.servocity.com/html/spt50_sub-micro_pan___tilt.html

Here comes the challenge:

I need a pair of motors that are able to move extremely fast to keep up with the time budget of the image processing being done. The whole process (should be less than 10ms) which means I need to find motors that can rotate as fast as 0.001s/60degrees. However, the motors also need to also be as precise as 0.5 degrees/step as the ball that were throwing would be around 2-5m in front away from the camera.

I found some nice steppers from "Oriental Motors" as well as "JVL" and "Ezi-STEP ALL", However I couldn't not find a ready made pan/tilt servo module that is fast enough.

I will be mounting a laser or a small torch on the motors.

Your help would greatly be appreciated on what kind of motors are best for this situation of speed and precision.

Thank you

I hope you're using state-feedback control for this problem, otherwise things will get really difficult.

There are no off the shelf setup available because the requirements (1ms / 60 deg) translate into a motor that has to spin its load at 10,000 RPM and stop dead when in fact all it has to do is spin at its max speed to get "near" the object then track it at a much slower speed.

I would use a servo motor + optical encoder for each axis.

 

[Modelworks]

The motor design you need is literally a few feet away from you. All you need to do is make a larger version of the same servo that is in every hard drive to move the heads. They are extremely quick and precise. They also scale up quite nicely . I saw a laser engraver made by an engineering student using the exact same design just with larger parts and currents.

You need two permanent magnets, some wire for the coil and electronics to control the movement. There are plenty of off the shelf controllers that can do that part. May require a transistor or two to handle the extra current.

 

[gsellis]

KISS - If I remember, Sprint used a paddle vane in the exhaust stream. It gave it incredible directional change. Or is the system doing interception in "coast' mode?

Sprint
http://www.youtube.com/watch?v=LsnkmpJhzlo

And real time

http://www.youtube.com/watch?v=uBLQ-f1i2uQ

 

[exdeath]

I'm more curious how you are going to track an object in the visible spectrum with a cheap CMOS/CCD camera? That goes into the realm of computer vision and is something else entirely. There is a reason all real missile trackers use radar and IR. Trying to seperate a moving object from a moving background in the visible spectrum just isn't feasable, especially with a typical cheap consumer high noise digital camera sensor.

Sure you can whip something up in Matlab to track a laser pointer on a white board with a web cam easily enough, but turn down the lights and point it at your carpet and try to track a rolling ball with all sorts of shadows and pixel jitter and watch the false detections and noise fly.

Unfortunately getting a liquid cooled high res IR sensor is cost prohibitive but that's what real systems use because objects of interest stand out from background and are infinitely easier to track.

http://en.wikipedia.org/wiki/Exoatmospheric_Kill_Vehicle

Also +1 on voice coil servos. Ditch the stepper servos and see if you can find a 2 axis voice coil operated gimbal that can instantly snap to any point in a cone/hemisphere. You don't need encoders, your feedback to modulate field current in the voice coil is based on the camera input and delta between the center of the sensor and the tracked object. The feedback system from the image tracking keeps the gimbal on track of the target, and you know physically where the gimbal is pointed based on how you are holding the voice coils. Encoders on the gimbal axis can always be used if you need to know absolute angle with greater precision.

Also it's alot harder to intercept when the tracker is on a base platform firing a remote interceptor with no onboard feedback. It's easier when the seeker is in the interceptor because the closed loop only need steer the interceptor such that the target stays in the center of the seeker.

 

[exdeath]

Your help would greatly be appreciated on what kind of motors are best for this situation of speed and precision.

Optical tracking gimbals in real missile seekers are free floating and use voice coils, like a hard drive head as mentioned.

 

[LiuKangBakinPie]

You dont need a motor you need a rocket....

Oh thats for intercepting it

Dont you have radar?

 

[LiuKangBakinPie]

Hello all,

To fill you in on my situation, I am working on an anti-missile defense system senior design project:

The project consists of a camera, micro controller and the physical interception system. Moving object detection will be done on the stable camera using image processing (image subtraction..etc). Information will then be sent through the micro controller to control either a stepper (in which I have to assemble to the 2 steppers to form the pan tilt module) or a ready-made pan tilt module such as:

http://www.servocity.com/html/spt50_sub-micro_pan___tilt.html

Here comes the challenge:

I need a pair of motors that are able to move extremely fast to keep up with the time budget of the image processing being done. The whole process (should be less than 10ms) which means I need to find motors that can rotate as fast as 0.001s/60degrees. However, the motors also need to also be as precise as 0.5 degrees/step as the ball that were throwing would be around 2-5m in front away from the camera.

I found some nice steppers from "Oriental Motors" as well as "JVL" and "Ezi-STEP ALL", However I couldn't not find a ready made pan/tilt servo module that is fast enough.

I will be mounting a laser or a small torch on the motors.

Your help would greatly be appreciated on what kind of motors are best for this situation of speed and precision.

Thank you

http://www.jameco.com/webapp/wcs/st...ics / Motors / Servos, Analog&category=754530

http://www.pyroelectro.com/tutorials/servo_motor/servomotor.html

Hope that help