Laser Model Rocket Tracker
This isn't so much a project as musings on an idea for a project. The inspiration comes from a (truck) portable laser tracking system that Sandia National Labarotories has. Very nice but a little more capable and expensive than what we need.
The basic idea is to try and come up with something that will automatically track a contest rocket and provide azimuth and elevation data. (Laser ranging of the rocket possibly being added later.)
The reflectivity of the average (or even exceptional) contest rocket isn't very good. Using a laser tracker would require adding a bit of reflective tape. The tape used with the Sandia tracker is the Scotchlite (TM) brand from 3M. It isn't quite as effective as a retroreflector (three orthoganel mirrors) but it is very good. Adding a bit of this tape would result in some decrease of the rockets performance but the tradeoff in ease of tracking should be worth it.
The Sandia tracker uses a green laser to illuminate the target. I don't know how much output power it has but it is supposed to be eye safe. I noticed recently that green solid state lasers are now available. They are more expensive than the now ubiquitous red laser pointers because no semiconductor laser puts out green light. They use a several step process to boost the laser output into the green spectrum and output powers are fairly low. The green laser pointer would have to be modified to spread the beam so that it filled the field of view of the optical system.
Tracking the target is done in two parts. The first is a stepper (or servo) motor controlled optical head. This is used to aim the optical system at the target. Feedback is used to keep the target in the field of view of the optical system. The reported azimuth and elevation angles are a combination of the mechanical pointing system and the location within the field of view of the optical system.
The optical system needs a camera. The field of view should be optomized for the purposes of tracking. The typical NAR contest range has pads spread around a 100 foot diameter circle. So a 200 foot horizontal field of view should be adequate. With a typical 4:3 aspect ratio this would provide a 150 foot vertical field of view.
The field of view and mechanical slew rates must be capable of keeping up with the target. At 30 frames per second, a target traveling at 1,000 feet per second (much faster than a typical contest rocket) travels 33.3 feet per frame. It would therefore appear in 4 to 5 succesive frames of the video system if it never moved. This should make sure that the rocket can be tracked optically.
The second part of tracking is that the mechanical pointing system has to be able to keep up with the rocket. The highest slew rates will be required shortly after launch. If the tracker is located 1,000 feet away and we again use 1,000 fps as the velocity, the peak slew rate is then 45 degrees per second. This might be more than can be achieved so it will require carefull study.
Digital cameras are available that output data streams suitable for direct processing by a computer while avoiding that messy step of conversion to NTSC (Never Twice the Same Color) video. A recent article in Circuit Cellar magazine detailed the interface of an OV6620 sensor to an AVR microcontroller.
The full RGB video data would be displayed on a small LCD display to assist the operator in aligning the unit with the rangehead. But only the green data would be processed by the tracking electronics.
I see that there is a product being sold to the robot community that is very close to what is required. It is the CMUCAM2. It has tracking features in its software and it has 5 outputs for RC servos with two dedicated to pan and tilt functions. I am not sure how well RC servos would work given the lack of position feedback plus the mounting schemes on the pan and tilt heads being sold aren't ideal. One has rubber grommets for servo mounting and one even uses sticky tape to mount the servo. Hard to acheive repeatable and accurate postitions that way.