Naturally, being most excitable about robots and autonomous things, the probes and rovers are my favorites. Rovers are easy to do, I build them all the time. Probes, on the other hand, not so much. You can do an underwater ROV, which is kind of similar, but still. Of course, there's another aspect of space travel that involves automated systems- launch vehicles. The rockets themselves are full of automated control hardware, guidance, monitoring, and tracking.
Of course, I can't put a model rocket into orbit. Yet, anyway. For now, I've been inspired by SpaceX's attempts to soft-land their first stage booster. I don't have access to throttleable rocket engines (Yet), but I figured I could design some control surfaces, implement tracking and control, guide the rocket to a given destination ballistically, and then trigger an inflight burn of some smaller braking motors to make the landing survivable.
Honestly, it's more of a guided missile, but NASA has used missiles to launch things before, so why can't I?
Basically, this project has three goals. First: Get the rocket high enough in the air. This might seem like a weak goal, but there's going to be a lot of instrumentation on this machine, more than most hobby rockets carry. It also needs to get sufficiently high to make it to a given destination. I'm going to shoot for a 100 yard target. Assuming a roughly parabolic flight path (course corrections will have to be minor, this isn't a rocket-propelled drone aircraft) after a vertical rise and roll maneuver (I know it will want to come straight down after a roll, let's pretend for estimation that the flight controlled path causes an approximation of a parabolic descent). Then because the apex is directly above the launch point, the rocket needs to reach a minimum of 100 yards (300 ft, the destination distance) height.
That doesn't seem like much, since most commercial model rockets advertise about 1200 ft apogee on a C motor, but remember most model rockets aren't carrying avionics, flight control and instrumentation. I figure, by looking at the net kinetic energy (since in the model rocket, propellant mass fraction is small- this is amplified by our greater payload weight), and assuming a net 15% loss due to air resistance (remember, it's not going as high, so less total wind friction) and a 3x weight load, I figure I can get approximately 340 ft (ish) out of a C motor. Obviously, one of the elements of early testing will be characterizing this flight profile, so these numbers aren't so made-up.
The second goal is to aim the rocket at its destination. This comes in two parts- first, correctly rolling the rocket so that its descent is in the appropriate direction, and then using small control-surface adjustments to guide it home. Guidance will require tracking, which I plan to implement as a combination of Accelerometer, Gyroscope, Altimeter and GPS systems. They'll approximate the position, orientation and dynamics, which will feed into a relatively simple (i.e. probably PID) controller, based on correcting the projected landing point. But that part is a long way away (and many flight tests). I think it will be a single servo driving a pair of flight wings- a very simple flight profile. Launch orientation will be required, naturally (I can only fit so much hardware into the little rocket, after all).
The first part of that second goal, the roll, is more immediate. My idea is to use the ejection charge, which normally deploys the parachute, as a sort of one-shot, pre-set RCS. I've built a small pressure vessel which includes 4 outlets. One such port is marginally wider than the others, which allows a higher rate of flow through it. Because the expansion is rapid, this should cause a proportionally larger thrust at that port, enabling the start of a ballistic pitch in the desired direction. I've built a prototype apparatus and plan to test it soon. I'm giving 5-1 odds it explodes.
The third goal is survivability. The notion is that as the rocket descends, it should use a set (I'm thinking three) of smaller motors to slow its descent. These would be triggered when flight systems detect that the rocket is within a certain distance of the ground (to be specified after the first flight profiles are made- the timing will depend very much on the specifics of the descent). The trick being that the rocket needs to jettison all the braking boosters when the first one goes out, so as not to start a spin close to the ground (likely fatal). This means I need a detection sensor to identify when a motor is on, and when it's off. That was the first experiment I ran, checking a thermistor, IR sensor, and Photosensor:
The other thing I need is in flight ignition of the braking motors, and on-cue activation of the deployment charges (which I'll use to jettison them when necessary, and potentially also as range safety on the main motor). These are the next tests I need to structure. The former has a technical challenge- getting sufficient current for ignition. I tried using small CR2303 batteries, but could not trigger ignition this way. I'm considering using capacitors charged on the pad, but have yet to test (planning to try it when I set up the roll trigger test). The latter I expect will, once the ignition is solved, be relatively simple by adding ignitors to the explosive charge directly, but I need to check if that is sufficient to trigger the charge. I plan to combine this test with another instrumentation experiment, and have the controller attempt to trigger the detonation as soon as it detects end-of-burn. I don't want to test this with a full compliment of propellant still in the tube.
That's where I'm at right now. I need to do these hold down tests to work out technical issues, and then I can do a launch to test avionics and get a flight profile, and work on control from there. I have some notions about a pair of wings for control surfaces, actuated by a single servo, but it's not well formed yet. Hopefully I'll be geared up to do a first actual launch by the end of the month, as long as not too much explodes.