I’m currently planning the design of an autonomous blimp (modified from a simple RC model), and I need to choose a microcontroller. Also, as it’s my 1st time building a robot, I also require a debugging board.
I figured the PicKIT2 might be a good starting point; however, because this is an airborne robot, weight is a huge issue. I’m sure I’d need the debugger board for land-testing, and making sure that my code works. But using the debugger board to power/use the microcontroller on the blimp would probably be too heavy, once the robot is in operation.
What I need is a smaller board that the PIC can attach to, that can also attach to the debugger board, and once the PIC program is working, can be removed from the debugger board to be directly attached to the robot, and run normally.
In other words, I’m wondering if there is a circuit board that can act as an intermediary between PIC and debugger board, such that all I need for actual operation of the robot is the PIC and intermediary board? By doing this, I believe I’d be saving a lot of weight. What do you guys think?
I’m reading over the documentation for the baby Pololu Orangutan (B-168), and it sounds to be just about the best controller for my project. However, there are two questions I have about the baby Pololu Orangutan.
First, is debugging ability. My guess is that to debug the baby controller, all you have to do is attach the controller to a breadboard and simulate inputs and outputs with LEDs to determine if the program actually works, before plugging it into the other sensors and motors? This may be fairly obvious to the more experienced robotician, but my first “experience” was with a debugger board with full out LEDs, buttons, number pad, and LED screen, and I`d rather not invest in a whole debugging board if the baby Orangutan board is capable of debugging itself.
Second is failure. If something goes horribly wrong and I burn the PIC on the baby Orangutan, can I replace just the PIC? That is, desolder the PIC from the board and solder a new Atmel mega168 AVR microcontroller to the baby Orangutan board? iI figure it`d be cheaper to just get the microcontroller, as opposed to getting the whole Orangutan board. I quickly tried looking online, and couldn’t actually find suppliers of the microcontroller itself (only debugging boards and Orangutan boards). Thoughts?
Oh yes, and to provide an update on the project, I figured I`d take the easy route and simply purchase an RC blimp and modify it with various components.
That’s exactly what we’re doing. I’m following along the lines of my second year design project, where we broke the design of an autonomous pipe-crawler into electromechanical, circuits and microcontroller subsystems, and then tied them together, so to speak.
The nice thing with buying an RC blimp is that the majority of the electromechanical/structural aspects of the project are already done. No hassle with leaks in the blimp’s envelope. The RC blimp also comes with propellers, 2 motors, and a remote controller. I’ll want to be able to use the remote controller to override the microcontroller in case the object-avoidance goes awry.
It’s been a while. The blimp project got put on hold because I got really busy with work. Anyways, a new thought has popped up on the autonomous blimp project: wireless control.
I gave it some thought, and realized that what I wanted was not just a programmable autonomous blimp, but an autonomous blimp that could act as a test platform for various sensors and algorithms. I want to use my computer as a ground station that could have a program on it, and send instructions to the blimp, have the blimp perform some action (like thrusters on), and then the blimp’s sensors pick up measurements, and send them back to the ground station / PC as feedback and decide what to do next.
I was reading an old issue of SERVO magazine (July 2004) that talked about PC to RC; however, they used a big, huge expensive remote controller that probably cost a couple hundred dollars (out of my price range :P). So I was thinking of using the SFE - RF Link - 2400bps - 315 MHz as a wireless communicator between blimp and ground station. I figure I would also need a PIC to translate computer instructions to PWM for more efficient control of the blimp’s motors. The tutorial for the SFE - RF Link - 2400bps - 315 MHz device references PIC18F452, although for no particular reason. So the architecture would be as follows:
PC -[instructions]-> PIC -[PWM instructions]-> SFE-RF Link transmitter -[wireless instructions]-> SFE-RF Link receiver -[PWM instructions]-> motors/sensors on blimp -[sensor data]-> …yadda yadda… -> PC
Does anyone have any suggestions on the wireless transmitter/receiver I should use, as well as the PIC? Do I even have the general architecture understood?
The ideal solution here is to make your own printed circtuit board with ISP capabilities. Making your own PCBs is fun and allows you to tailor it to your specific needs.
A great way to make your own PCBs at home is with a Photofabrication kit from MGChemicals:
If you can not debug from your software environment, using LEDs is a valid option but does not simulate actual project conditions. You could also use an osciloscope to debug outputs.
Ideally, work on each subsystem independently then integrate them together into one coherent system. There is however no problem debugging with motors and sensors attached.
The choice of programming environment is up to you. WinAVR is probably the best choice in terms of full featured environments but we suggest you try as many as possible to find the one that suits your specific needs.
The Orangutan (as well as many other microcontrollers) needs only a USB programmer for easy programming. As 710ml indicated, you would essentially connect the appropriate wires between each.
The PIC used on the Baby Orangutan is surface mounted. Unless you are very skilled at soldering, these connections are very, very difficult to make and break. You would be better buying a new board.The best way to avoid frying a board is to read all appropriate documentation as you are doing.
The difficulty is not in programming the PIC, but the the additional circuitry that is required. Many microcontrollers are PIC based, and include all the additional circuitry required. Your application calls for a tranceiver system (two-way communication). Consider the XBee transceiver module with antennae, and the XBee Explorer USB (which connects directly to your PC). XBee modules do not have standard pin spacing however. You would need a microcontroller (or PIC/Atmel IC only) and a motor controller (unless you are able to hack the motor controller that came with the blimp).
Although I’ve never built one, i have looked into its feasibility. To control the airship you will need a minimum of 3 motors, all operated separately (two for L/R, one to pivot the motors up and down). Consider using the Pololu Orangutan (which has two motor controllers built in, and you do not need to attach the pin headers to save weight). Very few Pics can match that for weight. Or, you can use the Pololu Dual Serial Motor Controller (featherweight) and Solarbotics micro motors (they can almost be inhaled). The whole setup should weight about 35g including 2 AAA batteries. You’d need to build a custom PCB to save weight. You would need ~3 cubic feet of helium to lift all of it, though simply buying a commercial (hobbyist) mini blimp will save you the hasstle of leaks.
The difficult part is powering the microcontroller, so you might want to use coin cells (maybe consider using spray-on solar panels?). You can adapt two tail rotors off a micro heli (~$6 each) for propulsion. Any programmer you choose can be connected to the microcontroller (or PIC) using just two or three wires (essentially you will need a "docking station). Hope you’ll add some pictures here and keep us updated.
A side note - make the blimp slightly negatively buoyant - otherwise you lose it!
