I have been working my next-generation UAV processor board. My V1 served me well: GPS signal processing, video overlay, power monitoring… and I learned a lot.
Now I’m working on a new design using an IMU to first support gyro-stabilization and eventually autopilot capability. I’m using a two-board design for now (which can be integrated into a single board in the future):
Cerebrum is the high-level control board based on Fez/Usbizi (Panda in this iteration). It will handle power management, communications, routing, UI, navigation.
I finished the v1 design of Cerebellum board incorporating compass, gyro and accelerometer on an arduino shield format. I should have the first PCBs in a week or two. Features:
Parallax Propeller 8-core processor connected to all sensors and shield TX/RX pins
3 sensors, 3 axes each (gyro, accel, mag compass)
connected via I2C
I2C breakout for direct access to IMU sensors and adding external sensors
video overlay circuit using Propeller’s video generation capability
RC 4x Radio in/ 4x servo out (connected to Propeller); can be reconfigured to 1x PPM in/7x servo out
5 voltage dividers connected to shield analog-in
GPS port
LEDs: power, GPS Fix, Video In sync
extra IO for debugging
It is also designed so it can be used in various stand-alone configurations by cutting off parts of the board. Cut off the edges where the headers are and it can be programmed to be a simple on-screen display. Then you can cut off the video section and use it as a gyro stabilization system. Cut off the RX/Servo connections and you can use it as a UART-connected IMU.
I’m looking forward to building this out. I’m also going to start a separate thread on a gyro order–an opportunity to buy in bulk and save money.
Making progress… my PCBs are in and I’m testing my assembly process. These sensor ICs are TINY and I’m going to need stencils to apply solder paste more precisely. I have another Digikey order in the hopper and with any luck I’ll have everything ready for mounting by next weekend.
The problem is that UAV boards tend to be rather specific applications. Although Andrej doesn’t show it right now, getting an autopilot board to work in an RC plane requires a number of very specific requirements. Size is one, power is another… I’m not so sure a merging of the Gameo and the UAV board would be a very good idea.
This has a video overlay/output circuit at the bottom and voltage dividers (5) on the analog inputs. But it doesn’t have any sound output but that could be useful in a UAV anyway (audio modem, variometer, audio alerts, etc.)
Most people flying FPV (first-person video) and UAV are carrying serious batteries on board so the weight isn’t really a big factor. This can definately evolve to something smaller though. I’ll probably end up with 5000mah of batteries on my UAV plane this Spring.
This isn’t just useful for UAV/RC planes though. It can also work well for anyone that wants an IMU for measuring movement/acceleration inside of a car, experimenting with self-stabilizing robots, boats, etc.
P.S. Anyone here geeky enough to enjoy a solder reflow video? Bah, anyone here NOT geeky enough? Action starts at 2:30ish
The problem is the uav is doing video overlay onto an exisiting video stream, it however cannot generate it’s own video. The Gameo would need to do this.
Sure, it can do video overlay or video gen. If the video gen software detects a video sync it uses that and if it doesn’t it generates a sync. Then theres a MOSFET switching circut that lets the input video through or not depending on if that paricular place on the screen should be transparent. The prop is cool because it can do this with just a few pins and software, and a pretty simple DAC circuit.
Gus pointed me at this a while back. This board can also generate audio but I didn’t include those circuits.
I found out tonight that I used the wrong pad size for my resistor arrays so I’m going to have to carefully solder individual SMT resistors on the same pads instead. That should be interesting.
