The RMV Chip is in commercial production in the avionics industry; therefore, I will not be able to write anything about it due to contractual obligations. I can write a little about its history, and the origins of the idea. This chip is my main and only commercial contribution to the avionics industry and the electronics industry in general. It is a name that I thought of and stands for Remote Management Vectoring.
I was a huge fan of Airwolf back in the 1990s, and was always interested in the futuristic electronics technology that they showed. There was a switch labelled RMV on the right-side control panel where the character Dominic Santini sat. I do not know if that switch did anything, but I was always very impressed at seeing it.
You can see this in the very first episode where they show a brief shot of a switch labelled 'Push to RMV', followed by a graphic display of the rotor blades going through automatic pitch adjustment. They also show it on the starting credits. Probably some sort of initialization procedure I used to think back then. I was very impressed by that and always wondered how it worked.
Since I studied Applied Mathematics at college, I remembered vectors, having direction and magnitude, so vectoring was a logical and correct choice of word. My chip is more than just an autopilot chip, because it filters all remote signals sent from the ground and manages them for the autopilot, hence RMV.
One of my hobbies has been Building and Flying Planes made of balsa. In 2002, I had just lost a plane due to transmitter failure when it flew out of range! This was a huge loss to me as glow engines, servos, and hardware are very expensive components on model aircraft, and I was not working so money was tight. After the loss I decided to leave model flying for a while, I did not have any money anyway.
A year later, I read about Atmel's new range of RISC based microcontrollers that featured built-in flash memory and I remembered one of my lost dreams was to build a development system many years prior. Since this was just within my budget, I decided to continue that dream and began work on designing the system on paper.
Whilst designing the ATMega32 Development System, I realized that it might be possible to build an autopilot chip that could control a model plane. RISC based microcontrollers require very little power, and can run from a single 9 V battery for hours.
I used simple tilt sensors installed inside the wings for maintaining a 5° bank, to turn its course by 180°, effectively returning it. The microcontroller chip controlled the servos directly, and all the remote control commands filtered through the chip. I wrote a procedure for every control surface movement. This was almost the first remote fly-by-wire design, because if the microcontroller chip failed, I would lose control of the model!
After much saving up and personal sacrifices that would have made Warren Buffett proud, I had the funds to begin a new balsa construction with a 2-foot wingspan that could house my autopilot chip.
My C programming has always been quite good, so I wrote the first procedure that would sense the carrier signal of the transmitter. As soon as the “carrier_sense ()” procedure returned zero, the autopilot routine would kick in. I called it “search_carrier ()”, which involved making a 180° turn on a 5° bank and continuing in that heading for 3 minutes. It continued flying in a circle with a slowly increasing radius until it detected the signal. It was a simple idea and code, which worked wonderfully every time, and I was able to test it by simply switching off the transmitter and seeing how the model flew on its own.
Since the chip had some more flash memory left I decided to write a safe landing routine called “M_ILS ()” function. Commercial aircraft use an automatic landing system called ILS, which consists of a set of beacons and sensors installed along the runway to bring the aircraft in on a safe glide slope.
Here is Me Landing a Boeing 737-400. Performing a textbook landing of a 747-400 without ILS assistance on FS to give you an idea of what is involved when landing any kind of craft.
Bringing a model plane in on an automatic glide-slope approach vector is complex because a model plane has no dedicated runway with beacons. The only thing I had to work with was the carrier signal of the transmitter. The strength of the signal is inversely proportional to the distance from the transmitter, as the Physics of these things generally go; hence, it is possible to use this information in a useful way.
My automatic landing routine simply hones in on the transmitter carrier and attempts a landing on a glide slope of 200 m length. I decided to use a large glide slope as a margin of safety for on-lookers to give them time to move out of the way. The plane always lands within 5 metres of the transmitter, and is completely safe.
After watching a James Bond film, one where he remotely controlled a BMW Series 5 car using the touch screen of his mobile phone, I was inspired to build a similar control using a Palm V touch screen to control the flying model. This worked very well but its battery does not last for long, so I only used it occasionally to impress people. However, I still prefer the dual joystick on the remote transmitter.
I demonstrated my gadgets to my friends who are also into building and flying model planes and they were completely amazed. There was a Japanese man in our group who was into real avionics research, and he put me in touch with his research firm in Japan. They bought my source code and circuit diagrams straight away!
After a few months, they contacted me again to ask if I would be interested in developing a more complex autopilot system involving a more powerful RISC microprocessor for an unmanned surveillance craft. Obviously, I jumped at the chance as I was looking for a job.
The RMV project took one and a half years, and the final code was just over 10 MB all written in C++ by yours truly. This was a huge project and easily beyond doctorate level. They were amazed that someone without an electronics degree was able to do so much, they were happy to go along with all of my work including the additional functions and enhancements, and even decided to call it the RMV Chip!
Cultures such as the Japanese are a highly evolved people and place skills and ability above everything else. In UK the colour, the right college, and the right name is more important.
Since this project is now in commercial development, I have decided not to publish any technical information about it on the Internet. However, there is my ATmega32 Development System that I have designed. I mainly get English students copying my code and circuits, often asking me to do their project work for them...