I have thought about this project for some time.  I was considering building the boat then figuring out how I was going to control it.  I have a good idea what size hull I am going to use for the test bed.  Not sure exactly which one though.  I’d like to keep cost and initial complexity down by choosing things like only using one motor.

The boat is going to be a stock out of the box kit.  No design problems there.  Just assemble or build it like all the others of its type.  It just has to be big enough for room to work and make changes.  Detailing the vessel is a whole other direction I’d like to include in this blog,  when the controls become less of the central focus.

The more I thought I realized I was going to tackle the problem from the wrong direction.  The controls ARE what is going to be the hardest (most complex) part to design.  They need to be the first part of the project.  If I can’t get them to work on the bench,  then I don’t need to put together a boat.

The computer components are very small and light so size,  weight and power usage are no problems.  But it is like scratch building a model or maybe a ham radio where there is no complete set of plans or all the parts in a box.  No kit with parts already selected that are ready to install.  That’s why the communication system is the cutting edge here.  I chose the Parallax folks for the hardware to keep the parts shopping at the minimum.  They do have a lot of the parts in the box for the robot kits,  but no boats in sight.

There is nothing sacred with my choice of supplier.  It just works for me

The first version control system will be on a prototyping breadboard and crude looking.  That’s why it will be built and tested on the bench,  using the 1 milliwatt transceivers and prototype plug in boards.  It will be easy to make changes.  Soldering parts on a production prototype board is not so friendly to changes.

I am trying to temper my enthusiasm for details by trying to have a suitable plan of action that won’t keep the project on the bench forever.  I see so much that can be,  that I have to slow down and think what has to be.  I don’t need to jump in the water with a fully functional space shuttle.  Say… there may be a surplus sale on one of those babies fairly soon…

Here is my thinking.  Stage one is a working communication network with rudder control (one servo) and motor control (one motor controller).  All inputs from the shore.  No more than R/C can do.  To prove bidirectional LON communications something will be monitored like maybe a compass heading or motor temperature displayed on shore.

Stage two will be to add the GPS.  Then the programming gets tough,  like PID loops on the rudder to hold course or turn to a course.  More feedback such as true course and distance and true speed can be added to the com link.

Prototype operation can be built into the programming to simulate real world reaction times.

Stage three would be the elimination of the full laptop on the shore terminal.  A hand held terminal box Like an R/C radio control supplies all the information using it own embedded microcomputer and a small display.  Wow, this could take some time,  trying out all the ideas.