Aber's Autonomous Sailing Robot

Lake Trials for Minty2

This week we’ve been frantically trying to get Minty2 ready to go to Greenland onboard the yacht Gambo next week. We’ve now had two days of testing at Bala lake in Snowdonia and are now fairly confident we’ve got a non leaking hull and reasonably working control, telemetry data, remote control and video telemetry systems.

We had to make a last minute change to our motor control system after we realised the motor controllers we’d chosen weren’t delivering enough current. After a frantic redesign we’ve chosen two 4QD VTX 40-12s (http://www.4qd.co.uk/prod/vtxbx.html) controllers that take analogue inputs. Fortunately our Labjack system that is being used for current monitoring also has two spare digital to analogue outputs that we can use for driving the motor controllers.

Introducing Minty2

Minty2 is a motorised robot boat intended for surveying calving glaciers in Greenland. This is to help a group of geographers who want to measure the amount of ice falling off the glaciers during calving events. They do this by building a 3D model of the glacier front with a 3D laser scanner and a resolution SONAR. To produce the 3D image requires travelling along the front of the glacier, a few hundred metres away from the ice. The problem is that this can be extremely dangerous should some ice fall off, as large chunks can fall off without warning and create very large waves or hit the boat directly.

Minty2 also helps us in performing research and acting as a demonstrator for power management techniques that we are currently researching with EADS Foundation Wales. Essentially we are treating Minty2 like an electric powered UAV operating in 2 dimensions. Thanks also go to EADS Foundation Wales for providing the money that has paid for most of the equipment essential to make Minty2 drive (but not the survey equipment) and for my time in getting it all to work.

We built Minty1 in 2010, it was just a radio controlled boat based on an optimist dinghy carrying a Swath Bathymetry high resolution SONAR, a long range 3D laser scanner and a PC. Minty1 went out to Greenland and performed several successful surveys, but suffered from four major problems. Firstly remote controlling Minty from some distance away (and the top of the mast of a yacht) proved quite difficult, so this was the key reason that Minty2 needed to be autonomous. Secondly the Optimist hull didn’t interact very well with the ice and often accumulated ice underneath it, to overcome this Minty2 uses a custom designed hull built with ice interaction in mind. Thirdly recharging the boat’s batteries required winching it onto the deck of the support yacht and connecting it to the yacht’s power systems. To solve this problem Minty2 has an onboard petrol generator allowing it to recharge while tied up alongside the yacht instead of being lifted out of the water. Unfortunately the generator creates too much noise to use it while surveying so we have to rely on batteries and electric motors for that. Finally Minty1 wasn’t fast enough to fight some currents and head winds or to push ice out of the way and a more powerful motor would be useful. Minty1 was propelled by a single Minn Kota Riptide RT45 electric outboard motor and steering was provided by rotating this with a Linak LA12 linear actuator. To increase performance Minty2 has two of the Minn Kotas and drives by altering their relative speeds instead of using a rudder.

Minty2 needs to be shipped to Greenland in late May, so development must be quite rapid. We have just received the hull plans from the US Naval Academy who have designed the hull and have spent the last couple of months deciding on and ordering the control system electronics. So basically we now have all the bits we need to make Minty work, its “just” a matter of putting it all together and testing it over the next 2-3 months. To speed things up we are trying to use hardware we’ve successfully used before on other boats and to reuse code. We’re using the same compass the Furuno PG-500 which we used initially on Pinta (but removed for the Microtransat due to its cost) and on BeagleB. The control system code will be based on BeagleB and Pinta’s code as well, but with some changes to remove the sail positioning and tacking logic and deal with a differential motor drive instead of a rudder. For the computer system we’ve chosen a FitPC, which is a full blown PC with a processor comparable to those in most netbooks, we’ll be running Ubuntu 10.04 on it as this is the same OS most of our lab PCs and field laptops are currently running. Motor control will be via a Pololu USB servo controller connected to a pair of Deventech MD03 motor controllers. We’re also going to use a Labjack U6 to provide high precision analogue to digital conversion to read from a series of current and voltage sensors that we’ll use to monitor the battery levels. A long range radio modem will be used for sending telemetry data and receiving radio control signals. A second FitPC will run Windows 7 for the SONAR and Laser Scanner control. A Linksys WRT54GL wifi router will connect up both PCs to some high gain antennas that will allow us to login to either PC from several kilometres away. To make remote controlling it at a distance easier we’ve also bought a GoPro camera to mount on the bow and a 5.8 GHz video transmitter so we can have live video from Minty’s point of view.

Setting up the FitPC

The plans for Minty2 printed out onto A1 paper. These will traced onto plywood to form the bulkheads for the hull.

The almost completed hull and in the background is Minty1 with two of the Minn Kota outboard motors on her deck. The yellow thing on the floor next to her is the generator we’ll be using for Mitny2.