2009
09.30

When we first started thinking about the design of our creature we took inspiration from many sources, mainly because to our knowledge mammals and reptiles do not have six legs and insects do not usually grow to 50cm in length. We were exploring the idea of the more unusual mammals such as the pangolin and the porcupine and then developing the concept further with the materials and techniques available to us.

Initially we decided to attempt to make a hexapod robot using 18 servos to achieve three degrees of rotation on each leg as well as one or two for the head and tail, to achieve this in a cost effective fashion we decided to use 9gram HTX900 servos which had remarkable strength and speed for their size. The fabrication was going to be done in 3mm acrylic that would be lasercut at AUT. In order to control all these servos we needed an Arduino Mega because of its ability to control more than 20 servos at once. To achieve full wireless operation we intended to use a battery pack, Xbee modems and a wireless camera to control the robot remotely.

After making the first lasercut and assembling we realised there were some things that needed changing, first of all we had to change the screw spacing on the servo template as it was .5mm out. The legs were also mounted far to close together for walking, so I spaced them out on the solidworks file so that four of the legs had ~200 degrees of freedom in the y-axis. In order to test degrees of freedom, etc. I created a solidworks file with circles marking the degrees of freedom between each joint and where they intersected. From this I determined that each step would travel a maximum of about 60mm. This is not bad for a model around 400mm long.

Programming went well, the servos responded fast and processing could be used to provide a serial connection over USB or the Xbee. The intention was to use serial commands to trigger preset movements on servos or sets of servos with the keyboard and x/y mouse values. A series of ‘modes’ would take care of the different poses we used in the final film.

We intended to fabricate the model using cotton wool and stretchy fabric (eg. Lycra) to provide something to add scales onto without damaging the model beneath. We also intended to use two RGB leds as eyes to help convey the mood and intentions of the robot at any point in the video.

Over the break the lasercutter was mostly unavailable so we didn’t get the final cut done until after I went to Wellington in week two. This limited our final development time to the first week after the break which wasn’t ideal. However we made good progress, creating a power rail system so that the current drawn by 18 servos didnt overwhelm the Arduino Mega’s voltage regulator chip. On thursday we had the code done and were ready to fabricate the model and film the final video. Unfortunately at this point the Arduino Mega shorted out, thereby ending three weeks of progress and rendering us completely unable to control 18 servos at once (most Arduinos can only control 12). Over the course of the afternoon things got worse, with first the Bluetooth Arduino and then the Lilypad causing their own technical issues.

With this in mind we decided to do what we could in the short time available, Hanan made an adjusted final render using Maya and fabricated the model then I filmed it using sets I scouted last year while exploring the area around the studio. Because the model wasn’t functioning we resorted to stop motion techniques to produce an acceptable result. If backstory for this film is required it can be said the story involves the creature escaping a lab after being discovered by scientists exploring Peru and it sneaking through the ventilation/piping facilities.

The final video seeks to produce heavy atmosphere using sound, lighting and camera techniques. While it doesn’t really show the weeks of hard work we put into programming and designing the model so that it could be remotely controlled and fluid in its motions we have included all the documentation photos/videos so that our progress up until the Arduino Mega failed are made clear.

Comments are closed.