Introduction:
While at the University of New South Wales, I am taking a course in Experimental Robotics. It is project-based: the initial introductory task is to make a small Rug Warrior floor-roving robot follow a white line on the ground while avoiding becoming stuck on any obstacles it encounters, and after that there is another project of your own choice. Together with Raymond Sheh I chose to attempt to build an interesting design of mobile robot platform I had found in Hans Moravec's book "Mind Children". He attributed it to a project Hitachi was working on in the early 1980s, but I have found no other reference to it, either on the Web or elsewhere.
The robot had five telescopic legs terminating in powered steerable wheels, all of which could be controlled independently. The aim of this design was to be able to travel rapidly over smooth ground that the wheels could traverse, but also to be able to climb over obstacles by lifting up each leg in turn, driving forward until the leg was over the obstacle and lowering the leg until it touched the surface. The legs could also be used as "active suspension" when driving over rough terrain.

A computer-generated picture of the whole robot, standing on flat ground.

A computer-generated picture of the robot climbing off a stack of blocks - the legs are on 3 different levels.
Our version:
This is quite a complex robot platform to attempt to make, especially for a time- and resource-limited project. Raymond and I therefore decided (for the time being, at least) not to build steering capability into the legs and make the robot able to skid-steer if necessary. The electronic organisation will be modular, with controllers at different levels. The name is, obviously, a corruption of "pentagon" - we invented it ourselves as no name for the design was mentioned in Moravec's book.
Current progress:
(29/4/2003)We have written and presented the mid-semester report (Microsoft Word format), and populated the final-version legs with electronic components.
(4/4/2003) We have designed and made all the circuit boards for the individual final-version legs (going through two revisions - the rev 2 boards are silver plated, which the original rev 1 board is not) and are preparing to populate them, both electronically and mechanically.
(20/3/2003) We have built a fully-functioning prototype leg module and are working on the designs, both electronic and mechanical, for the finished robot.
(14/3/2003) We are working on making 3d models of a single leg module, which we will build as a first step in the construction of the robot.