Well, the deadline for the project has been reached. There are still some things yet do be done, however:

• Electronics - Some work on the power supplies and main CPU.
• Hardware - New gearbox cases for the knee joints for added strength
• Software - Finish the SimMechanics model

The GA Server and client have been tested on the Broom Balancing Problem, and have successfully created a neural network capable of solving this problem.

Although the project is officially over, work will continue, but it will have to wait until after exams!

Three conference papers have been written describing the various aspects of the robot's design. These will be presented in SCARP 2003. The papers will be available for download from the project page.

Coding of the modelling software is continuing. A program to convert a 3D model of the robot in VRML format to a SimMechanics format model is nearing completion. It takes each solid structure in the robot (for example, the femur) and calculates the overall inertia tensor, centre of gravity and mass. This is then output as a SimMechanics Body structure.

It has been decided that processing will be done on a PC rather than on the PIC, as this will be easier to write. Data will be transferred between the two via an RS-232 link.

If any time is available later, processing could be moved back to the PIC

Since shortly before Christmas the construction of the robot has been at full speed. Much has now been completed including one complete leg and half of the second, plus all design work on computer has been completed. All that is left now is to put the feet and body together then install and test the compliance system. The construction of the legs of the robot is the larger of all these tasks.

SimMechanics has been made available, and so work is now beginning in creating a model of the robot. This will eventually be used to perform fitness evaulations on the GA clients.

The project was given a just-satisfactory mark in a very brief review with Paul Sharkey today. Unfortunately this was due to the current lack of hardware. The software aspect is coming along well, however, and the hardware is well planned and so progress should be rapid once physical construction is started.

Much time has so far been devoted to studying the structure of birds to assertain what aspects of their anatomy are important to benifit the robot's stability. This research is now complete and has also looked at the Therapod group of dinosaurs for more varied structures which could exhibit better walking gaits for the robot. The end result is a complete set of dimensions for the robot. Currently Torque analysis is being carried out on the designs of the robot to better select appropriate motors and gearboxes.

Construction of the Chicken Walker has begun. Graz is starting by testing the mechanical properties of various metals so a suitable one may be chosen for the joint compliance unit springs. Once chosen, fabrication of a single compliance unit will begin so it may be tested and mathematically modelled.

A new report, "Implementation of Genetic Algorithm", has been written and uploaded. This describes in detail how the GA will be written, and will be used as a basis for the code.

Coding is continuing, with the GA code now partially complete