Kernel Recompilation

Recompiled the kernel tonight (this morning) to get rid of kernel module errors. It should be now possible to boot the 'box and load up all RTD modules with correct (i.e. dedicated) IRQ channels. This came at the expense of the two on-board serial ports, however all functionality should (hopefully) be accomplished by the RTD boards and the USB 2.0 expansion board.

Wildfire Conversion

Spent the afternoon converting all of the robot models, manufacturing assemblies et cetera from Wildfire 2.0 to Wildfire 3.o.

Hopefully the improved features of Wildfire 3.0 will enable the creation of an even better 'bot.

Wireless

Removed the PC104 stack from its LAN umbilical, and it is now operating in wireless mode. A wireless network has been set up, facilitated by the use of a LINKSYS WAP54G access point.

Use of the access point has significantly improved the reliability of the network. Previously in the 3-node Ad-Hoc configuration, nodes had a tendency of fallling off the network quite regularly.

The included photo shows a departmental 'scope attached to the PWM board, showing the PWM signal from one of the boards 3 independent triple-channel PWM chips.

PC104 development

Here is the PC104 stack resting in its cradle. Also visible in the bottom region of the photograph are the webcams (Logitech Quickcam Pro 4000's) that will be used in the platform.

It seem that all the modules (USB, PCMCIA, RTD) have been loaded and are all generating correct interruots with no handling errors.

Electronics Cradle

Shown here is a cradle that was manufactured today for the electronics stack. It was becoming a serious mission disentangling the stack from the robot and setting it up every time some coding needed to be done.

The cradle has a normal computer power supply and fan mounts to ensure consistant cooling. It will hopefully allow the stack to be easily switched from bench to robot and speed up development time.

.mex Files

Successfully managed to use the system gcc compiler to compile a sample DM6420 program, including all relevant static libraries. This means that the functionality of the static library can be accessed from within the Matlab environment, allowing for quicker prototyping & behavioural changes.

Unfortunately, it means that Matlab must still be installed on the system and this means the hard-drive stays. When all (most) of the bugs in the system have been eliminated, it may be possible to reduce all code to C.