Name: Robert Goldsmith
Member since: 2002-10-05 20:34:13
Last Login: 2003-01-28 02:05:32
I am a Ph.D. student at the university of Birmingham, UK (School of Computer Science) and my actual Ph.D. research is into Evolutionary Hardware (the use of evolutionary techniques such as survival of the fittest to improve the design of hardware). For this research, I have needed to build my own technology platform from scratch. As well as being part of my research, robotics is a subject I am fascinated by in general and I am constantly learning new skills and thinking up new ideas for projects in my spare time. I have skills mostly in software and microcontroller systems but have rapidly learned hardware and electronics skills as need for my projects. To date, as my skills have been software, most of my designs have used kits or lego for the hardware and a verity of controllers for the software systems. As my Ph.D. forces me to learn hardware skills, I am moving into designing entire systems from scratch. I find my background and experience in cooperative systems, distributed architectures and evolutionary systems gives me quite a different outlook on robotics systems to many others :) All my interests in robotics are focused around mobile systems rather than industrial instillations.
Recent blog entries by nevyn
Gosh, quite a while since my last diary entry! Well, I've been busy but, unfortunately not with much robotics :( Let me see. Well, starting with my work on new 'toys' for the third year undergraduate course in robotics at my uni, most of it works, I am pleased to say :) We have both a Windows pocketpc 2002 pda and an intel compatible linux based pc104 board talking to my new Handyboard firmware but only via serial. We also have my firmware talking to my Mac (OS X) via USB so the firmware USB client drivers are working nicely (hacked together in a mix of C and ASM, in case you wondered!). The problem is linux and pocketpc usb drivers. Has anyone got any idea how to write them! You would have thought it would be easy but no: the documentation is non-existent and the example code usually totals the linux kernal. Hmm, fun. I've got a linux driver sending to the firmware but nothing is coming back atm.
On the same vein, we now have batteries and enclosures ordered (as well as 17 pc104 boards) so, as long as we stick to serial for the moment, we are doing fine. I'm sure I'll crack the usb driver problem somehow. Everything anyone would need to build the handyboard adapter (for usb), the new firmware source code (with comments and docs), the pc104 details, the custom linux install (including a java micro vm) and the interface low level and high level api's should all be available about end of january (once we've tested and written it!). Even if you are not interested in the complete project, little bits (like how to get a 2MHz HC11 to talk to a pc via USB) could be useful to you :) I will also be posting a very largely modified version of a windows HC11 c compiler I've revamped to work with the Handyboard interrupt vectors, ports and memory addressing. I've also got source for things like display drivers for the lcd display etc.
Next to my research and, overall, this is also going in the right direction - just very slowly! I now have a prototype of the USB to packet radio hub built and I'm currently programming it. I could really do with a good C compiler for HC908 code but they don't seem to exist (well, for less than a few hundred pounds). I tried to revamp the c compiler I used for the hc11 code but with little success to date. The problem is the hc908 don't have much ram but lots of flash (unlike the handyboard where it is all ram). Looks like it's asm all the way for now. I've also solved a few battery problems (found a very useful very-low-drop voltage regulators) and some motor interference problems (also solved using the regulators) and found some suitable batteries.
Lastly, the robot wars entry has ground to a halt. Basically the constant problems of time and finances have put a freeze on further development. Never mind - I am lucky enough to have two other projects to play with!
A final thanks, before I finish, to everyone who emailed about my last list of questions / problems - you were all very useful!
Hi again all :)
maybe I have missed a link somewhere in this site but I can't find any place to post simple questions. Myself and a couple of friends are looking to build an entry to Robot Wars (UK version) and this means working with technologies we have little experience with - such as high power controllers and batteries etc. If we can collect enough information and money and our tests go well (in the physics dept. at the uni ...) we will have a go and I will likely put up a project on this site. However, in the mean time, I would very much appreciate any comments anyone has on the following:
Batteries - most people say lead acid or gell cell but some comments about high capacity NiCad or NiMH are starting to surface. Any experience anyone?
Digital radio control on the 40MHz band - I'm used to radio control using the 433MHz band and packet switched radio systems. Can anyone point me towards such systems on the 40MHz band (I need a system that can work with changable crystals as well...)
Digital compass - I have spotted a couple. Has anyone had any experience of these? Esp. the accuracy around large metal objects (such as other robots!).
Oh, and can anyone suggest a better way to ask these questions than in a diary entry?!
you can contact me at firstname.lastname@example.org - Thanks :)
Well, Hi :)
As this is my first diary entry, maybe I should mention what I have been doing 'til this point.
First my research. This is into evolutionary hardware - specifically the placement of sensors on a mobile robot. A collection of robots (controlled via wireless by a computer) all try to find the most food they can while moving uses up energy. If two robots bump into each other (and if conditions such as their energy levels are acceptable), they may mate. Potential offspring from these matings are added to a pool and randomly selected when a physical robot becomes available (an old robot dies of lack of food, being poisoned by eating the wrong stuff or of old age, for example). The idea is that the robots better at finding food will mate more often and that the robots better at finding food consistantly do so because they have a better setup of sensors. Natural Selection :)
As for my other stuff, I have built a number of small robots using the MIT Handyboard controller (although I use ASM, not IC) and lego (fantastic stuff for prototyping), including a simple walking robot and a turtle (named Emily) I entered for a quite recent drawing robot competition in the elc eng dept. at my uni (uni of birmingham, uk). I came second (well, Emily was built of lego, unlike many of the other entries...).
I am also involved in developing new technologies for the Intelligent Robotics course run by my dept (Computer Science) for undergraduate students. Recent projects have included writing a custom Firmware for the MIT Handyboard, writing and building USB controller / expansion stuff for the Handyboard and linking the Handyboard up to pc104 and pda systems with webcams. Information on all of this, including pcb designs, the firmware and usb drivers etc. should be available in the near future (not sure where but I'll post when I know)
I think that's enough for now :)
nevyn certified others as follows:
Others have certified nevyn as follows:
- earlwb certified nevyn as Journeyer
- robotdojo certified nevyn as Journeyer
- jbm certified nevyn as Journeyer
- ROB.T. certified nevyn as Journeyer
- nitro certified nevyn as Journeyer
[ Certification disabled because you're not logged in. ]