Older blog entries for c6jones720 (starting at number 14)

Hi, I saw this interesting documentary/drama the other day on BBC1 (UK)about the prospect of a terrorist organisation exploding a Dirty Bomb in metropolitan london. At the end, the authorities cordoned off and abandoned a large section of London for a 30 year cooldown period. As soon as I saw that the first thing that sprung to mind mind was Wow! what an absolutely perfect opportunity to use robots for clearing and redeveloping a dangerous area! I should'nt say this, but that is one niche in robotics that needs to be exploited more.

Other thing, - I've just been reading some different books on neuroscience and artificial intelligence. I can recommend the books "Growing up with Lucy" by Steve Grand and "Conciousness" written by two neurologists. The Lucy book is basically a diary of a guy who builds a basic android its got all sorts of stuff about neural nets and vision, although his writng style is not so scientific. The conciousness book fills in some of the gaps in understanding minds. Its helped to explain to me why everytime we write an AI program it ends up being something somebody else already did and is never quite as good as we thought. I think were missing a couple of things in our software, but this is all in the book...

I was in a bookshop the other day and came across a book called AI game programmers wisdom volume 2. Its a big thick book with about 800 or so pages and it is quite expensive ($69.99 US or £45.00 UK). I got it and I'd say it was worth every penny. It is primarily about video game AI, but it is very relevant to robotics because there are many many similarities between game agents and physical robots in the real world. Its a great book and deals with all sorts of stuff like A* pathfinding (mobile navigation), PID control, Fuzzy and Neural nets...

If you get a chance to read it, it is definately worth it..

I just settled down in my new job as a Software Engineer (not much robotics or electronics but hey its not the end of the world..)Everythings running smooth, all the equipment is working and Im just starting to have ideas on how to produce a cheaper commercial version of the US Packbot. Then it rains! Our entire factory and my working area were flooded! Im gonna have to put my Packbot ideas to one side until we get dry....

I've got mixed feelings about my new robot. I've just moved from Plymouth (UK) to Bristol (UK) and I think one of my robots arms has broken in the process. I might have to replace the main shoulder motor. Wierd, the robot seems a lot weaker at liting now than it was in Plymouth, there must be the Plymouth electricity must be better than Bristol electricity or something. I've got half of the robot reflexes wired up now so now the left arm can respond if it hits something or if I prod it.

Hi guys, I sat down and read both "Brainchildren" by Dennet and "Minds Brains and Machines" by Geoffrey Brown. These books caused me to have a little think about the whole concept of AI. It seems to me that in fact when we actually look at the English language definitions of the words we associate with intelligent behaviour I seems clear that computers and robots employing computers are capable of many (low level) intelligent actions. From my experience it seems to me that there is always this brick wall we come to when we are designing an intelligent system. Im starting to think that it arrises because we focus too much on small areas and forget about the artificial mind entity as a whole. In short I think that nearly all aspects of AI are achieveable at some basic level but true "thought" as defined is difficult to achieve because although computers and robots have needs, they do not have wants or desires. If we can give them desires and wants then I beleive AI will benefit significantly... - Cheers Chris

I wasnt really sure how to guage the reliability of my animate vision system so I wrote a program that would cause the robot to move its head and point with its left arm to where it saw motion. Surprisingly it works with about 80% reliability its kind of odd to see a robot looking and pointing at you when you move about...

One more exam left - and thats the end of my robotics degree I won't have to feel guilty about spending time working on this robot then :-)

My new clas 6 robot is coming along fine.. I've used a combination of serial servo controller and home made control device to make the actuators work. I think it is okay.. The grippers can pick up light objects under 100g and visual gripper tracking works in at least 4 of 9 areas....

I've built a new version of my MR_Head robot (version 3.0) this one features a monochrome vision system and it works! The head is capable of tracking a person (or large object) around a room using only binarised images. I've also got a new metal detector for my car bomb inspection robot. i'll have to integrate this later on (after my exams)

I should be revising for my degree finals. I kind of am but I can't help being a compulsive roboticist!!! Im just halfway through a really simple collision avoidance and detection routine using vision. Im going to use this on a remake of one of my robot heads (Mr_Head 3.0)

Well despite the fact that my mates girlfriend dropped my robot from a great height and despite the fact that I couldnt get the camera pan tilt mechanism quite right, the academics really liked it!

The car bomb inspection robot I was working is finished and was put on display at tyhe University of Plymouth (England). To my surprise they gave me the 2004 award for engineering excellence - can you believe it?!

5 older entries...

X
Share this page