我提議一個新的Robot Competition
目標:活最久的(24Hr Competition)
工作:1.找食物 2.休息 3.避免被吃掉 4.可以攻擊別的Robot (Advance)
食物:一束Laser Beam會不定時出現在特定的區域 每個Robot背上都有一個LaserBeam Detector可以記錄你的覓食時間 你無法再當場覓食 必須到你的巢穴才能把食物算吃下去
休息:每個Robot都使用相同的電池 只有你的食物吃超過一個數量才能換成電池
避免被吃掉:不定時會有另一束Laser Beam會攻擊你 只要被攻擊的時間和次數超過一個量你就算死了
攻擊別的Robot:任何Robot都可以把其他的Robot推開讓他不能覓食 佔據他的巢穴
我覺得這樣才能讓Robot進步到下的Stage不然現在都只是Focus在機械控制上
I was a judge on one of the local Robot competition for the SouthEastern US colleges a couple of months ago. There is a sloped terrian and blocks (weight with different color and RF ID tags) to be picked up. The goal is to collect as many blocks with the highest score and fend off the competitions. Unfortunately, even the pea gravel surface can get several team stuck within minutes. So I think intelligent robotics with a lot of sensors is still elusive for college level. The real life is much harder than computer simulation as one of the competitor said to me ;)
Posted by: Fred | July 01, 2008 at 12:35 PM
You got the point. The terrain navigation problem is not the core problem. My game is play on a flat surface. We solve the 1st Hunt and Run problem then we move on to more complex terrain navigation issues.
I see a lot of game focus on these problems but actually that kind of problem is more mechanical centric then AI centric.
Posted by: Tze-Chien Chu | July 01, 2008 at 07:26 PM
All living thing can do Hunt&Run behvior before evolve to complex life form.
Posted by: Tze-Chien Chu | July 01, 2008 at 07:40 PM