Reynolds C.W. (1994) Evolution of Obstacle Avoidance Behavior, in: Advances in Genetic Programming, K. Kinnear, Jr. (ed.), MIT Press, Cambridge, MA  Home/Search   Document Not in Database   Summary   Related Articles   Check

....has been reported previously in a number of variants. Robotic controllers have, for instance, been evolved using dynamic recurrent neural nets [4] 12] Several experiments have also been performed where a controller program has been evolved directly through genetic programming [11] 15] [24]. Previous experiments with genetic programming and robotic control, however, have been performed with a simulated robot and a simulated environment. In such a set up, the environment and the robot can be reset easily into an initial state in order to ensure that each individual in the population ....

....system learning form past experiences. At the outset, the population of programs is initialized with random content. Tournaments are used for the competitive selection of individuals which are allowed to produce offspring. The GP system with its simple steady state tournament selection algorithm [24], 27] has the following execution cycle: 1. Select four arbitrary programs from the population. 2. For each of the programs calculate fitness. 3. Make two copies (offspring) of the two individuals with highest fitness and subject the copies to crossover and mutation 4. Replace the two individuals ....

Reynolds C.W. (1994). Evolution of Obstacle Avoidance Behavior. In Advances in Genetic Programming, K. Kinnear, Jr. (ed.), MIT Press, Cambridge, MA.

....through one specific obstacle course. They will not even work in the same obstacle course if any of the vehicle s initial conditions (position or orientation) is slightly perturbed. These brittle controllers are a bit like a house of cards which stands as long as absolutely nothing changes (Reynolds, 1994). Other experiments display a similar brittleness in behavior. One way to improve the simulation and to evolve more robust controllers is to add noise to the environment. This reduces the brittleness but does not eliminate it (Reynolds, 1994) Further examples of GP and simulated control include ....

....which stands as long as absolutely nothing changes (Reynolds, 1994) Other experiments display a similar brittleness in behavior. One way to improve the simulation and to evolve more robust controllers is to add noise to the environment. This reduces the brittleness but does not eliminate it (Reynolds, 1994). Further examples of GP and simulated control include (Koza, 1992; Atkin Cohen, 1994; Fraser Rush, 1994; Handley, 1994; Sims 1994; Spencer, 1994; Teller, 1994) More serious than the brittleness of controllers is the problems of moving the experiments from a simulated robot to a real one. The ....

[Article contains additional citation context not shown here]

Reynolds, C.W. (1994). Evolution of Obstacle Avoidance Behavior. In K. Kinnear (Ed.), Advances in Genetic Programming. MIT Press, Cambridge, MA, 221 --- 241.

....controllers have, for instance, been evolved using dynamic recurrent neural nets [5, 12] Genetic Algorithms have been used in [9] for generating wall following behavior. Several experiments have also been performed where a controller program has been evolved directly through genetic programming [11, 17, 28]. We, too, have reported earlier on our first experiments using GP to control a real robot that has been trained in real time with actual sensor values [23, 25] In a real environment our system had to evolve robust controllers because noise was present everywhere and the number of real life ....

C.W. Reynolds (1994) Evolution of Obstacle Avoidance Behavior, in: Advances in Genetic Programming., K. Kinnear (ed.) MIT Press, Cambridge, MA.

....has been reported previously in a number of variants. Robotic controllers have, for instance, been evolved using dynamic recurrent neural nets [4] 12] Several experiments have also been performed where a controller program has been evolved directly through genetic programming [15] [23], 11] Previous experiments with genetic programming and robotic control, however, have been performed with a simulated robot and a simulated environment. In such a set up, the environment and the robot can be reset easily to an initial state in order to ensure that each individual in the ....

....form past experiences. The population of programs is initialized with random content at the outset. A simple tournament is used for the competitive selection of individuals to produce the right distribution of offspring. The GPsystem with its simple steady state tournament selection algorithm [23], 26] has the following execution cycle: 1. Select four arbitrary programs from the population. 2. For each of the programs calculate fitness. 3. Make two copies (offspring) of the two individuals with highest fitness and let the copies be subject to crossover and mutation 4. Replace the two ....

Reynolds C.W. (1994) Evolution of Obstacle Avoidance Behavior, in: Advances in Genetic Programming, K. Kinnear, Jr. (ed.), MIT Press, Cambridge, MA

No context found.

Reynolds C.W. (1994) Evolution of Obstacle Avoidance Behavior, in: Advances in Genetic Programming, K. Kinnear, Jr. (ed.), MIT Press, Cambridge, MA

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC