Evolutionary Strategies Combined With Novel Binary Hill Climbing Used for Online Walking Pattern Generation in Two Legged Robot

Abstract

Evolutionary algorithms (EA) has often been proposed as a method for designing systems for real-world applications (Higuchi et al., 1999). Developing effective gaits for bipedal robots is a difficult task that requires optimization of many parameters in a highly irregular, multidimensional space. In recent years biologically inspired computation methods have been employed by several authors. For instance, Hornby et al. used genetic algorithms (GA) to generate robust gaits on the Aibo quadruped robot (Hornby et al., 2000). GA applied to bipedal locomotion was also proposed by Arakawa and Fukuda (Arakawa & Fukuda, 1996) who made a GA based on energy optimization in order to generate a natural, human-like bipedal gait. One of the main objections to applying EA’s in the search for gaits is the time consuming characteristic of these techniques due to the large fitness search space that is normally present. For this reason most approaches have been based on offline and simulator based searches. To reduce the time spent searching large search spaces with EA, various techniques for speeding up the algorithms have been presented. With the increased complexity evolution schema introduced by Torresen (Torresen, 1998), Torresen has shown how to increase the search speed by using a divide and conquer approach, by dividing the problem into subtasks in a character recognition system. Haddow and Tufte have also done experiments with reducing the genotype representation (Haddow & Tufte, 2000). Kalganova (Kalaganova, 2000) has shown how to increase the search speed by evolving incremental and bidirectional to achieve an overall complex behavior both for the complex system to the sub-system, and from the sub-system to the complex system. For an exhaustive description of other approaches readers may refer to Cantu-Paz (Cantu-Paz, 1998). The robot presented in this paper is a two-legged biped with binary operated pneumatic cylinders. The search space in our experiments was set up to describe the forward speed of the robot given the different gaits, and the goal was to find the most efficient gait with respect to speed. To enable efficient gaits the search space needed to be quite large as the accuracy of the pause lengths between the different leg positions is outmost critical, especially for gaits dominated by jumping movements. The focus has not been on evolving a balancing system as there have been no other sensory feedback than the forward position of the robot. The main goal for our work was to find a search algorithm fast enough to enable