2011 4 th International Conference on Mechatronics (ICOM‟11), 17-19 May 2011, Kuala Lumpur, Malaysia 978-1-61284-437-4/11/$26.00 ©2011 IEEE Geometrical Analysis on BIOLOID Humanoid System Standing on Single Leg Md. Akhtaruzzaman Department of Mechatronics Engineering, Kulliyyah of Engineering, International Islamic University Malaysia (IIUM), 53100 Kuala Lumpur, Malaysia. akhter900@yahoo.com Amir A. Shafie Department of Mechatronics Engineering, Kulliyyah of Engineering, International Islamic University Malaysia (IIUM), 53100 Kuala Lumpur, Malaysia aashafie@iium.edu.my Abstract — In near future humanoid robots will not only be able to socialize with the human being but will also be able to replace him even in the irksome and dangerous tasks, ranging from rescuing situations to interplanetary exploration. Nowadays many researchers are engaged on this field to make the humanoid more adaptable, intelligent and representable to the dynamic environment. Designing a suitable and efficient gait for the Biped Intelligent Machine (BIM) is a complex task. In this paper a geometrical analysis is presented to identify the movements and positions of the Center of Gravity (CoG) of a humanoid system while it balancing itself in the Single Support (SS) mode. SS mode while standing on single leg is a very critical job for a bipedal system. The paper also exemplifies the results based on the proposed Geometrical Analysis Technique (GAT) which is applied on the BIOLOID humanoid system. Keywords- BIOLOID Humanoid System, Humanoid Robot, Android, Biped Walking Robot, Geometrical Analysis, BIM, CoG, GAT. I. INTRODUCTION A human like autonomous robot which is capable to adapt itself with the changing of its environment and continue to reach its goal is considered as Humanoid Robot. These characteristics differs the Android from the other kind of robots. In recent years there has been much progress in the development of Humanoid and still there are a lot of scopes in this field. Though humanoids are neither intelligent enough nor autonomous, they currently represented as one of the mankind‟s splendid accomplishments. It is the single greatest attempt to produce an artificial, sentient being both in male and female like structures. In the recent years manufacturers are making various types of humanoid platforms which are more attainable to the general public. In present, the most advanced humanoid systems are devised with voice, vision and gesture recognition strategies. Dynamic advanced locomotion system makes the robots able to walk, run, climb stairs and even avoid static and dynamic obstacles while navigating. Basically locomotion is the main characteristic studied in humanoid robotics and only after achieving the natural walking and locomotion control of the humanoid systems in the natural environment, they will learn how to interact with the social surroundings using its artificial intelligence. In this paper the GAT is analyzed, described and implemented on a bipedal humanoid robot kit only for standing on a single leg during Single Support (SS) mode. The attempt of this simple project is considered as the initial step to establish a natural, human like gait for the BIM. The GAT focuses on changing characteristics of the CoG point of the bipedal system. The actuator movement patterns are also resulted and analyzed in the result section. II. PRELIMINARIES Bipedal humanoid systems have a better mobility to move on the rough terrains, steep stairs, inclined surface and obstacle environments. So, walking pattern synthesis is one of the major topics of interest in the field of humanoid robot research. Suitable gaits with well balancing control are also the important concern of the researchers. Various algorithms and techniques are established in these recent years to achieve the natural control on humanoid systems for its stable gait. CoG positioning technique, Zero Moment Point (ZMP), Contact Wrench Sum (CWS), Passive interaction of gravity based control, Balance control based on reaction force between the foot and ground, etc. are some of the popular techniques [1]. ZMP and CWS are the most common technique which are implemented in some of the most famous humanoid systems like QRIO from Sony, ASIMO from Honda, HRP-2 and HRP- 3 from Kawada to allow them walking on uneven terrain and inclined path, running and even climbing stairs [1] [3]. The main principle of these two techniques is to cancel the total inertial force acting on the humanoid with the floor reaction force. Jeong-Jung Kim and Ju-Jang Lee proposed a gait adaptation method for biped robot in various terrains where Central Pattern Generator (CPG) technique are adapted for creating desired joint angle of a biped robot [6]. For learning mechanism, Genetic Algorithm (GS) and Neural Network (NN) were used for this project experiment. A dynamic balancing strategy which is opposite of the ZMP technique, is implemented recently to the Dexter humanoid robot from Anybots. The system does not need preprogrammed foot prints, able to walk like human and also able to jump. Passive Dynamic Walking (PDW) is an interesting technique for a biped system which does not require any external control or any energy input. The system movement is governed by the natural swinging of the legs [10]. Sang-Ho Hyon, Jun Morimoto and Mitsuo Kawato integrated Artificial Central