Biped Model Based on Human Gait Pattern Parameters for Sagittal Plane Movement

Biped Model Based on Human Gait Pattern Parameters for Sagittal Plane Movement

1. INTRODUCTION

   • Human Push recovery capability is a learning process which human being learns through a continuous evolution process. Our two most sophisticated parts of central nervous system, right lobe and left lobe are used to control and recover from particular push. So it is one of the most frequent and quick decision taken by humans during normal walking.
     • adaptability, obstacle negotiation capability and stairs climbing capability we are using biped model. We as human prefer to learn through continuous learning by negotiating with environment and by effectively using different strategy to impart human learning robot so we are using fuzzy logic based controller which counters impreciseness of environment.
       • The major benefit of using fuzzy model
         • is to achieve low cost and more intuitive model. Fuzzy model as a tool is easily understandable by human i.e. it has sort of knowledge representation for better human understandability.

2. PERFORMANCE MEASURES

   • While dynamic models used a LIPM, which assume a point mass biped with fixed CoM value. We validate our method on the HOAP-2[2], which have 25 DoF with multibody mass better flexibility to avoid obstacle and less singularity issue as compared to LIPM model

3. HUMANOID OPEN ARCHITECTURE PLATFORM ROBOT

   • The HOAP-2 is one of the biped robots of HOPE series, developed by Fujitsu for commercial use. It is very much popular between research community due to resemblance with the human with more DoF and mobility. The DoF which is actually defining the number of parameter to defined the all joint configuration which help us to analyze the problem of push recovery.

4. EXPERIMENTAL RESULT

   • The experimentally result using in house development proved that humans have preference for left and right legs analogous to left and right handedness[14]. . It gives great insight to develop a reactive controller for a bipedal humanoid robot, to make robot model different from the architecture of plain robot, which robot is helpful in many real world interface like swimming robot or a flying robot. So, It is desirable to use the mass instead of the density so we set density =-1 mass =0.5 density in kilogram/cubic meter and mass in kilometer. When we specified density it ignored mass. The below mentioned table gives details about the various parameters used in our model
     • • Denavit–Hartenberg parameter or D-H parameter are well known and highly useful concept in robotics. With the help of D-H parameter, Right ankle and left ankle of left handed person shows little involvement in balancing. This experiment shows a very different result that one side of human body is more active during push recovery i.e. for left hander person, his right leg joints are more active in push recovery pattern and vice versa. Vital role has played by knee in balancing.

5. OUR APPROACH

   • The objective of this paper is to describe the model which can perform the human like Gait cycle and analyze the behavior using hierarchical fuzzy algorithm, control the balance of lateral plane movements of humanoid robots. We taken consideration of the HOAP2 is model because it is resemble to human
     • organized our methodology in the following steps:
       • Step1
         • To simulate the liner inverted pendulum equivalent biped model i.e. HOAP2 similar model in Webots. And calibrate the all the part and joint with their respective physical parameter like degree of freedom, mass, centre of mass, height, length of body part etc.
       • Step2
         • To compute the control torques for the joints in a bipedal humanoid, change in angle of different joint, gait pattern taken from HMCD is of human and finally calculate Inverse kinematic and D-H parameter of model.
       • Step3
         • Apply all the above calculated parameter to our model i.e. Halo and make it to follow the human gait cycle based on real time data captured through HMCD.
       • Step4
         • To gather the sufficient amount of data by our model to impart the learning technique.
           • he experimental data collected through in house developed HMCD device used in development of software based model. Then to make it follow the human gait cycle, Inverse kinematic have been solved for its legs, the model also has been programmed to be visible dynamically. To make advance we involved the two infrared sensors on the eyes of biped to identify obstacle in the way. To make it free from intensive computation mechanism and impart learning we proposed the hierarchical fuzzy system .gait cycle style in which a human being walks.
             • Gait cycle starts where one foot leaves the ground and end it again touches the ground [12]. The control torques can be computed for the joints in a bipedal humanoid using the following equations:
               • • For normal walking pattern, torques can be generated as in eqn.(1) which is sum of inertial torque (M), Centrifugal and Coriolis torque(C) and gravitational torque (G) together with some frictional torques which has been neglected. When external force is applied on human, these ideal walking patterns of different joint torques get disturbed. To regain original torques, control torques required to be applied.
       • Step5
         • To develop a hierarchical fuzzy module based on the data gather through indigenous developed suit HMCD and associate same computing on Webots controller to guide the our model similar to HOAP2 to follow the gait cycle and predict the output angle and velocity on the basis of input force and direction around our rule defined in paper.

6. JUAN PABLO GARCÍA CALDERÓN DANNY MONTAÑES MARTINES

Imágenes y archivos del mapa