Online Gait Research Articles

On-Line Gait Adjustment for Humanoid Robot Robust Walking Based on Divergence Component of Motion

As the first step for biped robots to enter the human life, robust walking is a difficult problem to be solved owing to the various algorithms and practical engineering issues being involved. This paper studies the robust walking problem for humanoid robots by using the divergence component of motion (DCM) method based on linear inverted pendulum model. Firstly, we implement a DCM trajectory planning method to simplify the planning process. It calculates the DCM trajectories under the requirements of walking speed and initial state of the system. Then, a DCM feedback controller with anti-disturbance ability is proposed to realize the tracking control of the planned trajectory. Finally, the optimization method and DCM feedback control are integrated into a hybrid optimization controller, which takes into account the step adjustment and the step duration adjustment of the robot. Simulation results demonstrates that the technique can act naturally stable inside an enormous scope of effect unsettling influences, the maximum recoverable impact of a humanoid robots with a mass of 70Kg can reach 85Ns, which is much better than the 20Ns of the existing model-based prediction control method.

Quick foot placement adjustments during gait are less accurate in individuals with focal cerebellar lesions

Online gait corrections are frequently used to restore gait stability and prevent falling. They require shorter response times than voluntary movements which suggests that subcortical pathways contribute to the execution of online gait corrections. To evaluate the potential role of the cerebellum in these pathways we tested the hypotheses that online gait corrections would be less accurate in individuals with focal cerebellar damage than in neurologically intact controls and that this difference would be more pronounced for shorter available response times and for short step gait corrections. We projected virtual stepping stones on an instrumented treadmill while some of the approaching stepping stones were shifted forward or backward, requiring participants to adjust their foot placement. Varying the timing of those shifts allowed us to address the effect of available response time on foot placement error. In agreement with our hypothesis, individuals with focal cerebellar lesions were less accurate in adjusting their foot placement in reaction to suddenly shifted stepping stones than neurologically intact controls. However, the cerebellar lesion group’s foot placement error did not increase more with decreasing available response distance or for short step versus long step adjustments compared to the control group. Furthermore, foot placement error for the non-shifting stepping stones was also larger in the cerebellar lesion group as compared to the control group. Consequently, the reduced ability to accurately adjust foot placement during walking in individuals with focal cerebellar lesions appears to be a general movement control deficit, which could contribute to increased fall risk.

Online Gait Learning for Modular Robots with Arbitrary Shapes and Sizes.

Evolutionary robotics using real hardware is currently restricted to evolving robot controllers, but the technology for evolvable morphologies is advancing quickly. Rapid prototyping (3D printing) and automated assembly are the main enablers of robotic systems where robot offspring can be produced based on a blueprint that specifies the morphologies and the controllers of the parents. This article addresses the problem of gait learning in newborn robots whose morphology is unknown in advance. We investigate a reinforcement learning method and conduct simulation experiments using robot morphologies with different size and complexity. We establish that reinforcement learning does the job well and that it outperforms two alternative algorithms. The experiments also give insights into the online dynamics of gait learning and into the influence of the size, shape, and morphological complexity of the modular robots. These insights can potentially be used to predict the viability of modular robotic organisms before they are constructed.

Robust model predictive control of biped robots with adaptive on-line gait generation

In this paper, an on-line gait control scheme is proposed for the biped robots for walking up and down the stairs. In the proposed strategy, the nonlinear model predictive control approach is used for the trajectory planning and as well as for the control of the robot. The motion of the robot is expressed in the form of a cost function and some constraints that are related to the stable walking of the robot. The main feature of this method is that it does not need any off-line trajectory planning and the walking gait is formulated such that the environmental and stability constraints of the robot are satisfied. This on-line trajectory planning gives the important ability to the robot to adjust its gait lengths. In this way, the robot is able to ascend and descend the stairs without knowing the height and depth of the stairs in advance. In the control algorithm, the Radial-Basis Function (RBF) neural network with on-line training method is used to model the behavior of the robot over the prediction horizon. The stability analysis of the closed-loop system is performed using the Lyapunov method as well as the Poincare map. The proposed method is applied to a 5-DOF biped robot in the sagittal plane. The simulation results show effectiveness of the proposed method.

An online gait generator for quadruped walking using motor primitives

This article presents implementation of an online gait generator on a quadruped robot. Firstly, the design of a quadruped robot is presented. The robot contains four leg modules each of which is constructed by a 2 degrees of freedom (2-DOF) five-bar parallel linkage mechanism. Together with other two rotational DOF, the leg module is able to perform 4-DOF movement. The parallel mechanism of the robot allows all the servos attached on the body frame, so that the leg mass is decreased and motor load can be balanced. Secondly, an online gait generator based on dynamic movement primitives for the walking control is presented. Dynamic movement primitives provide an approach to generate periodic trajectories and they can be modulated in real time, which makes the online adjustment of walking gaits possible. This gait controller is tested by the quadruped robot in regulating walking speed, switching between forward\backward movements and steering. The controller is easy to apply, expand and is quite effective on phase coordination and online trajectory modulation. Results of simulated experiments are presented.

Variable Admittance Control for Climbing Stairs in Human-Powered Exoskeleton Systems

Online gait control in human-powered exoskeleton systems is still rich research field and represents a step towards fully autonomous, safe and intelligent navigation. Admittance Controller performs well on flat terrain walking in human-powered exoskeleton systems for acceleration and slowdown. We are the first who proposed Variable Admittance Controller (VAC) for smooth stair climbing control in Human-Powered Exoskeleton Systems. Trajectory correction technique transforms the interaction forces exerted on the exoskeleton from the pilot to appropriate intended joint flexion angles through dynamic viscoelastic models. We demonstrate the proposed control strategy on one degree-of-freedom (1-DOF) platform first, and then extend to the Human power Augmentation Lower Exoskeleton (HUALEX). The experimental results show that the proposed gait transition control strategy can minimize the interaction dynamics with less interaction force between the pilot and the exoskeleton compared to the traditional admittance controller. Compared to Ordinary Admittance Controller, the proposed VAC significantly improve the normalized Mean Squared Error (nMSE) of trajectory tracking from 2.751° to1.105°.

Survey of On-line Control Strategies of Human-Powered Augmentation Exoskeleton Systems

On-line gait control in human-powered exoskeleton systems is still rich research field and represents a step towards fully autonomous, safe and intelligent indoor and outdoor navigation. It is still a big challenge to develop a control strategy which makes the exoskeleton supply an efficient tracking for pilot intended trajectories on-line. Considering the number of degrees of freedom the lower limb exoskeletons are simpler to design, compared to upper limb. The comparison between lower limb and upper limb is useless when consider the control issues, because of the differences in missions and applications. Based on the literature, we aim to give an overview about control strategies of some famous lower limb human power exoskeleton systems. In the state of the art, different control strategies and approaches for different types of lower limb exoskeletons will be compared consider the efficiency and economic issues. Exact estimation of needed joints torques to execute human intended motions on-line with efficient performance, low cost and reliable way is the main goal of studied system’s control strategies. We have study different control strategies used for wide known human power augmentation exoskeletons and compare between them in graphs and tables.

Energy-Efficient Gait Planning and Control for Biped Robots Utilizing Vertical Body Motion and Allowable ZMP Region

An energy-efficient gait planning (EEGP) and control system is established for biped robots with three-mass inverted pendulum mode (3MIPM), which utilizes both vertical body motion (VBM) and allowable zero-moment-point (ZMP) region (AZR). Given a distance to be traveled, we newly designed an online gait synthesis algorithm to construct a complete walking cycle, i.e., a starting step, multiple cyclic steps, and a stopping step, in which: 1) ZMP was fully manipulated within AZR; and 2) vertical body movement was allowed to relieve knee bending. Moreover, gait parameter optimization is effectively performed to determine the optimal set of gait parameters, i.e., average body height and amplitude of VBM, number of steps, and average walking speed, which minimizes energy consumption of actuation motors for leg joints under practical constraints, i.e., geometrical constraints, friction force limit, and yawing moment limit. Various simulations were conducted to identify the effectiveness of the proposed method and verify energy-saving performance for various ZMP regions. Our control system was implemented and tested on the humanoid robot DARwIn-OP.

Parameter Synthesis of Coupled Nonlinear Oscillators for CPG-Based Robotic Locomotion

This paper presents a numerical method for parameter synthesis of a central pattern generator (CPG) network to acquire desired locomotor patterns. The CPG network is modeled as a chain of unidirectionally or bidirectionally coupled Hopf oscillators with a novel coupling scheme that eliminates the influence of afferent signals on amplitude of the oscillator. The method converts the related CPG parameters into dynamic systems that evolve as part of the CPG network dynamics. The frequency, amplitude, and phase relations of teaching signals can be encoded by the CPG network with the proposed learning rules. The ability of the method to learn instructed locomotor pattern is proven with simulations. Application of the proposed method to online gait synthesis of a robotic fish is also presented.

Energy-Efficient Gait Planning and Control for Biped Robots Utilizing the Allowable ZMP Region

Energy-efficient gait planning and control is established for biped robots, which utilizes the allowable zero moment point (ZMP) region. Based on 3-D linear inverted pendulum mode (LIPM), we construct a practical gait planning algorithm for a given travel distance minimizing the energy consumed by the actuators of humanoid joints with 1) an online gait synthesis (GSYN) algorithm to generate a complete walking cycle (a starting step, several cyclic steps, and a stopping step) compromising waking stability and energy efficiency at the fully utilizing allowable ZMP region and with 2) effective gait parameter optimization to maximize the energy efficiency of the gaits generated by GSYN, finding two optimal parameters-number of steps and average walking speed-satisfying geometrical constraints, friction force limit, and yawing moment limit to guarantee feasible motions. The proposed algorithm was verified through simulations, and the gait control system was implemented on a DARwIn-OP humanoid robot.

Development of a new spinning gait for a planar snake robot using central pattern generators

In this paper, we first present dynamic equation of n-link snake robot using Lagrange's method in a simplified matrix form and verify them experimentally. Next, we introduce a new locomotion mode called spinning gait. Central pattern generators (CPGs) are used for online gait generation. To realize spinning gait, genetic algorithm is used to find optimal CPG network parameters. We illustrate both theoretically, using derived robot dynamics and experimentally that the CPG-based online gait generation method allows continuous and rather smooth transitions between gaits. Lastly, we present an application where the snake robot is guided from an initial to final position while avoiding obstacles by changing CPG parameters.

A framework for online gait recognition based on multilinear tensor analysis

The gait recognition is to recognize an individual based on the characteristics extracted from the gait image sequence. There are many researches for the gait recognition which use diverse kinds of information such as shape of gait silhouette, motion variation caused by walking, and so on. In general, shape information is more useful for recognition. However, shape information is influenced by a variety of factors, which degrade the recognition performance. Moreover, the information used in most of those studies might be able to be extracted after all of one or more sequences of the gait cycle are known. And it is also hard to discriminate the gait cycle from given gait sequences exactly by the online approach. In regard to these difficulties, we propose a novel gait recognition method based on the multilinear tensor analysis. To recognize the cyclic characteristic of gait without an exact division for the gait cycle, this paper's propose is the method to form the accumulated silhouette and then describes those as the tensor. For the accumulated silhouette proposed by this paper, the image sequence of one gait cycle is divided into four sections in the training phase. However, discrimination for the gait cycle in the training phase is not directly related to the recognition phase, thus the online approach is possible. We first form the accumulated silhouettes for every individual using gait silhouettes within each section. And then, we represent these accumulated silhouettes as the tensor. Using a multilinear tensor analysis, we compute the core tensor which governs the interaction between factors organizing the original tensor, and then compose the basis to recognize the individual in the online recognition framework. Finally, we recognize the individual using the computation of similarity based on the Euclidean distance, which is more suitable to our method. We verify the superiority of the proposed approach via experiments with real gait sequences.

On the Suitability of Different Online Gait Generation Methods for Pre‐Defined Footsteps

AbstractDetailed dynamical modeling is the basis for simulation and model based control. In this contribution the Projection Equation is used for the modeling of a biped walking machine, resulting in the equations of motion which are needed for gait generation and verification of its stability. For biped robots one main issue is the generation of stable trajectories for the center of mass (CoM). Several different approaches based on the Zero Moment Point (ZMP) scheme have been presented in the past. Due to the complex dynamic structure of bipedal robots most of the considered algorithms use a linear inverted pendulum as a simplified model. This results in a decoupling of the ZMP equations in lateral and forward direction, but limits the trajectories to a constant height of the CoM. An extension of the well known LQR theory by future reference values has been proposed. This model based approach seems to perform quite well, but does not allow the consideration of constraints on the position of the ZMP. This limitation is removed by the use of Model Predictive Control (MPC) with inequality constraints. By extending this approach to a time invariant one the restriction to a constant height of the CoM is no longer necessary. Both methods as well as the time invariant approach for variable CoM heights have been evaluated in simulations and will be experimentally verified on a real robot soon. (© 2011 Wiley‐VCH Verlag GmbH & Co. KGaA, Weinheim)

Motion Planning for Humanoid Robot Based on Hybrid Evolutionary Algorithm

In this paper, online gait control system is designed for walking-up-stairs movement according to the features of humanoid robot, the hybrid evolutionary approach based on neural network optimized by particle swarm is employed for the offline training of the movement process, and the optimal gait of the stability is generated. Additionally, through embedded monocular vision, on-site environmental information is collected as neural network input, so necessary joint trajectory is output for the movement. Simulations and experiment testify the efficiency of the method.

Online gait event detection using a large force platform embedded in a treadmill

Gait research and clinical gait training may benefit from movement-dependent event control, that is, technical applications in which events such as obstacle appearance or visual/acoustic cueing are (co)determined online on the basis of current gait properties. A prerequisite for successful gait-dependent event control is accurate online detection of gait events such as foot contact (FC) and foot off (FO). The objective of the present study was to assess the feasibility of online FC and FO detection using a single large force platform embedded in a treadmill. Center-of-pressure, total force output and kinematic data were recorded simultaneously in 12 healthy participants. Online FC and FO estimates and spatial and temporal gait parameters estimated from the force platform data—i.e., center-of-pressure profiles—were compared to offline kinematic counterparts, which served as the gold standard. Good correspondence was achieved between online FC detections using center-of-pressure profiles and those derived offline from kinematic data, whereas FO was detected 31 ms too late. A good relative and absolute agreement was achieved for both spatial and temporal gait parameters, which was improved further by applying more fine-grained FO estimation procedures using characteristic local minima in the total force output time series. These positive results suggest that the proposed system for gait-dependent event control may be successfully implemented in gait research as well as gait interventions in clinical practice.

ONLINE DYNAMICALLY BALANCED ASCENDING AND DESCENDING GAIT GENERATIONS OF A BIPED ROBOT USING SOFT COMPUTING

In the present paper, two algorithms based on soft computing have been developed for dynamically balanced gait generations of a biped robot ascending and descending a staircase. The utility of the soft computing tools is best justified, when the data available for the problem to be solved are imprecise in nature, difficult to model and exhibit large-scale solution spaces. The problem of online gait generation of a biped robot exhibits such a complex phenomenon, and ultimately soft computing has become a natural choice for solving it. The gait generation problems of a biped robot have been solved using two different approaches, namely genetic-neural (GA-NN) and genetic-fuzzy (GA-FLC) systems. In GA-NN, the gait generation problem of a two-legged robot has been modeled using two modules of Neural Network (NN), whose weights are optimized offline using a Genetic Algorithm (GA), whereas in GA-FLC, the above problem is modeled utilizing two modules of Fuzzy Logic Controller (FLC) and their rule bases are optimized offline using a GA. Once optimized, the GA-NN and GA-FLC systems will be able to generate dynamically balanced gaits of the biped robot online. The performances of the two approaches are compared with respect to the Dynamic Balance Margin (DBM).

Multi-legged multi-roped walking and climbing robots: online static equilibrium analysis

The paper proposes a fast method to solve the equilibrium problem of statically indeterminate walking and climbing robots with quasi-static locomotion and whatever number of legs and ropes connecting the robot frame to the ground. The configuration is instantaneously assigned. Legs and rope winches are imagined blocked. Due to the number of legs and ropes, the robot is a statically indeterminate system under unilateral constraints: in order to solve the system of the equilibrium equations, it is necessary to take into account the compliance of the robot and of the terrain. This can be done by non-linear analysis of a finite element model of the robot, including the ground characteristics, but this is very time consuming. The method presented in the paper seems an effective alternative. To assess its performances, the method is applied to the heavy-duty climbing robot Roboclimber with satisfactory results. The running time is low enough to allow the application for the online gait planning and for real-time control of the robot. The method will be soon used as decision support for online gait planning within the remote control system of Roboclimber.