Mark A. Hoepflinger

State Estimation for Legged Robots - Consistent Fusion of Leg Kinematics and IMU

This paper introduces a state estimation framework for legged robots that allows estimating the full pose of the robot without making any assumptions about the geometrical structure of its environment. This is achieved by means of an Observability Constrained Extended Kalman Filter that fuses kinematic encoder data with on-board IMU measurements. By including the absolute position of all footholds into the filter state, simple model equations can be formulated which accurately capture the uncertainties associated with the intermittent ground contacts. The resulting filter simultaneously estimates the position of all footholds and the pose of the main body. In the algorithmic formulation, special attention is paid to the consistency of the linearized filter: it maintains the same observability properties as the nonlinear system, which is a prerequisite for accurate state estimation. The presented approach is implemented in simulation and validated experimentally on an actual quadrupedal robot.

Quadrupedal locomotion using hierarchical operational space control

This paper presents the application of operational space control based on hierarchical task optimization for quadrupedal locomotion. We show how the behavior of a complex robotic machine can be described by a simple set of least squares problems with different priorities for motion, torque, and force optimization. Using projected dynamics of floating base systems with multiple contact points, the optimization dimensionality can be reduced or decoupled such that the formulation is purely based on the inversion of kinematic system properties. The present controller is extensively tested in various experiments using the fully torque controllable quadrupedal robot StarlETH. The load distribution is optimized for static walking gaits to improve contact stability and/or actuator efficiency under various terrain conditions. This is augmented with simultaneous joint position and torque limitations as well as with an interpolation method to ensure smooth contact transitions. The same control structure is further used to stabilize dynamic trotting gaits under significant external disturbances such as uneven ground or pushes. To the best of our knowledge, this work is the first documentation of static and dynamic locomotion with pure task-space inverse dynamics (no joint position feedback) control.

Efficient and Versatile Locomotion With Highly Compliant Legs

Drawing inspiration from nature, this paper introduces and compares two compliant robotic legs that are able to perform precise joint torque and position control, enable passive adaption to the environment, and allow for the exploitation of natural dynamic motions. We report in detail on the design and control of both prototypes and elaborate specifically on the problem of precise foot placement during flight without the sacrifice of efficient energy storage during stance. This is achieved through an integrated design and control approach that incorporates series elastic actuation, series damping actuation, and active damping through torque control. The two legs are employed in efficient hopping/running motions for which they achieve performance similar to humans or animals. This paper is concluded by a comparison of the various design choices with respect to performance and applicability, as well as an outlook on the usage of these legs in a fully actuated quadruped.

State estimation for legged robots on unstable and slippery terrain

This paper presents a state estimation approach for legged robots based on stochastic filtering. The key idea is to extract information from the kinematic constraints given through the intermittent contacts with the ground and to fuse this information with inertial measurements. To this end, we design an unscented Kalman filter based on a consistent formulation of the underlying stochastic model. To increase the robustness of the filter, an outliers rejection methodology is included into the update step. Furthermore, we present the nonlinear observability analysis of the system, where, by considering the special nature of 3D rotations, we obtain a relatively simple form of the corresponding observability matrix. This yields, that, except for the global position and the yaw angle, all states are in general observable. This also holds if only one foot is in contact with the ground. The presented filter is evaluated on a real quadruped robot trotting over an uneven and slippery terrain.

Practice Makes Perfect: An Optimization-Based Approach to Controlling Agile Motions for a Quadruped Robot

This article approaches the problem of controlling quadrupedal running and jumping motions with a parameterized, model-based, state-feedback controller. Inspired by the motor learning principles observed in nature, our method automatically fine tunes the parameters of our controller by repeatedly executing slight variations of the same motion task. This learn-through-practice process is performed in simulation to best exploit computational resources and to prevent the robot from damaging itself. To ensure that the simulation results match the behavior of the hardware platform, we introduce and validate an accurate model of the compliant actuation system. The proposed method is experimentally verified on the torque-controllable quadruped robot StarlETH by executing squat jumps and dynamic gaits, such as a running trot, pronk, and a bounding gait.

Quadrupedal robots with stiff and compliant actuation

Abstract In the broader context of quadrupedal locomotion, this overview article introduces and compares two platforms that are similar in structure, size, and morphology, yet differ greatly in their concept of actuation. The first, ALoF, is a classically stiff actuated robot that is controlled kinematically, while the second, StarlETH, uses a soft actuation scheme based on Changedhighly compliant series elastic actuators. We show how this conceptual difference influences design and control of the robots, compare the hardware of the two systems, and show exemplary their advantages in different applications. Zusammenfassung Der vorliegende Beitrag vergleicht zwei Laufroboter, die sich in Hinblick auf Struktur, Größe und Morphologie stark ähneln, jedoch im Antriebskonzept klar unterscheiden. Während es sich beim ersten System, ALoF, um einen klassisch angetriebenen Roboter handelt der kinematisch geregelt wird, besitzt der zweite Roboter, StarlETH, Federelemente im Antriebsstrang. Diese ermöglichen eine weiche, kraftgeregelte Aktuierung. Der Beitrag zeigt wie dieser Unterschied Design und Regelung der Roboter beeinflusst, vergleicht die Hardware und erläutert Vor- und Nachteile in verschiedenen Anwendungsfällen.

HAPTIC TERRAIN CLASSIFICATION ON NATURAL TERRAINS FOR LEGGED ROBOTS

In this paper, we are presenting a method to estimate terrain properties (such as small-scale geometry or surface friction) to improve the assessment of stability and the guiding of foot placement of legged robots in rough terrain. Haptic feedback, expressed through joint motor currents and ground contact force measurements that arises when prescribing a predefined motion was collected for a variety of ground samples (four different shapes and four different surface properties). Features were extracted from this data and used for training and classification by a multiclass AdaBoost machine learning algorithm. In a single leg testbed, the algorithm could correctly classify about 94% of the terrain shapes, and about 73% of the surface samples.

State Estimation for Legged Robots

The Dynamic Legged Systems Lab (DLS) is currently working on the implementation of state estimation algorithms for a range of diverse applications for quadruped robots. Some examples include probabilistic foot contact estimation [1] and hetrogeneous sensor fusion for accurate state estimation [2]. Our hydraulic quadruped robot series HyQ is a fully torque-controlled system, capable of locomotion over rough terrain and performing highly dynamic tasks such as jumping and running with a variety of gaits. It is a unique research platform, designed for unstructured environments. We are currently looking for a highly motivated, full-time internship position to work on the implementation, evaluation, and further development of state estimation algorithms into our legged robot framework.

WALKING AND CRAWLING WITH ALoF - A ROBOT FOR AUTONOMOUS LOCOMOTION ON FOUR LEGS

Purpose – The purpose of this paper is to introduce the robotic quadrupedal platform ALoF that is designed to aid research on perception in legged locomotion. Design/methodology/approach – A well-balanced size and complexity of the robot results in a robust platform that is easy to handle, yet able to perform complex maneuvers as well as to carry sophisticated 3D sensors. A very large range of motion allows the robot to actively explore its surroundings through haptic interaction, and to choose between a wide range of planning options. Findings – This robot was employed and tested in the lunar robotics challenge organized by the European Space Agency, for which the authors also developed a novel crawling gait, in which the weight of the robot is alternately supported by scaled plates under the main body and the four shank segments. This allowed for stable locomotion in steep terrain with very loose soil. Originality/value – The paper describes how a very large range of motion allows the robot to actively explore its surroundings through haptic interaction, and to choose between a wide range of planning options. The paper describes how the authors developed a novel crawling gait, in which the weight of the robot is alternately supported by scaled plates under the main body and the four shank segments.

Extrinsic RGB-D camera calibration for legged robots

This paper describes a method to identify the extrinsic parameters of RGB-D cameras mounted on legged robots. Since the calculation of the parameters is based on the detection of the robots feet in the camera images, no special calibration objects are required. Therefore the method is simple to use and can even be applied automatically. Experiments demonstrate the precision and the robustness of the method.