Header logo is am


2010


no image
Constrained Accelerations for Controlled Geometric Reduction: Sagittal-Plane Decoupling for Bipedal Locomotion

Gregg, R., Righetti, L., Buchli, J., Schaal, S.

In 2010 10th IEEE-RAS International Conference on Humanoid Robots, pages: 1-7, IEEE, Nashville, USA, 2010 (inproceedings)

Abstract
Energy-shaping control methods have produced strong theoretical results for asymptotically stable 3D bipedal dynamic walking in the literature. In particular, geometric controlled reduction exploits robot symmetries to control momentum conservation laws that decouple the sagittal-plane dynamics, which are easier to stabilize. However, the associated control laws require high-dimensional matrix inverses multiplied with complicated energy-shaping terms, often making these control theories difficult to apply to highly-redundant humanoid robots. This paper presents a first step towards the application of energy-shaping methods on real robots by casting controlled reduction into a framework of constrained accelerations for inverse dynamics control. By representing momentum conservation laws as constraints in acceleration space, we construct a general expression for desired joint accelerations that render the constraint surface invariant. By appropriately choosing an orthogonal projection, we show that the unconstrained (reduced) dynamics are decoupled from the constrained dynamics. Any acceleration-based controller can then be used to stabilize this planar subsystem, including passivity-based methods. The resulting control law is surprisingly simple and represents a practical way to employ control theoretic stability results in robotic platforms. Simulated walking of a 3D compass-gait biped show correspondence between the new and original controllers, and simulated motions of a 16-DOF humanoid demonstrate the applicability of this method.

link (url) DOI [BibTex]

2010

link (url) DOI [BibTex]


no image
Inverse dynamics with optimal distribution of ground reaction forces for legged robot

Righetti, L., Buchli, J., Mistry, M., Schaal, S.

In Proceedings of the 13th International Conference on Climbing and Walking Robots (CLAWAR), pages: 580-587, Nagoya, Japan, sep 2010 (inproceedings)

Abstract
Contact interaction with the environment is crucial in the design of locomotion controllers for legged robots, to prevent slipping for example. Therefore, it is of great importance to be able to control the effects of the robots movements on the contact reaction forces. In this contribution, we extend a recent inverse dynamics algorithm for floating base robots to optimize the distribution of contact forces while achieving precise trajectory tracking. The resulting controller is algorithmically simple as compared to other approaches. Numerical simulations show that this result significantly increases the range of possible movements of a humanoid robot as compared to the previous inverse dynamics algorithm. We also present a simplification of the result where no inversion of the inertia matrix is needed which is particularly relevant for practical use on a real robot. Such an algorithm becomes interesting for agile locomotion of robots on difficult terrains where the contacts with the environment are critical, such as walking over rough or slippery terrain.

DOI [BibTex]

DOI [BibTex]


no image
Variable impedance control - a reinforcement learning approach

Buchli, J., Theodorou, E., Stulp, F., Schaal, S.

In Robotics Science and Systems (2010), Zaragoza, Spain, June 27-30, 2010, clmc (inproceedings)

Abstract
One of the hallmarks of the performance, versatility, and robustness of biological motor control is the ability to adapt the impedance of the overall biomechanical system to different task requirements and stochastic disturbances. A transfer of this principle to robotics is desirable, for instance to enable robots to work robustly and safely in everyday human environments. It is, however, not trivial to derive variable impedance controllers for practical high DOF robotic tasks. In this contribution, we accomplish such gain scheduling with a reinforcement learning approach algorithm, PI2 (Policy Improvement with Path Integrals). PI2 is a model-free, sampling based learning method derived from first principles of optimal control. The PI2 algorithm requires no tuning of algorithmic parameters besides the exploration noise. The designer can thus fully focus on cost function design to specify the task. From the viewpoint of robotics, a particular useful property of PI2 is that it can scale to problems of many DOFs, so that RL on real robotic systems becomes feasible. We sketch the PI2 algorithm and its theoretical properties, and how it is applied to gain scheduling. We evaluate our approach by presenting results on two different simulated robotic systems, a 3-DOF Phantom Premium Robot and a 6-DOF Kuka Lightweight Robot. We investigate tasks where the optimal strategy requires both tuning of the impedance of the end-effector, and tuning of a reference trajectory. The results show that we can use path integral based RL not only for planning but also to derive variable gain feedback controllers in realistic scenarios. Thus, the power of variable impedance control is made available to a wide variety of robotic systems and practical applications.

link (url) [BibTex]

link (url) [BibTex]

2009


Thumb xl synchro
Synchronized Oriented Mutations Algorithm for Training Neural Controllers

Berenz, V., Suzuki, K.

In Advances in Neuro-Information Processing: 15th International Conference, ICONIP 2008, Auckland, New Zealand, November 25-28, 2008, Revised Selected Papers, Part II, pages: 244-251, Springer Berlin Heidelberg, Berlin, Heidelberg, 2009 (inbook)

link (url) DOI [BibTex]

2009

link (url) DOI [BibTex]


Thumb xl screen shot 2015 08 23 at 14.50.55
Grasping familiar objects using shape context

Bohg, J., Kragic, D.

In Advanced Robotics, 2009. ICAR 2009. International Conference on, pages: 1-6, 2009 (inproceedings)

Abstract
We present work on vision based robotic grasping. The proposed method relies on extracting and representing the global contour of an object in a monocular image. A suitable grasp is then generated using a learning framework where prototypical grasping points are learned from several examples and then used on novel objects. For representation purposes, we apply the concept of shape context and for learning we use a supervised learning approach in which the classifier is trained with labeled synthetic images. Our results show that a combination of a descriptor based on shape context with a non-linear classification algorithm leads to a stable detection of grasping points for a variety of objects. Furthermore, we will show how our representation supports the inference of a full grasp configuration.

pdf slides [BibTex]

pdf slides [BibTex]


Thumb xl screen shot 2015 08 23 at 14.45.26
Integration of Visual Cues for Robotic Grasping

Bergström, N., Bohg, J., Kragic, D.

In Computer Vision Systems, 5815, pages: 245-254, Lecture Notes in Computer Science, Springer Berlin Heidelberg, 2009 (incollection)

Abstract
In this paper, we propose a method that generates grasping actions for novel objects based on visual input from a stereo camera. We are integrating two methods that are advantageous either in predicting how to grasp an object or where to apply a grasp. The first one reconstructs a wire frame object model through curve matching. Elementary grasping actions can be associated to parts of this model. The second method predicts grasping points in a 2D contour image of an object. By integrating the information from the two approaches, we can generate a sparse set of full grasp configurations that are of a good quality. We demonstrate our approach integrated in a vision system for complex shaped objects as well as in cluttered scenes.

pdf link (url) DOI [BibTex]

pdf link (url) DOI [BibTex]


Thumb xl screen shot 2015 08 23 at 15.10.20
Towards Grasp-Oriented Visual Perception of Humanoid Robots

Bohg, J., Barck-Holst, C., Huebner, K., Ralph, M., Rasolzadeh, B., Song, D., Kragic, D.

International Journal of Humanoid Robotics, 06(03):387-434, 2009 (article)

Abstract
A distinct property of robot vision systems is that they are embodied. Visual information is extracted for the purpose of moving in and interacting with the environment. Thus, different types of perception-action cycles need to be implemented and evaluated. In this paper, we study the problem of designing a vision system for the purpose of object grasping in everyday environments. This vision system is firstly targeted at the interaction with the world through recognition and grasping of objects and secondly at being an interface for the reasoning and planning module to the real world. The latter provides the vision system with a certain task that drives it and defines a specific context, i.e. search for or identify a certain object and analyze it for potential later manipulation. We deal with cases of: (i) known objects, (ii) objects similar to already known objects, and (iii) unknown objects. The perception-action cycle is connected to the reasoning system based on the idea of affordances. All three cases are also related to the state of the art and the terminology in the neuroscientific area.

pdf DOI [BibTex]

pdf DOI [BibTex]


Thumb xl 5420560 fig 1 glance
Sensory-objects network driven by intrinsic motivation for survival abilities

Berenz, V., Suzuki, K.

In Robotics and Biomimetics (ROBIO), 2009 IEEE International Conference on, pages: 871-876, 2009 (inproceedings)

DOI [BibTex]

DOI [BibTex]


no image
A Limiting Property of the Matrix Exponential with Application to Multi-loop Control

Trimpe, S., D’Andrea, R.

In Proceedings of the Joint 48th IEEE Conference on Decision (CDC) and Control and 28th Chinese Control Conference, 2009 (inproceedings)

PDF DOI [BibTex]

PDF DOI [BibTex]


no image
Path integral-based stochastic optimal control for rigid body dynamics

Theodorou, E. A., Buchli, J., Schaal, S.

In Adaptive Dynamic Programming and Reinforcement Learning, 2009. ADPRL ’09. IEEE Symposium on, pages: 219-225, 2009, clmc (inproceedings)

Abstract
Recent advances on path integral stochastic optimal control [1],[2] provide new insights in the optimal control of nonlinear stochastic systems which are linear in the controls, with state independent and time invariant control transition matrix. Under these assumptions, the Hamilton-Jacobi-Bellman (HJB) equation is formulated and linearized with the use of the logarithmic transformation of the optimal value function. The resulting HJB is a linear second order partial differential equation which is solved by an approximation based on the Feynman-Kac formula [3]. In this work we review the theory of path integral control and derive the linearized HJB equation for systems with state dependent control transition matrix. In addition we derive the path integral formulation for the general class of systems with state dimensionality that is higher than the dimensionality of the controls. Furthermore, by means of a modified inverse dynamics controller, we apply path integral stochastic optimal control over the new control space. Simulations illustrate the theoretical results. Future developments and extensions are discussed.

link (url) [BibTex]

link (url) [BibTex]


no image
Learning locomotion over rough terrain using terrain templates

Kalakrishnan, M., Buchli, J., Pastor, P., Schaal, S.

In Intelligent Robots and Systems, 2009. IROS 2009. IEEE/RSJ International Conference on, pages: 167-172, 2009, clmc (inproceedings)

Abstract
We address the problem of foothold selection in robotic legged locomotion over very rough terrain. The difficulty of the problem we address here is comparable to that of human rock-climbing, where foot/hand-hold selection is one of the most critical aspects. Previous work in this domain typically involves defining a reward function over footholds as a weighted linear combination of terrain features. However, a significant amount of effort needs to be spent in designing these features in order to model more complex decision functions, and hand-tuning their weights is not a trivial task. We propose the use of terrain templates, which are discretized height maps of the terrain under a foothold on different length scales, as an alternative to manually designed features. We describe an algorithm that can simultaneously learn a small set of templates and a foothold ranking function using these templates, from expert-demonstrated footholds. Using the LittleDog quadruped robot, we experimentally show that the use of terrain templates can produce complex ranking functions with higher performance than standard terrain features, and improved generalization to unseen terrain.

link (url) Project Page [BibTex]

link (url) Project Page [BibTex]


Valero-Cuevas, F., Hoffmann, H., Kurse, M. U., Kutch, J. J., Theodorou, E. A.

IEEE Reviews in Biomedical Engineering – (All authors have equally contributed), (2):110?135, 2009, clmc (article)

Abstract
Computational models of the neuromuscular system hold the potential to allow us to reach a deeper understanding of neuromuscular function and clinical rehabilitation by complementing experimentation. By serving as a means to distill and explore specific hypotheses, computational models emerge from prior experimental data and motivate future experimental work. Here we review computational tools used to understand neuromuscular function including musculoskeletal modeling, machine learning, control theory, and statistical model analysis. We conclude that these tools, when used in combination, have the potential to further our understanding of neuromuscular function by serving as a rigorous means to test scientific hypotheses in ways that complement and leverage experimental data.

link (url) [BibTex]

link (url) [BibTex]


no image
Compact models of motor primitive variations for predictible reaching and obstacle avoidance

Stulp, F., Oztop, E., Pastor, P., Beetz, M., Schaal, S.

In IEEE-RAS International Conference on Humanoid Robots (Humanoids 2009), Paris, Dec.7-10, 2009, clmc (inproceedings)

Abstract
over and over again. This regularity allows humans and robots to reuse existing solutions for known recurring tasks. We expect that reusing a set of standard solutions to solve similar tasks will facilitate the design and on-line adaptation of the control systems of robots operating in human environments. In this paper, we derive a set of standard solutions for reaching behavior from human motion data. We also derive stereotypical reaching trajectories for variations of the task, in which obstacles are present. These stereotypical trajectories are then compactly represented with Dynamic Movement Primitives. On the humanoid robot Sarcos CB, this approach leads to reproducible, predictable, and human-like reaching motions.

link (url) [BibTex]

link (url) [BibTex]


no image
Human optimization strategies under reward feedback

Hoffmann, H., Theodorou, E., Schaal, S.

In Abstracts of Neural Control of Movement Conference (NCM 2009), Waikoloa, Hawaii, 2009, 2009, clmc (inproceedings)

Abstract
Many hypothesis on human movement generation have been cast into an optimization framework, implying that movements are adapted to optimize a single quantity, like, e.g., jerk, end-point variance, or control cost. However, we still do not understand how humans actually learn when given only a cost or reward feedback at the end of a movement. Such a reinforcement learning setting has been extensively explored theoretically in engineering and computer science, but in human movement control, hardly any experiment studied movement learning under reward feedback. We present experiments probing which computational strategies humans use to optimize a movement under a continuous reward function. We present two experimental paradigms. The first paradigm mimics a ball-hitting task. Subjects (n=12) sat in front of a computer screen and moved a stylus on a tablet towards an unknown target. This target was located on a line that the subjects had to cross. During the movement, visual feedback was suppressed. After the movement, a reward was displayed graphically as a colored bar. As reward, we used a Gaussian function of the distance between the target location and the point of line crossing. We chose such a function since in sensorimotor tasks, the cost or loss function that humans seem to represent is close to an inverted Gaussian function (Koerding and Wolpert 2004). The second paradigm mimics pocket billiards. On the same experimental setup as above, the computer screen displayed a pocket (two bars), a white disk, and a green disk. The goal was to hit with the white disk the green disk (as in a billiard collision), such that the green disk moved into the pocket. Subjects (n=8) manipulated with the stylus the white disk to effectively choose start point and movement direction. Reward feedback was implicitly given as hitting or missing the pocket with the green disk. In both paradigms, subjects increased the average reward over trials. The surprising result was that in these experiments, humans seem to prefer a strategy that uses a reward-weighted average over previous movements instead of gradient ascent. The literature on reinforcement learning is dominated by gradient-ascent methods. However, our computer simulations and theoretical analysis revealed that reward-weighted averaging is the more robust choice given the amount of movement variance observed in humans. Apparently, humans choose an optimization strategy that is suitable for their own movement variance.

[BibTex]

[BibTex]


no image
Bayesian Methods for Autonomous Learning Systems (Phd Thesis)

Ting, J.

Department of Computer Science, University of Southern California, Los Angeles, CA, 2009, clmc (phdthesis)

PDF [BibTex]

PDF [BibTex]


no image
On-line learning and modulation of periodic movements with nonlinear dynamical systems

Gams, A., Ijspeert, A., Schaal, S., Lenarčič, J.

Autonomous Robots, 27(1):3-23, 2009, clmc (article)

Abstract
Abstract  The paper presents a two-layered system for (1) learning and encoding a periodic signal without any knowledge on its frequency and waveform, and (2) modulating the learned periodic trajectory in response to external events. The system is used to learn periodic tasks on a humanoid HOAP-2 robot. The first layer of the system is a dynamical system responsible for extracting the fundamental frequency of the input signal, based on adaptive frequency oscillators. The second layer is a dynamical system responsible for learning of the waveform based on a built-in learning algorithm. By combining the two dynamical systems into one system we can rapidly teach new trajectories to robots without any knowledge of the frequency of the demonstration signal. The system extracts and learns only one period of the demonstration signal. Furthermore, the trajectories are robust to perturbations and can be modulated to cope with a dynamic environment. The system is computationally inexpensive, works on-line for any periodic signal, requires no additional signal processing to determine the frequency of the input signal and can be applied in parallel to multiple dimensions. Additionally, it can adapt to changes in frequency and shape, e.g. to non-stationary signals, such as hand-generated signals and human demonstrations.

link (url) [BibTex]

link (url) [BibTex]


no image
The SL simulation and real-time control software package

Schaal, S.

University of Southern California, Los Angeles, CA, 2009, clmc (techreport)

Abstract
SL was originally developed as a Simulation Laboratory software package to allow creating complex rigid-body dynamics simulations with minimal development times. It was meant to complement a real-time robotics setup such that robot programs could first be debugged in simulation before trying them on the actual robot. For this purpose, the motor control setup of SL was copied from our experience with real-time robot setups with vxWorks (Windriver Systems, Inc.)Ñindeed, more than 90% of the code is identical to the actual robot software, as will be explained later in detail. As a result, SL is divided into three software components: 1) the generic code that is shared by the actual robot and the simulation, 2) the robot specific code, and 3) the simulation specific code. The robot specific code is tailored to the robotic environments that we have experienced over the years, in particular towards VME-based multi-processor real-time operating systems. The simulation specific code has all the components for OpenGL graphics simulations and mimics the robot multi-processor environment in simple C-code. Importantly, SL can be used stand-alone for creating graphics an-imationsÑthe heritage from real-time robotics does not restrict the complexity of possible simulations. This technical report describes SL in detail and can serve as a manual for new users of SL.

link (url) [BibTex]

link (url) [BibTex]


no image
Local dimensionality reduction for non-parametric regression

Hoffman, H., Schaal, S., Vijayakumar, S.

Neural Processing Letters, 2009, clmc (article)

Abstract
Locally-weighted regression is a computationally-efficient technique for non-linear regression. However, for high-dimensional data, this technique becomes numerically brittle and computationally too expensive if many local models need to be maintained simultaneously. Thus, local linear dimensionality reduction combined with locally-weighted regression seems to be a promising solution. In this context, we review linear dimensionality-reduction methods, compare their performance on nonparametric locally-linear regression, and discuss their ability to extend to incremental learning. The considered methods belong to the following three groups: (1) reducing dimensionality only on the input data, (2) modeling the joint input-output data distribution, and (3) optimizing the correlation between projection directions and output data. Group 1 contains principal component regression (PCR); group 2 contains principal component analysis (PCA) in joint input and output space, factor analysis, and probabilistic PCA; and group 3 contains reduced rank regression (RRR) and partial least squares (PLS) regression. Among the tested methods, only group 3 managed to achieve robust performance even for a non-optimal number of components (factors or projection directions). In contrast, group 1 and 2 failed for fewer components since these methods rely on the correct estimate of the true intrinsic dimensionality. In group 3, PLS is the only method for which a computationally-efficient incremental implementation exists. Thus, PLS appears to be ideally suited as a building block for a locally-weighted regressor in which projection directions are incrementally added on the fly.

link (url) [BibTex]

link (url) [BibTex]


no image
The SL simulation and real-time control software package

Schaal, S.

University of Southern California, Los Angeles, CA, 2009, clmc (techreport)

Abstract
SL was originally developed as a Simulation Laboratory software package to allow creating complex rigid-body dynamics simulations with minimal development times. It was meant to complement a real-time robotics setup such that robot programs could first be debugged in simulation before trying them on the actual robot. For this purpose, the motor control setup of SL was copied from our experience with real-time robot setups with vxWorks (Windriver Systems, Inc.)â??indeed, more than 90% of the code is identical to the actual robot software, as will be explained later in detail. As a result, SL is divided into three software components: 1) the generic code that is shared by the actual robot and the simulation, 2) the robot specific code, and 3) the simulation specific code. The robot specific code is tailored to the robotic environments that we have experienced over the years, in particular towards VME-based multi-processor real-time operating systems. The simulation specific code has all the components for OpenGL graphics simulations and mimics the robot multi-processor environment in simple C-code. Importantly, SL can be used stand-alone for creating graphics an-imationsâ??the heritage from real-time robotics does not restrict the complexity of possible simulations. This technical report describes SL in detail and can serve as a manual for new users of SL.

link (url) [BibTex]

link (url) [BibTex]


no image
Learning and generalization of motor skills by learning from demonstration

Pastor, P., Hoffmann, H., Asfour, T., Schaal, S.

In International Conference on Robotics and Automation (ICRA2009), Kobe, Japan, May 12-19, 2009, 2009, clmc (inproceedings)

Abstract
We provide a general approach for learning robotic motor skills from human demonstration. To represent an observed movement, a non-linear differential equation is learned such that it reproduces this movement. Based on this representation, we build a library of movements by labeling each recorded movement according to task and context (e.g., grasping, placing, and releasing). Our differential equation is formulated such that generalization can be achieved simply by adapting a start and a goal parameter in the equation to the desired position values of a movement. For object manipulation, we present how our framework extends to the control of gripper orientation and finger position. The feasibility of our approach is demonstrated in simulation as well as on a real robot. The robot learned a pick-and-place operation and a water-serving task and could generalize these tasks to novel situations.

link (url) [BibTex]

link (url) [BibTex]


no image
Compliant quadruped locomotion over rough terrain

Buchli, J., Kalakrishnan, M., Mistry, M., Pastor, P., Schaal, S.

In Intelligent Robots and Systems, 2009. IROS 2009. IEEE/RSJ International Conference on, pages: 814-820, 2009, clmc (inproceedings)

Abstract
Many critical elements for statically stable walking for legged robots have been known for a long time, including stability criteria based on support polygons, good foothold selection, recovery strategies to name a few. All these criteria have to be accounted for in the planning as well as the control phase. Most legged robots usually employ high gain position control, which means that it is crucially important that the planned reference trajectories are a good match for the actual terrain, and that tracking is accurate. Such an approach leads to conservative controllers, i.e. relatively low speed, ground speed matching, etc. Not surprisingly such controllers are not very robust - they are not suited for the real world use outside of the laboratory where the knowledge of the world is limited and error prone. Thus, to achieve robust robotic locomotion in the archetypical domain of legged systems, namely complex rough terrain, where the size of the obstacles are in the order of leg length, additional elements are required. A possible solution to improve the robustness of legged locomotion is to maximize the compliance of the controller. While compliance is trivially achieved by reduced feedback gains, for terrain requiring precise foot placement (e.g. climbing rocks, walking over pegs or cracks) compliance cannot be introduced at the cost of inferior tracking. Thus, model-based control and - in contrast to passive dynamic walkers - active balance control is required. To achieve these objectives, in this paper we add two crucial elements to legged locomotion, i.e., floating-base inverse dynamics control and predictive force control, and we show that these elements increase robustness in face of unknown and unanticipated perturbations (e.g. obstacles). Furthermore, we introduce a novel line-based COG trajectory planner, which yields a simpler algorithm than traditional polygon based methods and creates the appropriate input to our control system.We show results from bot- h simulation and real world of a robotic dog walking over non-perceived obstacles and rocky terrain. The results prove the effectivity of the inverse dynamics/force controller. The presented results show that we have all elements needed for robust all-terrain locomotion, which should also generalize to other legged systems, e.g., humanoid robots.

link (url) [BibTex]

link (url) [BibTex]


no image
Incorporating Muscle Activation-Contraction dynamics to an optimal control framework for finger movements

Theodorou, Evangelos A., Valero-Cuevas, Francisco J.

Abstracts of Neural Control of Movement Conference (NCM 2009), 2009, clmc (article)

Abstract
Recent experimental and theoretical work [1] investigated the neural control of contact transition between motion and force during tapping with the index finger as a nonlinear optimization problem. Such transitions from motion to well-directed contact force are a fundamental part of dexterous manipulation. There are 3 alternative hypotheses of how this transition could be accomplished by the nervous system as a function of changes in direction and magnitude of the torque vector controlling the finger. These hypotheses are 1) an initial change in direction with a subsequent change in magnitude of the torque vector; 2) an initial change in magnitude with a subsequent directional change of the torque vector; and 3) a simultaneous and proportionally equal change of both direction and magnitude of the torque vector. Experimental work in [2] shows that the nervous system selects the first strategy, and in [1] we suggest that this may in fact be the optimal strategy. In [4] the framework of Iterative Linear Quadratic Optimal Regulator (ILQR) was extended to incorporate motion and force control. However, our prior simulation work assumed direct and instantaneous control of joint torques, which ignores the known delays and filtering properties of skeletal muscle. In this study, we implement an ILQR controller for a more biologically plausible biomechanical model of the index finger than [4], and add activation-contraction dynamics to the system to simulate muscle function. The planar biomechanical model includes the kinematics of the 3 joints while the applied torques are driven by activation?contraction dynamics with biologically plausible time constants [3]. In agreement with our experimental work [2], the task is to, within 500 ms, move the finger from a given resting configuration to target configuration with a desired terminal velocity. ILQR does not only stabilize the finger dynamics according to the objective function, but it also generates smooth joint space trajectories with minimal tuning and without an a-priori initial control policy (which is difficult to find for highly dimensional biomechanical systems). Furthemore, the use of this optimal control framework and the addition of activation-contraction dynamics considers the full nonlinear dynamics of the index finger and produces a sequence of postures which are compatible with experimental motion data [2]. These simulations combined with prior experimental results suggest that optimal control is a strong candidate for the generation of finger movements prior to abrupt motion-to-force transitions. This work is funded in part by grants NIH R01 0505520 and NSF EFRI-0836042 to Dr. Francisco J. Valero- Cuevas 1 Venkadesan M, Valero-Cuevas FJ. 
Effects of neuromuscular lags on controlling contact transitions. 
Philosophical Transactions of the Royal Society A: 2008. 2 Venkadesan M, Valero-Cuevas FJ. 
Neural Control of Motion-to-Force Transitions with the Fingertip. 
J. Neurosci., Feb 2008; 28: 1366 - 1373; 3 Zajac. Muscle and tendon: properties, models, scaling, and application to biomechanics and motor control. Crit Rev Biomed Eng, 17 4. Weiwei Li., Francisco Valero Cuevas: ?Linear Quadratic Optimal Control of Contact Transition with Fingertip ? ACC 2009

PDF [BibTex]

PDF [BibTex]


no image
Inertial parameter estimation of floating-base humanoid systems using partial force sensing

Mistry, M., Schaal, S., Yamane, K.

In IEEE-RAS International Conference on Humanoid Robots (Humanoids 2009), Paris, Dec.7-10, 2009, clmc (inproceedings)

Abstract
Recently, several controllers have been proposed for humanoid robots which rely on full-body dynamic models. The estimation of inertial parameters from data is a critical component for obtaining accurate models for control. However, floating base systems, such as humanoid robots, incur added challenges to this task (e.g. contact forces must be measured, contact states can change, etc.) In this work, we outline a theoretical framework for whole body inertial parameter estimation, including the unactuated floating base. Using a least squares minimization approach, conducted within the nullspace of unmeasured degrees of freedom, we are able to use a partial force sensor set for full-body estimation, e.g. using only joint torque sensors, allowing for estimation when contact force measurement is unavailable or unreliable (e.g. due to slipping, rolling contacts, etc.). We also propose how to determine the theoretical minimum force sensor set for full body estimation, and discuss the practical limitations of doing so.

link (url) [BibTex]

link (url) [BibTex]


no image
On-line learning and modulation of periodic movements with nonlinear dynamical systems

Gams, A., Ijspeert, A., Schaal, S., Lenarčič, J.

Autonomous Robots, 27(1):3-23, 2009, clmc (article)

Abstract
Abstract  The paper presents a two-layered system for (1) learning and encoding a periodic signal without any knowledge on its frequency and waveform, and (2) modulating the learned periodic trajectory in response to external events. The system is used to learn periodic tasks on a humanoid HOAP-2 robot. The first layer of the system is a dynamical system responsible for extracting the fundamental frequency of the input signal, based on adaptive frequency oscillators. The second layer is a dynamical system responsible for learning of the waveform based on a built-in learning algorithm. By combining the two dynamical systems into one system we can rapidly teach new trajectories to robots without any knowledge of the frequency of the demonstration signal. The system extracts and learns only one period of the demonstration signal. Furthermore, the trajectories are robust to perturbations and can be modulated to cope with a dynamic environment. The system is computationally inexpensive, works on-line for any periodic signal, requires no additional signal processing to determine the frequency of the input signal and can be applied in parallel to multiple dimensions. Additionally, it can adapt to changes in frequency and shape, e.g. to non-stationary signals, such as hand-generated signals and human demonstrations.

link (url) [BibTex]

link (url) [BibTex]

2008


no image
Learning to control in operational space

Peters, J., Schaal, S.

International Journal of Robotics Research, 27, pages: 197-212, 2008, clmc (article)

Abstract
One of the most general frameworks for phrasing control problems for complex, redundant robots is operational space control. However, while this framework is of essential importance for robotics and well-understood from an analytical point of view, it can be prohibitively hard to achieve accurate control in face of modeling errors, which are inevitable in com- plex robots, e.g., humanoid robots. In this paper, we suggest a learning approach for opertional space control as a direct inverse model learning problem. A first important insight for this paper is that a physically cor- rect solution to the inverse problem with redundant degrees-of-freedom does exist when learning of the inverse map is performed in a suitable piecewise linear way. The second crucial component for our work is based on the insight that many operational space controllers can be understood in terms of a constrained optimal control problem. The cost function as- sociated with this optimal control problem allows us to formulate a learn- ing algorithm that automatically synthesizes a globally consistent desired resolution of redundancy while learning the operational space controller. From the machine learning point of view, this learning problem corre- sponds to a reinforcement learning problem that maximizes an immediate reward. We employ an expectation-maximization policy search algorithm in order to solve this problem. Evaluations on a three degrees of freedom robot arm are used to illustrate the suggested approach. The applica- tion to a physically realistic simulator of the anthropomorphic SARCOS Master arm demonstrates feasibility for complex high degree-of-freedom robots. We also show that the proposed method works in the setting of learning resolved motion rate control on real, physical Mitsubishi PA-10 medical robotics arm.

link (url) DOI [BibTex]

2008

link (url) DOI [BibTex]


no image
Adaptation to a sub-optimal desired trajectory

M. Mistry, E. A. G. L. T. Y. S. S. M. K.

Advances in Computational Motor Control VII, Symposium at the Society for Neuroscience Meeting, Washington DC, 2008, 2008, clmc (article)

PDF [BibTex]

PDF [BibTex]


no image
Human movement generation based on convergent flow fields: A computational model and a behavioral experiment

Hoffmann, H., Schaal, S.

In Advances in Computational Motor Control VII, Symposium at the Society for Neuroscience Meeting, Washington DC, 2008, 2008, clmc (inproceedings)

link (url) [BibTex]

link (url) [BibTex]


no image
Operational space control: A theoretical and emprical comparison

Nakanishi, J., Cory, R., Mistry, M., Peters, J., Schaal, S.

International Journal of Robotics Research, 27(6):737-757, 2008, clmc (article)

Abstract
Dexterous manipulation with a highly redundant movement system is one of the hallmarks of hu- man motor skills. From numerous behavioral studies, there is strong evidence that humans employ compliant task space control, i.e., they focus control only on task variables while keeping redundant degrees-of-freedom as compliant as possible. This strategy is robust towards unknown disturbances and simultaneously safe for the operator and the environment. The theory of operational space con- trol in robotics aims to achieve similar performance properties. However, despite various compelling theoretical lines of research, advanced operational space control is hardly found in actual robotics imple- mentations, in particular new kinds of robots like humanoids and service robots, which would strongly profit from compliant dexterous manipulation. To analyze the pros and cons of different approaches to operational space control, this paper focuses on a theoretical and empirical evaluation of different methods that have been suggested in the literature, but also some new variants of operational space controllers. We address formulations at the velocity, acceleration and force levels. First, we formulate all controllers in a common notational framework, including quaternion-based orientation control, and discuss some of their theoretical properties. Second, we present experimental comparisons of these approaches on a seven-degree-of-freedom anthropomorphic robot arm with several benchmark tasks. As an aside, we also introduce a novel parameter estimation algorithm for rigid body dynamics, which ensures physical consistency, as this issue was crucial for our successful robot implementations. Our extensive empirical results demonstrate that one of the simplified acceleration-based approaches can be advantageous in terms of task performance, ease of parameter tuning, and general robustness and compliance in face of inevitable modeling errors.

link (url) [BibTex]

link (url) [BibTex]


no image
Movement reproduction and obstacle avoidance with dynamic movement primitives and potential fields

Park, D., Hoffmann, H., Pastor, P., Schaal, S.

In IEEE International Conference on Humanoid Robots, 2008., 2008, clmc (inproceedings)

PDF [BibTex]

PDF [BibTex]


no image
The dual role of uncertainty in force field learning

Mistry, M., Theodorou, E., Hoffmann, H., Schaal, S.

In Abstracts of the Eighteenth Annual Meeting of Neural Control of Movement (NCM), Naples, Florida, April 29-May 4, 2008, clmc (inproceedings)

Abstract
Force field experiments have been a successful paradigm for studying the principles of planning, execution, and learning in human arm movements. Subjects have been shown to cope with the disturbances generated by force fields by learning internal models of the underlying dynamics to predict disturbance effects or by increasing arm impedance (via co-contraction) if a predictive approach becomes infeasible. Several studies have addressed the issue uncertainty in force field learning. Scheidt et al. demonstrated that subjects exposed to a viscous force field of fixed structure but varying strength (randomly changing from trial to trial), learn to adapt to the mean disturbance, regardless of the statistical distribution. Takahashi et al. additionally show a decrease in strength of after-effects after learning in the randomly varying environment. Thus they suggest that the nervous system adopts a dual strategy: learning an internal model of the mean of the random environment, while simultaneously increasing arm impedance to minimize the consequence of errors. In this study, we examine what role variance plays in the learning of uncertain force fields. We use a 7 degree-of-freedom exoskeleton robot as a manipulandum (Sarcos Master Arm, Sarcos, Inc.), and apply a 3D viscous force field of fixed structure and strength randomly selected from trial to trial. Additionally, in separate blocks of trials, we alter the variance of the randomly selected strength multiplier (while keeping a constant mean). In each block, after sufficient learning has occurred, we apply catch trials with no force field and measure the strength of after-effects. As expected in higher variance cases, results show increasingly smaller levels of after-effects as the variance is increased, thus implying subjects choose the robust strategy of increasing arm impedance to cope with higher levels of uncertainty. Interestingly, however, subjects show an increase in after-effect strength with a small amount of variance as compared to the deterministic (zero variance) case. This result implies that a small amount of variability aides in internal model formation, presumably a consequence of the additional amount of exploration conducted in the workspace of the task.

[BibTex]

[BibTex]


no image
Dynamic movement primitives for movement generation motivated by convergent force fields in frog

Hoffmann, H., Pastor, P., Schaal, S.

In Adaptive Motion of Animals and Machines (AMAM), 2008, clmc (inproceedings)

PDF [BibTex]

PDF [BibTex]


no image
Efficient inverse kinematics algorithms for highdimensional movement systems

Tevatia, G., Schaal, S.

CLMC Technical Report: TR-CLMC-2008-1, 2008, clmc (techreport)

Abstract
Real-time control of the endeffector of a humanoid robot in external coordinates requires computationally efficient solutions of the inverse kinematics problem. In this context, this paper investigates methods of resolved motion rate control (RMRC) that employ optimization criteria to resolve kinematic redundancies. In particular we focus on two established techniques, the pseudo inverse with explicit optimization and the extended Jacobian method. We prove that the extended Jacobian method includes pseudo-inverse methods as a special solution. In terms of computational complexity, however, pseudo-inverse and extended Jacobian differ significantly in favor of pseudo-inverse methods. Employing numerical estimation techniques, we introduce a computationally efficient version of the extended Jacobian with performance comparable to the original version. Our results are illustrated in simulation studies with a multiple degree-offreedom robot, and were evaluated on an actual 30 degree-of-freedom full-body humanoid robot.

link (url) [BibTex]

link (url) [BibTex]


no image
Behavioral experiments on reinforcement learning in human motor control

Hoffmann, H., Theodorou, E., Schaal, S.

In Abstracts of the Eighteenth Annual Meeting of Neural Control of Movement (NCM), Naples, Florida, April 29-May 4, 2008, clmc (inproceedings)

Abstract
Reinforcement learning (RL) - learning solely based on reward or cost feedback - is widespread in robotics control and has been also suggested as computational model for human motor control. In human motor control, however, hardly any experiment studied reinforcement learning. Here, we study learning based on visual cost feedback in a reaching task and did three experiments: (1) to establish a simple enough experiment for RL, (2) to study spatial localization of RL, and (3) to study the dependence of RL on the cost function. In experiment (1), subjects sit in front of a drawing tablet and look at a screen onto which the drawing pen's position is projected. Beginning from a start point, their task is to move with the pen through a target point presented on screen. Visual feedback about the pen's position is given only before movement onset. At the end of a movement, subjects get visual feedback only about the cost of this trial. We choose as cost the squared distance between target and virtual pen position at the target line. Above a threshold value, the cost was fixed at this value. In the mapping of the pen's position onto the screen, we added a bias (unknown to subject) and Gaussian noise. As result, subjects could learn the bias, and thus, showed reinforcement learning. In experiment (2), we randomly altered the target position between three different locations (three different directions from start point: -45, 0, 45). For each direction, we chose a different bias. As result, subjects learned all three bias values simultaneously. Thus, RL can be spatially localized. In experiment (3), we varied the sensitivity of the cost function by multiplying the squared distance with a constant value C, while keeping the same cut-off threshold. As in experiment (2), we had three target locations. We assigned to each location a different C value (this assignment was randomized between subjects). Since subjects learned the three locations simultaneously, we could directly compare the effect of the different cost functions. As result, we found an optimal C value; if C was too small (insensitive cost), learning was slow; if C was too large (narrow cost valley), the exploration time was longer and learning delayed. Thus, reinforcement learning in human motor control appears to be sen

[BibTex]

[BibTex]


no image
Movement generation by learning from demonstration and generalization to new targets

Pastor, P., Hoffmann, H., Schaal, S.

In Adaptive Motion of Animals and Machines (AMAM), 2008, clmc (inproceedings)

PDF [BibTex]

PDF [BibTex]


no image
Combining dynamic movement primitives and potential fields for online obstacle avoidance

Park, D., Hoffmann, H., Schaal, S.

In Adaptive Motion of Animals and Machines (AMAM), Cleveland, Ohio, 2008, 2008, clmc (inproceedings)

link (url) [BibTex]

link (url) [BibTex]


no image
A library for locally weighted projection regression

Klanke, S., Vijayakumar, S., Schaal, S.

Journal of Machine Learning Research, 9, pages: 623-626, 2008, clmc (article)

Abstract
In this paper we introduce an improved implementation of locally weighted projection regression (LWPR), a supervised learning algorithm that is capable of handling high-dimensional input data. As the key features, our code supports multi-threading, is available for multiple platforms, and provides wrappers for several programming languages.

link (url) [BibTex]

link (url) [BibTex]


no image
Computational model for movement learning under uncertain cost

Theodorou, E., Hoffmann, H., Mistry, M., Schaal, S.

In Abstracts of the Society of Neuroscience Meeting (SFN 2008), Washington, DC 2008, 2008, clmc (inproceedings)

Abstract
Stochastic optimal control is a framework for computing control commands that lead to an optimal behavior under a given cost. Despite the long history of optimal control in engineering, it has been only recently applied to describe human motion. So far, stochastic optimal control has been mainly used in tasks that are already learned, such as reaching to a target. For learning, however, there are only few cases where optimal control has been applied. The main assumptions of stochastic optimal control that restrict its application to tasks after learning are the a priori knowledge of (1) a quadratic cost function (2) a state space model that captures the kinematics and/or dynamics of musculoskeletal system and (3) a measurement equation that models the proprioceptive and/or exteroceptive feedback. Under these assumptions, a sequence of control gains is computed that is optimal with respect to the prespecified cost function. In our work, we relax the assumption of the a priori known cost function and provide a computational framework for modeling tasks that involve learning. Typically, a cost function consists of two parts: one part that models the task constraints, like squared distance to goal at movement endpoint, and one part that integrates over the squared control commands. In learning a task, the first part of this cost function will be adapted. We use an expectation-maximization scheme for learning: the expectation step optimizes the task constraints through gradient descent of a reward function and the maximizing step optimizes the control commands. Our computational model is tested and compared with data given from a behavioral experiment. In this experiment, subjects sit in front of a drawing tablet and look at a screen onto which the drawing-pen's position is projected. Beginning from a start point, their task is to move with the pen through a target point presented on screen. Visual feedback about the pen's position is given only before movement onset. At the end of a movement, subjects get visual feedback only about the cost of this trial. In the mapping of the pen's position onto the screen, we added a bias (unknown to subject) and Gaussian noise. Therefore the cost is a function of this bias. The subjects were asked to reach to the target and minimize this cost over trials. In this behavioral experiment, subjects could learn the bias and thus showed reinforcement learning. With our computational model, we could model the learning process over trials. Particularly, the dependence on parameters of the reward function (Gaussian width) and the modulation of movement variance over time were similar in experiment and model.

[BibTex]

[BibTex]


no image
Optimization strategies in human reinforcement learning

Hoffmann, H., Theodorou, E., Schaal, S.

Advances in Computational Motor Control VII, Symposium at the Society for Neuroscience Meeting, Washington DC, 2008, 2008, clmc (article)

PDF [BibTex]

PDF [BibTex]


no image
A Bayesian approach to empirical local linearizations for robotics

Ting, J., D’Souza, A., Vijayakumar, S., Schaal, S.

In International Conference on Robotics and Automation (ICRA2008), Pasadena, CA, USA, May 19-23, 2008, 2008, clmc (inproceedings)

Abstract
Local linearizations are ubiquitous in the control of robotic systems. Analytical methods, if available, can be used to obtain the linearization, but in complex robotics systems where the the dynamics and kinematics are often not faithfully obtainable, empirical linearization may be preferable. In this case, it is important to only use data for the local linearization that lies within a ``reasonable'' linear regime of the system, which can be defined from the Hessian at the point of the linearization -- a quantity that is not available without an analytical model. We introduce a Bayesian approach to solve statistically what constitutes a ``reasonable'' local regime. We approach this problem in the context local linear regression. In contrast to previous locally linear methods, we avoid cross-validation or complex statistical hypothesis testing techniques to find the appropriate local regime. Instead, we treat the parameters of the local regime probabilistically and use approximate Bayesian inference for their estimation. This approach results in an analytical set of iterative update equations that are easily implemented on real robotics systems for real-time applications. As in other locally weighted regressions, our algorithm also lends itself to complete nonlinear function approximation for learning empirical internal models. We sketch the derivation of our Bayesian method and provide evaluations on synthetic data and actual robot data where the analytical linearization was known.

link (url) [BibTex]

link (url) [BibTex]


no image
Do humans plan continuous trajectories in kinematic coordinates?

Hoffmann, H., Schaal, S.

In Abstracts of the Society of Neuroscience Meeting (SFN 2008), Washington, DC 2008, 2008, clmc (inproceedings)

Abstract
The planning and execution of human arm movements is still unresolved. An ongoing controversy is whether we plan a movement in kinematic coordinates and convert these coordinates with an inverse internal model into motor commands (like muscle activation) or whether we combine a few muscle synergies or equilibrium points to move a hand, e.g., between two targets. The first hypothesis implies that a planner produces a desired end-effector position for all time points; the second relies on the dynamics of the muscular-skeletal system for a given control command to produce a continuous end-effector trajectory. To distinguish between these two possibilities, we use a visuomotor adaptation experiment. Subjects moved a pen on a graphics tablet and observed the pen's mapped position onto a screen (subjects quickly adapted to this mapping). The task was to move a cursor between two points in a given time window. In the adaptation test, we manipulated the velocity profile of the cursor feedback such that the shape of the trajectories remained unchanged (for straight paths). If humans would use a kinematic plan and map at each time the desired end-effector position onto control commands, subjects should adapt to the above manipulation. In a similar experiment, Wolpert et al (1995) showed adaptation to changes in the curvature of trajectories. This result, however, cannot rule out a shift of an equilibrium point or an additional synergy activation between start and end point of a movement. In our experiment, subjects did two sessions, one control without and one with velocity-profile manipulation. To skew the velocity profile of the cursor trajectory, we added to the current velocity, v, the function 0.8*v*cos(pi + pi*x), where x is the projection of the cursor position onto the start-goal line divided by the distance start to goal (x=0 at the start point). As result, subjects did not adapt to this manipulation: for all subjects, the true hand motion was not significantly modified in a direction consistent with adaptation, despite that the visually presented motion differed significantly from the control motion. One may still argue that this difference in motion was insufficient to be processed visually. Thus, as a control experiment, we replayed control and modified motions to the subjects and asked which of the two motions appeared 'more natural'. Subjects chose the unperturbed motion as more natural significantly better than chance. In summary, for a visuomotor transformation task, the hypothesis of a planned continuous end-effector trajectory predicts adaptation to a modified velocity profile. The current experiment found no adaptation under such transformation.

[BibTex]

[BibTex]

2007


no image
Towards Machine Learning of Motor Skills

Peters, J., Schaal, S., Schölkopf, B.

In Proceedings of Autonome Mobile Systeme (AMS), pages: 138-144, (Editors: K Berns and T Luksch), 2007, clmc (inproceedings)

Abstract
Autonomous robots that can adapt to novel situations has been a long standing vision of robotics, artificial intelligence, and cognitive sciences. Early approaches to this goal during the heydays of artificial intelligence research in the late 1980s, however, made it clear that an approach purely based on reasoning or human insights would not be able to model all the perceptuomotor tasks that a robot should fulfill. Instead, new hope was put in the growing wake of machine learning that promised fully adaptive control algorithms which learn both by observation and trial-and-error. However, to date, learning techniques have yet to fulfill this promise as only few methods manage to scale into the high-dimensional domains of manipulator robotics, or even the new upcoming trend of humanoid robotics, and usually scaling was only achieved in precisely pre-structured domains. In this paper, we investigate the ingredients for a general approach to motor skill learning in order to get one step closer towards human-like performance. For doing so, we study two ma jor components for such an approach, i.e., firstly, a theoretically well-founded general approach to representing the required control structures for task representation and execution and, secondly, appropriate learning algorithms which can be applied in this setting.

PDF DOI [BibTex]

2007

PDF DOI [BibTex]


no image
Reinforcement Learning for Optimal Control of Arm Movements

Theodorou, E., Peters, J., Schaal, S.

In Abstracts of the 37st Meeting of the Society of Neuroscience., Neuroscience, 2007, clmc (inproceedings)

Abstract
Every day motor behavior consists of a plethora of challenging motor skills from discrete movements such as reaching and throwing to rhythmic movements such as walking, drumming and running. How this plethora of motor skills can be learned remains an open question. In particular, is there any unifying computa-tional framework that could model the learning process of this variety of motor behaviors and at the same time be biologically plausible? In this work we aim to give an answer to these questions by providing a computational framework that unifies the learning mechanism of both rhythmic and discrete movements under optimization criteria, i.e., in a non-supervised trial-and-error fashion. Our suggested framework is based on Reinforcement Learning, which is mostly considered as too costly to be a plausible mechanism for learning com-plex limb movement. However, recent work on reinforcement learning with pol-icy gradients combined with parameterized movement primitives allows novel and more efficient algorithms. By using the representational power of such mo-tor primitives we show how rhythmic motor behaviors such as walking, squash-ing and drumming as well as discrete behaviors like reaching and grasping can be learned with biologically plausible algorithms. Using extensive simulations and by using different reward functions we provide results that support the hy-pothesis that Reinforcement Learning could be a viable candidate for motor learning of human motor behavior when other learning methods like supervised learning are not feasible.

[BibTex]

[BibTex]


no image
Machine Learning of Motor Skills for Robotics

Peters, J.

University of Southern California, Los Angeles, CA, USA, University of Southern California, Los Angeles, CA, USA, 2007, clmc (phdthesis)

Abstract
Autonomous robots that can assist humans in situations of daily life have been a long standing vision of robotics, artificial intelligence, and cognitive sciences. A first step towards this goal is to create robots that can accomplish a multitude of different tasks, triggered by environmental context or higher level instruction. Early approaches to this goal during the heydays of artificial intelligence research in the late 1980s, however, made it clear that an approach purely based on reasoning and human insights would not be able to model all the perceptuomotor tasks that a robot should fulfill. Instead, new hope was put in the growing wake of machine learning that promised fully adaptive control algorithms which learn both by observation and trial-and-error. However, to date, learning techniques have yet to fulfill this promise as only few methods manage to scale into the high-dimensional domains of manipulator robotics, or even the new upcoming trend of humanoid robotics, and usually scaling was only achieved in precisely pre-structured domains. In this thesis, we investigate the ingredients for a general approach to motor skill learning in order to get one step closer towards human-like performance. For doing so, we study two major components for such an approach, i.e., firstly, a theoretically well-founded general approach to representing the required control structures for task representation and execution and, secondly, appropriate learning algorithms which can be applied in this setting. As a theoretical foundation, we first study a general framework to generate control laws for real robots with a particular focus on skills represented as dynamical systems in differential constraint form. We present a point-wise optimal control framework resulting from a generalization of Gauss' principle and show how various well-known robot control laws can be derived by modifying the metric of the employed cost function. The framework has been successfully applied to task space tracking control for holonomic systems for several different metrics on the anthropomorphic SARCOS Master Arm. In order to overcome the limiting requirement of accurate robot models, we first employ learning methods to find learning controllers for task space control. However, when learning to execute a redundant control problem, we face the general problem of the non-convexity of the solution space which can force the robot to steer into physically impossible configurations if supervised learning methods are employed without further consideration. This problem can be resolved using two major insights, i.e., the learning problem can be treated as locally convex and the cost function of the analytical framework can be used to ensure global consistency. Thus, we derive an immediate reinforcement learning algorithm from the expectation-maximization point of view which leads to a reward-weighted regression technique. This method can be used both for operational space control as well as general immediate reward reinforcement learning problems. We demonstrate the feasibility of the resulting framework on the problem of redundant end-effector tracking for both a simulated 3 degrees of freedom robot arm as well as for a simulated anthropomorphic SARCOS Master Arm. While learning to execute tasks in task space is an essential component to a general framework to motor skill learning, learning the actual task is of even higher importance, particularly as this issue is more frequently beyond the abilities of analytical approaches than execution. We focus on the learning of elemental tasks which can serve as the "building blocks of movement generation", called motor primitives. Motor primitives are parameterized task representations based on splines or nonlinear differential equations with desired attractor properties. While imitation learning of parameterized motor primitives is a relatively well-understood problem, the self-improvement by interaction of the system with the environment remains a challenging problem, tackled in the fourth chapter of this thesis. For pursuing this goal, we highlight the difficulties with current reinforcement learning methods, and outline both established and novel algorithms for the gradient-based improvement of parameterized policies. We compare these algorithms in the context of motor primitive learning, and show that our most modern algorithm, the Episodic Natural Actor-Critic outperforms previous algorithms by at least an order of magnitude. We demonstrate the efficiency of this reinforcement learning method in the application of learning to hit a baseball with an anthropomorphic robot arm. In conclusion, in this thesis, we have contributed a general framework for analytically computing robot control laws which can be used for deriving various previous control approaches and serves as foundation as well as inspiration for our learning algorithms. We have introduced two classes of novel reinforcement learning methods, i.e., the Natural Actor-Critic and the Reward-Weighted Regression algorithm. These algorithms have been used in order to replace the analytical components of the theoretical framework by learned representations. Evaluations have been performed on both simulated and real robot arms.

[BibTex]

[BibTex]


no image
Reinforcement learning by reward-weighted regression for operational space control

Peters, J., Schaal, S.

In Proceedings of the 24th Annual International Conference on Machine Learning, pages: 745-750, ICML, 2007, clmc (inproceedings)

Abstract
Many robot control problems of practical importance, including operational space control, can be reformulated as immediate reward reinforcement learning problems. However, few of the known optimization or reinforcement learning algorithms can be used in online learning control for robots, as they are either prohibitively slow, do not scale to interesting domains of complex robots, or require trying out policies generated by random search, which are infeasible for a physical system. Using a generalization of the EM-base reinforcement learning framework suggested by Dayan & Hinton, we reduce the problem of learning with immediate rewards to a reward-weighted regression problem with an adaptive, integrated reward transformation for faster convergence. The resulting algorithm is efficient, learns smoothly without dangerous jumps in solution space, and works well in applications of complex high degree-of-freedom robots.

link (url) DOI [BibTex]

link (url) DOI [BibTex]


no image
Policy gradient methods for machine learning

Peters, J., Theodorou, E., Schaal, S.

In Proceedings of the 14th INFORMS Conference of the Applied Probability Society, pages: 97-98, Eindhoven, Netherlands, July 9-11, 2007, 2007, clmc (inproceedings)

Abstract
We present an in-depth survey of policy gradient methods as they are used in the machine learning community for optimizing parameterized, stochastic control policies in Markovian systems with respect to the expected reward. Despite having been developed separately in the reinforcement learning literature, policy gradient methods employ likelihood ratio gradient estimators as also suggested in the stochastic simulation optimization community. It is well-known that this approach to policy gradient estimation traditionally suffers from three drawbacks, i.e., large variance, a strong dependence on baseline functions and a inefficient gradient descent. In this talk, we will present a series of recent results which tackles each of these problems. The variance of the gradient estimation can be reduced significantly through recently introduced techniques such as optimal baselines, compatible function approximations and all-action gradients. However, as even the analytically obtainable policy gradients perform unnaturally slow, it required the step from ÔvanillaÕ policy gradient methods towards natural policy gradients in order to overcome the inefficiency of the gradient descent. This development resulted into the Natural Actor-Critic architecture which can be shown to be very efficient in application to motor primitive learning for robotics.

[BibTex]

[BibTex]


no image
Less Conservative Polytopic LPV Models for Charge Control by Combining Parameter Set Mapping and Set Intersection

Kwiatkowski, A., Trimpe, S., Werner, H.

In Proceedings of the 46th IEEE Conference on Decision and Control, 2007 (inproceedings)

DOI [BibTex]

DOI [BibTex]


no image
Policy Learning for Motor Skills

Peters, J., Schaal, S.

In Proceedings of 14th International Conference on Neural Information Processing (ICONIP), pages: 233-242, (Editors: Ishikawa, M. , K. Doya, H. Miyamoto, T. Yamakawa), 2007, clmc (inproceedings)

Abstract
Policy learning which allows autonomous robots to adapt to novel situations has been a long standing vision of robotics, artificial intelligence, and cognitive sciences. However, to date, learning techniques have yet to fulfill this promise as only few methods manage to scale into the high-dimensional domains of manipulator robotics, or even the new upcoming trend of humanoid robotics, and usually scaling was only achieved in precisely pre-structured domains. In this paper, we investigate the ingredients for a general approach policy learning with the goal of an application to motor skill refinement in order to get one step closer towards human-like performance. For doing so, we study two major components for such an approach, i.e., firstly, we study policy learning algorithms which can be applied in the general setting of motor skill learning, and, secondly, we study a theoretically well-founded general approach to representing the required control structures for task representation and execution.

PDF DOI [BibTex]

PDF DOI [BibTex]


no image
Reinforcement learning for operational space control

Peters, J., Schaal, S.

In Proceedings of the 2007 IEEE International Conference on Robotics and Automation, pages: 2111-2116, IEEE Computer Society, ICRA, 2007, clmc (inproceedings)

Abstract
While operational space control is of essential importance for robotics and well-understood from an analytical point of view, it can be prohibitively hard to achieve accurate control in face of modeling errors, which are inevitable in complex robots, e.g., humanoid robots. In such cases, learning control methods can offer an interesting alternative to analytical control algorithms. However, the resulting supervised learning problem is ill-defined as it requires to learn an inverse mapping of a usually redundant system, which is well known to suffer from the property of non-convexity of the solution space, i.e., the learning system could generate motor commands that try to steer the robot into physically impossible configurations. The important insight that many operational space control algorithms can be reformulated as optimal control problems, however, allows addressing this inverse learning problem in the framework of reinforcement learning. However, few of the known optimization or reinforcement learning algorithms can be used in online learning control for robots, as they are either prohibitively slow, do not scale to interesting domains of complex robots, or require trying out policies generated by random search, which are infeasible for a physical system. Using a generalization of the EM-based reinforcement learning framework suggested by Dayan & Hinton, we reduce the problem of learning with immediate rewards to a reward-weighted regression problem with an adaptive, integrated reward transformation for faster convergence. The resulting algorithm is efficient, learns smoothly without dangerous jumps in solution space, and works well in applications of complex high degree-of-freedom robots.

link (url) DOI [BibTex]

link (url) DOI [BibTex]


no image
Relative Entropy Policy Search

Peters, J.

CLMC Technical Report: TR-CLMC-2007-2, Computational Learning and Motor Control Lab, Los Angeles, CA, 2007, clmc (techreport)

Abstract
This technical report describes a cute idea of how to create new policy search approaches. It directly relates to the Natural Actor-Critic methods but allows the derivation of one shot solutions. Future work may include the application to interesting problems.

PDF link (url) [BibTex]

PDF link (url) [BibTex]


no image
Using reward-weighted regression for reinforcement learning of task space control

Peters, J., Schaal, S.

In Proceedings of the 2007 IEEE International Symposium on Approximate Dynamic Programming and Reinforcement Learning, pages: 262-267, Honolulu, Hawaii, April 1-5, 2007, 2007, clmc (inproceedings)

Abstract
In this paper, we evaluate different versions from the three main kinds of model-free policy gradient methods, i.e., finite difference gradients, `vanilla' policy gradients and natural policy gradients. Each of these methods is first presented in its simple form and subsequently refined and optimized. By carrying out numerous experiments on the cart pole regulator benchmark we aim to provide a useful baseline for future research on parameterized policy search algorithms. Portable C++ code is provided for both plant and algorithms; thus, the results in this paper can be reevaluated, reused and new algorithms can be inserted with ease.

link (url) DOI [BibTex]

link (url) DOI [BibTex]