Uncertainties in contact dynamics and object geometry remain significant barriers to robust robotic manipulation. Caging helps mitigate these uncertainties by constraining an object’s mobility without requiring precise contact modeling. Existing caging research often treats morphology and policy optimization as separate problems, overlooking their synergy. In this paper, we introduce CageCoOpt, a hierarchical framework that jointly optimizes manipulator morphology and control policy for robust caging-based manipulation. The framework employs reinforcement learning for policy optimization at the lower level and multitask Bayesian optimization for morphology optimization at the upper level. We incorporate a caging metric into both optimization levels to encourage caging configurations and thereby improve manipulation robustness. The evaluation consists of four manipulation tasks and demonstrates that co-optimizing morphology and policy improves task performance under uncertainties, establishing caging-guided co-optimization as a viable approach for robust manipulation.

The integration of variable stiffness actuators (VSA) in robotic systems endows them with intrinsic flexibility and therefore robustness to unknown disturbances. However, this characteristic presents a challenge: choosing the best intrinsic stiffness setting guaranteeing the required force ap-plication capability while keeping the system as adaptable to uncertainties as possible. This paper proposes a method to set the optimal stiffness for a multi-finger VSA hand to perform a desired manipulation task. The task is generically represented as a force (with unknown magnitude) applied along a reference direction. According to the force application's direction and the hand's kinematic state, the fingers assume a certain role to split the collective force application. We employ the endpoint stiffness ellipsoid to analyze the required finger stiffness to fulfill the task. We evaluate the optimized stiffness settings in a door opening application with an iterative adaption of the stiffness behavior to handle the unknown force requirement. The results show a successful collective behavior of the fingers, where the stiffness setting considers a task-oriented force-adaptability trade-off and effective use of independent VSA fingers.

Although instance-aware perception is a key prerequisite for many autonomous robotic applications, most of the methods only partially solve the problem by focusing solely on known object categories. However, for robots interacting in dynamic and cluttered environments, this is not realistic and severely limits the range of potential applications. Therefore, we propose a novel object instance segmentation approach that does not require any semantic or geometric information of the objects beforehand. In contrast to existing works, we do not explicitly use depth data as input, but rely on the insight that slight viewpoint changes, which for example are provided by stereo image pairs, are often sufficient to determine object boundaries and thus to segment objects. Focusing on the versatility of stereo sensors, we employ a transformer-based architecture that maps directly from the pair of input images to the object instances. This has the major advantage that instead of a noisy, and potentially incomplete depth map as an input, on which the segmentation is computed, we use the original image pair to infer the object instances and a dense depth map. In experiments in several different application domains, we show that our Instance Stereo Transformer (INSTR) algorithm outperforms current state-of-the-art methods that are based on depth maps. Training code and pretrained models are available at https://github.com/DLR-RM/instr.

This paper deals with the problem of planning grasp strategies on constrained and cluttered scenarios. The planner sequences the objects for grasping by considering multiple factors: (i) possible environmental constraints that can be exploited to grasp an object, (ii) object neighborhood, (iii) capability of the arm, and (iv) confidence score of the vision algorithm. To successfully exploit the environmental constraints, this work uses the CLASH hand, a compliant hand that can vary its passive stiffness. The hand can be softened such that it can comply with the object shape, or it can be stiffened to pierce between the objects in clutter. A stiffness decision tree is introduced to choose the best stiffness setting for each particular scenario. In highly cluttered scenarios, a finger position planner is used to find a suitable orientation for the hand such that the fingers can slide in the free regions around the object. Thus, the grasp strategy planner predicts not only the sequence in which the objects can be grasped, but also the required stiffness of the end effector, and the appropriate positions for the fingers around the object. Different experiments are carried out in the context of grocery handling to test the performance of the planner in scenarios that require different grasping strategies.

This paper presents the first two members of the new generation of CLASH hands, which exploit low cost actuation and rapid prototyping to create antagonistic modular and lightweight hands and grippers. The hands approach the robustness of the DLR Awiwi hand with a much lower complexity and cost. To reduce the number of required actuators, a differential coupling mechanism for underactuated fingers was developed, along with a new mechanism that uses variable stiffness actuation in order to increase the workspace of underactuated fingers. The hands provide a research platform for both hand-in-hand and robotic grasping. Design aspects are discussed, and an initial experimental validation verifies the hands' performance.

We address the operation of robotic legs with intrinsic elasticity in hopping cycles determined by the mechanical resonant properties of the system. This ensures energy efficiency and high jumping velocity and distance. Recently, we have shown in simulation that a simple, biologically inspired bang-bang controller operating in the local coordinate of the first resonant mode leads to limit cycles which are robust with respect to leg model uncertainties and ground properties. In this paper we address the velocity control of the hopping and the planning of the bang-bang control parameters for the case that the systems should not move at steady state velocity, but should have different step lengths and heights. We exploit the discrete structure and the small number of parameters of the controller to develop a fast optimization procedure for generating an arbitrary sequence of steps. This approach can provide high motion performance, robustness and substantial computational time saving compared to continuous trajectory and controller gain planning. The stationary and the aperiodic hopping is validated by experiments on a new planar elastic leg.

This video presents the recent upgrades of the mobile humanoid Agile Justin, bringing it closer to an ideal platform for research in autonomous manipulation. Significant upgrades have been made in the fields of mechatronics, 3D sensors, tactile skin, massive GPGPU based computing power, and software communication framework. In addition, first algorithms and two experimental scenarios are presented that take advantage of these new capabilities.

This paper discusses tendon friction effects regarding guiding and material selection. In order to extract valuable information for designers of tendon driven systems, several experiments are conducted to investigate e. g. the intrinsic friction or sliding effect. The results are used to build an experimental friction model and to derive a set of guidelines. The mechanical designer can use the proposed models to anticipate the friction for a given tendon path. Additionally, the guidelines help the mechanical designer to systematically verify the numerous constraints involved in the design process of a tendon driven system.

Intrinsically elastic robots, which technically implement some key characteristics of the human muskoskeletal system, have become a major research topic in nowadays robotics. These novel devices open up entirely new control approaches. They base on temporary storage of potential energy and its timed transformation into kinetic energy. In legged locomotion, such considerations have been a common tool for unveiling the respective fundamental physical processes. However, in arm control, elasticities were typically considered parasitic. In this video we outline our efforts in exploiting the inherent capabilities of intrinsically elastic robots in order to bring them closer to human performance. Instead of applying purely kinematic learing-by-demonstration approaches, which are certainly suboptimal, we argue for using model based techniques in order to optimally exploit the system dynamics such that highly dynamic motion and manipulation capabilities can be achieved. In particular, the explicit use of elasticities as temporary energy tanks can be fully exploited, if they are modeled adequately as an integral part of the mechanism. We also believe that such approaches can substantially contribute to the understanding of human motion biomechanics.

The mobile humanoid Rollin'Justin is a versatile experimental platform for research in manipulation tasks. Previously, different state of the art control methods and first autonomous task execution scenarios have been demonstrated. In this video two new applications with challenging task requirements are presented. One is the catching of one or even two flying balls using all of Justin's degrees of freedom. The other is the autonomous preparation of coffee. Both applications need adequate sensors to support local referencing. The required precision in position and timing is realized in software, using the sensor information, taking the varying precision of Justin's kinematic sub-chains into account and handling all timings in sub-millisecond range.

An anthropomorphic hand arm system using variable stiffness actuation has been developed at DLR. It is aimed to reach its human archetype regarding size, weight and performance. The main focus of our development is put on robustness, dynamic performance and dexterity. Therefore, a paradigm change from impedance controlled, but mechanically stiff joints to robots using intrinsic variable compliance joints is carried out. Collisions of the rigid joint robot at high speeds with stiff objects induce the energy too fast for an active controller to prevent damages. In contrast, passively compliant robots are able to temporarily store energy. In this case the resulting internal forces applied to the robot structure and the drive trains are reduced. Furthermore, the energy storage allows to outperform the dynamics of stiff robots. The hand drives and the electronics are completely integrated within the forearm. Extremely miniaturized electronics have been developed to drive the 52 motors of the system and interface their sensors. Several variable stiffness actuation principles used in the arm joints and the hand are presented. The paper highlights the different requirements that they have to fulfill. A first test of the systems robustness and dynamics has been performed by driving nails with a grasped hammer and is demonstrated in the attached video.

A large workspace and proper force capabilities of a robotic thumb can be obtained using a tensegrity structure for the actuation, similar to the human thumb base muscles. Using nonlinear stiffness elements and an antagonistic architecture, the joint stiffness can be adjusted by variation of the tendon pre-tension. However, the highly nonlinear actuation creates new control challenges and in particular the nonlinear tendon kinematics must be accounted for. Despite the challenges, the nonlinear structure is required to achieve the desired torques. In this paper, the dynamic equations of a tendon driven thumb are established. An efficient formulation is proposed to generate the pretension forces in order to preserve the torques and approximate the stiffness matrix. A cascaded structure is used for the controller. The equations for the inner tendon force control loop and the outer impedance control loop are presented. Because of the absence of link side position sensors, an iterative estimation algorithm is proposed and implemented in real-time. It is shown that, using the mechanical joint flexibility, the controller impedance gain can be adjusted to improve the steady-state effective impedance. The search algorithm robustness is evaluated through a set of simulations. Finally, experimental results and equivalent simulations demonstrate the effectiveness of our controller.

In robotic hands design tendon driven systems have been considered for years. The main advantage is a small end effector inertia e.g. a light, small hand with high dynamics due to remote actuators. To protect the actuators from impact in unknown environments a compliant mechanism can be used. It absorbs energy during an impact or saves energy to enhance the joint dynamics. In this paper an antagonistic tendon mechanism is presented. It fits 38 times in the DLR Hand Arm System forearm and enables is adapted to the different finger joints and different tendon lengths. A magnetic sensor was developed for the force measurement of the tendons. Finally, the calibration and the robustness are demonstrated through a set of experiments.

The DLR Hand Arm System is based upon the variable stiffness concept which has been recently developed to improve impact robustness and energy efficiency of modern robots. This paper continues the work on the bidirectional antagonistic variable stiffness (BAVS) joint concept which is an extension of antagonistic joints. Three mechanical setups utilizing different spring and cam disc combinations to implement a desired torque-stiffness characteristic are analyzed. Two BAVS joint solutions as used for the wrist and forearm rotation of the DLR Hand Arm System are presented. Furthermore in the experimental section torque-deflection calibration and drive redundancy are validated.

The variable stiffness actuation concept is considered to provide a human-friendly robot technology. This paper examines a joint concept called the bidirectional antagonistic joint which is a extension of antagonistic joints. A new operating mode called the helping mode is introduced, which increases the joint load range. Although the joint can not be pretensioned in the helping mode, it is shown that a stiffness variation is possible, assuming a suitable torque-stiffness characteristic of the elastic elements. A methodology to design such characteristics is presented along with several example cases interpreted in a torque-stiffness plot. Furthermore, a stiffness adaptation control scheme which ensures mechanism safety is described. Finally, the design methodology and the control are evaluated on an implementation of a bidirectional antagonistic joint.