The humanoid robot LOLA walks over uneven terrain without using visual information TUM Chair of Applied Mechanics 2019-03-21 | This video shows how the humanoid robot LOLA walks over uneven terrain without using any vision-based information. This reactive robustness to unknown terrain is achieved by a ground-force control scheme that handles ground-height variations and unplanned partial footholds with the environment.A preprint of the paper describing the force-control scheme is available online:https://mediatum.ub.tum.de/doc/1482152/1482152.pdfAn additional video to the development of the force-control scheme is available here: youtu.be/ifuD-ETTi_IFor more information please see our project's website: https://www.mw.tum.de/en/am/research/current-projects/robotics/humanoid-robot-lola/
Joint project advances prosthetic evaluation TUM Chair of Applied Mechanics 2024-04-17 | Advances in prosthetic evaluation have been achieved through a joint project with the IPD CAD Lab at the University of Maribor and the Chair of Applied Mechanics at the Technical University of Munich.Exciting possibilities in prosthetics are emerging through additive manufacturing, especially with the advent of 3D printing, which enables personalized and cost-effective production. The IPD CAD Lab focuses on the advancement of 3D printing for lower limb prostheses.A novel testing approach is being pursued to evaluate lower limb prostheses using a robotic arm. In this approach, a cyclic gait motion is simulated by a robotic arm to test durability. Both a prosthetic limb and a 3D printed footprosthesis are mounted on the robot for testing. The gait trajectory for the robot is taken from walking lab data, and the force signals are monitored to detect potential failures.While initially relying on pre-recorded gait data, our collaboration works towards the exploration of Real-Time Hybrid Substructuring (RTHS) for prosthesis testing, a method previously investigated at the Chair of Applied Mechanics. This approach promises more adaptable and versatile testing techniques. First steps have been taken with simple 1-dimensional mass spring damper experiments using the industrial robot. However, further research is needed before fully implementing an RTHS approach in prosthesis testing.In summary, this collaborative effort is making significant strides in prosthesis evaluation through the integration of robotics and innovative testing techniques. Stay tuned for updates on this research journey!
Centralized vs. Decoupled Dual-Arm Planning Taking into Account Path Quality TUM Chair of Applied Mechanics 2023-12-21 | This video presents our work on coordinated dual-arm manipulation. First, we show our decoupled planning approach that uses cubic spline interpolation to generate C2-continuous velocity control commands. Second, we show our method for minimizing the rotational motion of the end-effectors. Third, we benchmark our decoupled planning approach against a centralized planning approach. Both use the method to minimize rotational motions. Fourth, we apply our decoupled planning approach to a real manipulation scenario.
A Mobile Dual-Arm System That Makes Intralogistics More Efficient TUM Chair of Applied Mechanics 2023-12-21 | This video shows the mobile dual-arm system we have developed. The two robot arms perform manipulation tasks during transport, making intralogistics more efficient. We worked on the synchronization of the two robot arms and investigated the impact of the platform dynamics on the control of the two robot arms. Our project partner youtube.com/@fmlTUMunich worked on the autonomous navigation of the mobile platform and developed robust task planning with error recovery strategies.
Motion Planning Framework for Human-Centered Environments (Presentation) TUM Chair of Applied Mechanics 2023-12-21 | ...
Real-Time Predictive Kinematics Control (Presentation) TUM Chair of Applied Mechanics 2023-12-21 | ...
Spherical Cubic Blends: Smooth Trajectory Generation with Quaternions (Presentation) TUM Chair of Applied Mechanics 2023-12-21 | ...
Efficient Zero-Clamped Cubic Splines Based on Analytical Gradients (Presentation) TUM Chair of Applied Mechanics 2023-12-21 | ...
Hierarchical Control of First-Order Differential Kinematics TUM Chair of Applied Mechanics 2023-08-23 | This video shows an experiment and a simulation. Both use our developed first-order hierarchical control approach. However, the experiment uses continuous task repriorization and the simulation uses discrete task repriorization.The experiment demonstrates that our approach with continuous task repriorization leads to continuous joint velocity control commands.The simulation shows that the approach with discrete task repriorization leads to discontinuous joint velocity control commands that can not be executed on the real robot.
Mobile Preassembly Systems with Cooperative Dual-Arm Manipulation TUM Chair of Applied Mechanics 2023-08-23 | This video shows the current state of our mobile preassembly system.We show the autonomous navigation of the mobile base within a logistics facility.We show the coordinated manipulation of the two mounted manipulators in a shared work space.We show the cooperative manipulation when both robots handle a box. We show this scenario only in the laboratory, i.e. the robot arms are not mounted on the mobile base.We show the cooperative manipulation when both robots perform an insertion assembly operation. We show this scenario only in the laboratory, i.e. the robot arms are not mounted on the mobile base.
Spherical Cubic Blends: Smooth Trajectory Generation with Quaternions TUM Chair of Applied Mechanics 2023-08-23 | This video shows a benchmark of three trajectory generation approaches for quaternions. The state-of-the-art methods SQUAD and SPB are modified such that they generate C2-continuous and zero-clamped trajectory profiles. We benchmark these two state-of-the-art approaches with our proposed method Spherical Cubic Blends on a FRANKA EMIKA robot.
Real-Time Predictive Kinematics Control TUM Chair of Applied Mechanics 2023-08-23 | This video benchmarks our Predictive Kinematics Controller (PKC), that uses model predictive control, with two local redundancy resolution approaches: Resolved Motion Rate Control (RMRC) and Automatic Supervisory Control (ASC). In the simulation, the PKC avoids the obstacle with a bigger distance than the local approaches as it early predicts the approaching obstacle. In the experiment, the PKC early predicts the configuration of low manipulability and appropriately updates the nullspace motion. For the PKC, we use a horizon length of 1.5 seconds and a step size of 0.1 seconds for the underlying MPC.
Nonlinear Optimization-Based Path Length Improvement TUM Chair of Applied Mechanics 2023-08-23 | This video shows the benefit of our Path Length Post-Processor (PLPP) when applied to the solution path of a standard RRT path planner. The input path is jerky with long motions and thus energy consuming and not intuitive for human coworkers. Our PLPP decreases the path length and makes the motion more intuitive for human coworkers.
Robotic Framework for Autonomous Assembly TUM Chair of Applied Mechanics 2023-08-23 | This video shows a benchmark of our autonomous robotic assembly framework with a human and an easy-to-implement app-based program. The application scenario is taken from the Robothon Grand Challenge 2021 competition. Our framework can not compete with a human. However, it is faster than the app-based program and also more flexible as it allows a random position of the task board. We show an uncut, but accelerated, sequence of ten task board runs with our framework.
Efficient Zero-Clamped Cubic Splines Based on Analytical Gradients TUM Chair of Applied Mechanics 2023-08-23 | This video shows the benefits of our trajectory generation approach that computes fast cubic splines. The motion duration is optimized while respecting the actuators' limits. In contrast to the four other presented state-of-the-art approaches (IPTP, ISP, TOTG) our computed trajectory profiles respect the actuators' limits throughout the whole trajectory and the motion is less jerky. A FRANKA EMIKA robot is used for the benchmark.
Stereoscopic High Speed Camera Based Operational Modal Analysis using a One-Camera Setup TUM Chair of Applied Mechanics 2022-03-11 | In this study, we demonstrate a very simple setup that enables 3D measurements of a vibrating structure with only one single high-speed camera using a mirror to generate a stereoscopic image. Using a very specific viewing angle and making some simplifying assumptions, the 3D displacement reconstruction is very easy and doesn't require camera calibration procedures as usually used for similiar tasks.The measurement results are used for an output-only modal analysis of the structure which yields mode shapes that look as expected, so the general approach could be validated.The work was presented at the conference IMAC-XL. The content is published in the conference proceedings.
Real-Time Hybrid Substructuring for Shock Applications Considering Effective Actuator Control TUM Chair of Applied Mechanics 2022-03-11 | In this video, Real-Time Hybrid Substructuring (RTHS) is applied to a shock application. Since shock is a short event, actuator control has to be effective. This talk presents the combined use of a minimum phase inverse compensator and Iterative Learning Control for accurate actuator tracking. Furthermore, challenges associated with RTHS testing of impacts are addressed.The work was presented at the conference IMAC-XL. The content is published in the conference proceedings.
Feasibility Study for Testing Prosthetic Feet Using Real-Time Hybrid Substructuring (RTHS) TUM Chair of Applied Mechanics 2022-03-11 | This video describes a proof-of-concept test, where a prosthetic foot is tested using Real-Time Hybrid Substructuring (RTHS). The goal is to investigate the dynamic interplay between an amputee and a prosthetic foot. The amputee is modeled using a modified version of the Virtual Pivot Point (VPP) model. The experimental setup is presented in this talk. The work was presented at the conference IMAC-XL. The content is published here: mdpi.com/2076-3417/11/20/9492
Humanoid Robot LOLA - Vision Guided Autonomous Multi-Contact Locomotion TUM Chair of Applied Mechanics 2022-01-31 | In this video we demonstrate fully autonomous multi-contact locomotion for our humanoid robot LOLA. In contrast to our previous multi-contact videos where contact points for the feet and hands had to be specified manually by the user, this time all contacts are autonomously planned by the robot itself based on the perceived environment. The only input by the user is the desired final goal position (a horizontal position and rotation around the vertical axis). The robot then automatically computes a feasible contact sequence (if possible) and connects the discrete poses with kinematically and dynamically feasible trajectories while considering multi-contact effects (external forces applied at the hands of the robot).Through this experiments we demonstrate the coupling of LOLA's new computer vision (Chair for Computer Aided Medical Procedures & Augmented Reality, TUM) and walking pattern generation (Chair of Applied Mechanics, TUM) systems. All algorithms run onboard an in real-time. The scene is not known to the robot (it has to detect it on its own).For more information on LOLA please see our project's website:https://www.mec.ed.tum.de/en/am/research/current-projects/robotics/humanoid-robot-lola/This work is supported by the German Research Foundation (DFG, project number 407378162).
Humanoid Robot LOLA - Walking Pattern Generation for Autonomous Multi-Contact Locomotion TUM Chair of Applied Mechanics 2022-01-31 | In this video we explain how LOLA's walking pattern generation (=motion planning) for multi-contact locomotion (=additional hand support) works. Moreover, a series of benchmark simulations is shown. The video is divided into the following sections:00:00 Motivation00:36 Overview LOLA01:16 Phase 1: Environment Model Processing01:54 Phase 2: Contact Planning04:02 Phase 3: Motion Generation06:05 Multi-Contact Configurations06:46 Scenario: Right Wall07:08 Scenario: Right Table07:31 Scenario: Corridor07:54 Scenario: Obstacles08:27 Scenario: Ramps08:53 Scenario: Stairs - Full09:20 Scenario: Stairs - Partial09:49 Scenario: Trap10:28 Scenario: Dynamic Replanning - Abort10:47 Scenario: Dynamic Replanning - UpdateFor more information on LOLA please see our project's website:https://www.mec.ed.tum.de/en/am/research/current-projects/robotics/humanoid-robot-lola/This work is supported by the German Research Foundation (DFG, project number 407378162).
Inverse Pendulum Christmas Star TUM Chair of Applied Mechanics 2022-01-31 | A simple pendulum can exhibit a second stable equlibrium when the hanging point moves up and down harmonically. In order to better see this phenomenon, a strobe light is used that only lights up the scene at a certain point of the harmonic movement. These short light flashes (or the short moments in between) cause the black bars in the video as the camera's CMOS image sensor sometimes gets exposed when the light is off.Happy holidays!
LOLA v1.1 - An Upgrade in Hardware and Software Design for Dynamic Multi-Contact Locomotion TUM Chair of Applied Mechanics 2021-07-20 | In this video we present our recent upgrades to the hard- and software of our humanoid robot LOLA. The changes were made in order to make LOLA capable of dynamic multi-contact locomotion (by which we mean additional hand-support during fast walking). This is meant to increase overall stability and robustness against model inaccuracies and disturbances.This talk was held as oral presentation at the 2020 IEEE-RAS International Conference on Humanoid Robots (July 19-21 2021). For details, see the related conference paper "LOLA v1.1 - An Upgrade in Hardware and Software Design for Dynamic Multi-Contact Locomotion" by Philipp Seiwald, Shun-Cheng Wu, Felix Sygulla, Tobias Berninger, Nora-Sophie Staufenberg,Moritz Sattler, Nicolas Neuburger, Daniel Rixen and Federico Tombari.For more information on LOLA please see our project's website: https://www.mw.tum.de/en/am/research/current-projects/robotics/humanoid-robot-lola/This work is supported by the German Research Foundation (DFG, project number 407378162).
Humanoid Robot LOLA - Balancing on Soft and Hard Ground TUM Chair of Applied Mechanics 2021-04-08 | This video shows LOLA balancing on different terrain when being pushed in different directions. The robot is technically blind, not using any camera-based or prior information on the terrain (hard ground is assumed). Technical Details:Height of the foam sheets: 5cmDensity of the foam: 40 kg/m^3Compression hardness of the foam: 6 kPa @ 40% compressionFor more information on LOLA please see our project's website: https://www.mw.tum.de/en/am/research/current-projects/robotics/humanoid-robot-lola/This work is supported by the German Research Foundation (DFG, project number 407378162).
Humanoid Robot LOLA – Dynamic Multi-Contact Locomotion TUM Chair of Applied Mechanics 2021-03-16 | This video shows our humanoid robot LOLA performing various multi-contact maneuvers. The robot is technically blind, not using any camera-based input. For these experiments the navigation module is disabled, thus the foothold positions and hand contact points are manually set according to the current environment setup. However, disturbances (pushing, uneven terrain, rolling board) are not known to the robot and are compensated by our online stabilization methods. Note that ALL algorithms run in real-time and onboard. The robot's joints are position-controlled.Technical Details:Walking speed: up to 0.5 m/s = 1.8 km/h (depends on maneuver)Runtime of walking pattern generation (once): (30ms for 18.72s of motion in most complex scenario – realtime-factor ~ 600!)Runtime of stabilization and inverse kinematics (cyclic @1kHz): 350us (mean)For more information on LOLA please see our project's website: https://www.mw.tum.de/en/am/research/current-projects/robotics/humanoid-robot-lola/This work is supported by the German Research Foundation (DFG, project number 407378162).
Humanoid Robot LOLA - Robustness to Soft and Hard Obstacles TUM Chair of Applied Mechanics 2021-03-07 | In this video, LOLA reacts to randomly arranged, undetected soft and hard obstacles. The robot is technically blind, not using any camera-based or prior information on the terrain (hard ground is assumed). Due to the limited transmission of contact forces, unexpectedly soft obstacles are typically more difficult to overcome. When LOLA steps on soft obstacles, the control adapts itself to achieve similar performance to walking on hard ground.Technical Details:All runs use the same parametrization for the walking controller and identical footstep locations.Walking speed: 0.5 m/sHeight of the foam sheets: 5cmDensity of the foam: 40 kg/m^3Compression hardness of the foam: 6 kPa @ 40% compressionFor more information on LOLA please see our project's website: https://www.mw.tum.de/en/am/research/current-projects/robotics/humanoid-robot-lola/This work is supported by the German Research Foundation (DFG, project number 407378162).
Humanoid Robot LOLA - Robustness to Unexpected Ground Height Changes TUM Chair of Applied Mechanics 2021-02-27 | In this video, LOLA reacts to undetected ground height changes, including a drop and leg-in-hole experiment. Further tests show the robustness to vertical disturbances using a seesaw. The robot is technically blind, not using any camera-based or prior information on the terrain.Technical Details:Walking speed: 0.5 m/sHeight of the board: 6.7cmGround height change leg-in-hole: 6cmNegative ground height change platform: 9cmFor more information on LOLA please see our project's website: https://www.mw.tum.de/en/am/research/current-projects/robotics/humanoid-robot-lola/This work is supported by the German Research Foundation (DFG, project number 407378162).
Humanoid Robot LOLA Walks on Loose Wooden Boards TUM Chair of Applied Mechanics 2021-02-19 | This video shows the latest achievements for LOLA walking on undetected uneven terrain. The robot is technically blind, not using any camera-based or prior information on the terrain.Technical Details:Walking Speed: 0.5 m/sMax. height of the inclined wooden obstacles: 6 cmHeight of the platform: 12 cmMax. height of the pile of wooden plates: 9 cmFor more information on LOLA please see our project's website: https://www.mw.tum.de/en/am/research/current-projects/robotics/humanoid-robot-lola/This work is supported by the German Research Foundation (DFG, project number 407378162).
The Application of Neural Networks for Fidelity Assessment of Real-Time Hybrid Substructuring (RTHS) TUM Chair of Applied Mechanics 2021-02-08 | In this video, the idea of using Artificial Neural Networks (ANNs) to assess the fidelity of Real-Time Hybrid Simulation/Substructuring (RTHS) is proposed. It is investigated, whether ANNs could be a meaningful tool to predict the test fidelity---when no reference solution is avaialble---solely on data that can be measured during the test. An ANN was trained on data from 280 simulated RTHS tests and applied to data from real RTHS tests. The results revealed that the training process is successful and a relation between the selected input features (error indicators from literature, dynamical properties of the investigated system) and the target (test fidelity) found. The application of the trained ANN to a different dynamical system showed that more data have to be included in the training process such that the fidelity of never-seen dynamical systems can be predicted robustly. Therefore, a next step could be to gather data from historical RTHS tests and use all available tests for the training process. The work was presented at the conference IMAC-XXXIX.
Humanoid Robot LOLA v1.1 - Validation of Hardware Upgrade - Initial Tests TUM Chair of Applied Mechanics 2020-12-18 | In this video we present initial tests of our humanoid robot LOLA, which has recently received a major hardware upgrade. In particular the upper body of the robot was redesigned from scratch. The video is meant to demonstrate its basic operation. Note that this is raw footage of initial tests, i.e., WITHOUT fine-tuning of control parameters.Balancing:The robot is standing in an idle pose and receives external forces. The controller is able to compensate the disturbances without exciting oscillations.Stamping:The robot is commanded to walk in place, i.e., straight walking with a step length of zero. This kind of motion is typically more difficult for our robot when compared to "normal" walking.Walking:Basic straight walking of a fixed step sequence.For more information please see our project's website: https://www.mw.tum.de/en/am/research/current-projects/robotics/humanoid-robot-lola/This work is supported by the German Research Foundation (DFG, project number 407378162).
Winter Holiday Gyro TUM Chair of Applied Mechanics 2020-12-16 | Here's our little christmas gift to the fans of strange dynamic behavior.The gyro will follow any given shape as soon as the tip touches its edge and the rotation is fast enough.The friction between tip and shape generates a tangential force, creating a moment such that the gyroscopic reaction pushes the tip towards the shape.The resulting normal force produces a moment that guides the tip along the shape's edge.
Humanoid Robot LOLA v1.1 - Hardware Upgrade for Multi-Contact Locomotion TUM Chair of Applied Mechanics 2020-10-26 | In this video we present recent efforts to make our humanoid robot LOLA ready for multi-contact locomotion, i.e. additional hand-environment support for extra stabilization during walking. We focus on the hardware changes of the upper body (the pelvis and the legs remain unchanged).Comparison of degrees of freedom (DoFs):The new upper body features two additional joints, alias "arm rotation", which significantly increase the reachable taskspace. This is essential for supporting against walls in the lateral proximity of the robot.Assembly:We explain the composition of the new torso and arm design. The torso contains core components, like the inertial measurement unit (IMU), the two onboard PCs (control and vision), bus couplers and power bridges. Note that the cabling is not modeled and thus not visible in the renderings.Finite Element Analysis:A finite element analysis of two exemplary structural parts is shown. These parts have shown to be critical and were developed in an iterative process. The whole robot is designed for minimal weight.Taskspace Analysis:The reachable taskspace of the new arm design is compared to the previous topology. We define the torso segment as fixed "base" and the center of the hand as tool center point (TCP). The volume describes the reachable region while the surface is colored according to certain metrics. These metrics indicate "how well" a certain point in 3D space can be reached by the TCP. For this analysis the taskspace is defined as the cartesian position of the TCP relative to the base.For more information please see our project's website: https://www.mw.tum.de/en/am/research/current-projects/robotics/humanoid-robot-lola/This work is supported by the German Research Foundation (DFG, project number 407378162).
Hierarchical Motion Planning Framework for Manipulators in Human-Centered Environments TUM Chair of Applied Mechanics 2020-09-21 | This video shows the application of our real-time capable motion planning framework in a human-centered environment. A FRANKA EMIKA Panda robot is used in a Pick-and-Place scenario with human disturbances and dynamic obstacles. The framework has implemented a new Obstacle-related Sampling Rejection Probabilistic Roadmap planner that represents the free workspace in an efficient way. During online motion queries, dynamic obstacles can be avoided in real-time using an attractor-based online trajectory generation. The resulting motions satisfy kinematic and dynamic joint limits.The work was presented at the International Symposium on Robot and Human Interactive Communication (RO-MAN) 2020, http://ro-man2020.unina.it/.
Conference presentation Normalized Passivity Control for HiL with Contact TUM Chair of Applied Mechanics 2020-08-24 | This video explains the use of passivity control to stabilize Hardware-in-the-Loop tests (also known as Real-Time Hybrid Simulation (RTHS)). Our special focus is on mechanical systems with contact. The work was presented at the IFAC World Congress 2020, ifac2020.org
Motion Planning Framework for Human-Centered Environments TUM Chair of Applied Mechanics 2020-04-27 | This video shows the application of our real-time capable motion planning framework in a human-centered environment. A FRANKA EMIKA Panda robot is used in a Pick-and-Place scenario with human disturbances and dynamic obstacles. The framework has implemented a new Obstacle-related Sampling Rejection Probabilistic Roadmap planner that represents the free workspace in an efficient way. During online motion queries, dynamic obstacles can be avoided in real-time using an attractor-based online trajectory generation. The resulting motions satisfy kinematic and dynamic joint limits.For more information please see our project's website: https://www.mw.tum.de/en/am/research/current-projects/robotics/real-time-motion-planning-for-manipulators/
Smooth Real-Time Walking-Pattern Generation for Humanoid Robot LOLA TUM Chair of Applied Mechanics 2019-07-03 | This video demonstrates our new approach for planning smooth center-of-mass trajectories for biped walking robots. The method is based on quintic spline interpolation and collocation and generates dynamically and kinematically feasible motions in real-time.Method/Testscenario:Our humanoid robot Lola steps up and down a platform of 12.5cm height. The planned center-of-mass motion respects the dynamics and kinematic limits of the robot using simplified models. The complete motion lasts more than 17 seconds and is planned in less than 9 milliseconds (CPU only, single-core).Simulation:Custom multi-body simulationVisualization with Blender (custom interface using Blenders Python API)Experiments: Planning and control of the robot runs in real-time (onboard). External communication only triggers a signal to start and stop walking. The vision system is not active, thus foothold sequence is predefined.For more information please see our project's website: https://www.mw.tum.de/en/am/research/current-projects/robotics/humanoid-robot-lola/This work is supported by the German Research Foundation (DFG, project number 407378162).
The Humanoid Robot Lola walks Outside TUM Chair of Applied Mechanics 2019-05-24 | This video shows our humanoid robot Lola walking on asphalt, grass, and cobblestone. All motions were generated online and on board of the robot. The cable is needed for power supply as the robot doesn't have a battery yet. Lola is safeguarded by a safety harness - this rope is loose during walking.For this video, Lola was either operated by a human via a wireless controller or commanded simple objectives, e.g. "walk straight". The motion planner assumes a flat ground and the different ground properties are previously unknown to the robot's feedback control methods.For more information please see our project's website: http://www.amm.mw.tum.de/en/research/current-projects/humanoid-robot-lola/Or follow Lola on Instagram: @lola_the_robot
Overview Humanoid Robot LOLA 2018 TUM Chair of Applied Mechanics 2019-04-15 | This video gives an overview on the humanoid robot LOLA and the motion planning procedure required for biped walking. It shows the kinematic and the collision model of the robot and gives some insight to the planned trajectories of the ZMP and the feet. In addition, several experiments with LOLA from the year 2017 are shown.For more information please see our project's website: https://www.amm.mw.tum.de/en/research/current-projects/humanoid-robot-lola/
A Force Control Scheme for Biped Robots to Walk over Uneven Terrain Including Partial Footholds TUM Chair of Applied Mechanics 2019-03-21 | This video shows several experiments of walking over unknown, uneven terrain with the biped robot LOLA. Three different methods for ground-force control are experimentally validated in early, late and partial contact situations. The methods are combined to a force-control scheme to allow for the traversal of highly uneven terrain without vision-based information.Additional experiments for highly uneven terrain are available here: youtu.be/pmtKv8VEItYA preprint of our paper describing the force-control scheme is available online:https://mediatum.ub.tum.de/doc/1482152/1482152.pdfFor more information please see our project's website: https://www.mw.tum.de/en/am/research/current-projects/robotics/humanoid-robot-lola/
Automatic swing - a parametric oscillator TUM Chair of Applied Mechanics 2018-10-16 | An oscillation that automatically grows in amplitude through a servomotor moving a weigth up and down - similar to a child on a swing.
Test Rig Levitating in Active Magnetic Bearings TUM Chair of Applied Mechanics 2018-09-06 | Test Rig Levitating in Active Magnetic Bearings
Robot LOLA Autonomous Walking Compilation + HoloLens Mixed Reality App TUM Chair of Applied Mechanics 2017-10-29 | In this video we show a compilation of our research for the last 4 years on autonomous navigation of bipedal robots.It is part of the DFG-founded project "Versatile and Robust Walking in Uneven Terrain" (German Research Foundation) and includes developments in environment perception and modeling, motion planning and stability control. Additionally, we give a sneak peak of our HoloLens Mixed Reality App, "HoLola" which will be published soon!Check out our open source repositories!github.com/am-lolaFor more information, see our project's website:http://www.amm.mw.tum.de/en/research/current-projects/humanoid-robots/
Vision System of Humanoid Robot LOLA: Platforms and Dynamic Obstacles TUM Chair of Applied Mechanics 2017-10-23 | In this video, we test the capabilities of our Vision System and Motion Planning System, by letting the robot walk in previously unknown dynamic environments. The vision system approximates "walkable" surfaces as polygons and obstacles as capsules or spheres and the motion planning finds safe paths in real-time.Check out our open source repositories!github.com/am-lola
Our Humanoid Robot LOLA pre-simulates motions to better walk over complicated terrain TUM Chair of Applied Mechanics 2017-08-23 | Kinematic Optimization for Bipedal RobotsIn this video we show the result of our latest optimization algorithm: by pre-simulating the robot motion, taking its kinematics into account, we can generate better movements, exploiting the redundant kinematics of the robot.Check out our open source repositories!github.com/am-lola
Time-Variable Control for our Humanoid Robot LOLA (simulation) TUM Chair of Applied Mechanics 2017-07-24 | In this video we shortly present our new time-variable, event-based walking control for biped robots. We add two phases to our walking control which are activated when the robot finds itself in an early-contact or late-contact scenario. The newest phase, "glide", specifically deals with a late-contact scenario by preventing the tilting over of the robot. With it, our humanoid lola can deal with leg-in-hole situations. Real experiments coming soon!Check out our open source repositories!github.com/am-lola
Hybrid Position/Force Control for Biped Robot Stabilization with Integrated Center of Mass Dynamics TUM Chair of Applied Mechanics 2017-07-23 | In this video we show the performance of a new force control approach for our humanoid LOLA. By using an explicit contact model and integrated CoM dynamics, we are able to walk from an unexpected platform of 5.5 cm height.Check out our open source repositories!github.com/am-lolaCheck out our paper!http://mediatum.ub.tum.de/node?id=1394924
Our Humanoid Robot LOLA walks up and down platforms and stairs TUM Chair of Applied Mechanics 2017-04-12 | Usual control of humanoid robots uses heuristically pre-defined torso trajectories. When walking up and down, however, these trajectories may result in joints violating kinematic constraints. If we integrate the robot's motion beforehand the torso trajectories can be adapted such that kinematic limits are never reached, achieving a more precise, versatile and stable motion.Check out our open source repositories!github.com/am-lola
ALE-Beam Model: Reference/Shifting Motion for a CVT TUM Chair of Applied Mechanics 2017-03-15 | The ALE beam model consists of reference degrees and overlaid deformations. In this video the overlaid degrees of freedom are deactivated which only leaves the possibility to move in the reference configuration. This example shows the reference motion for a continuously variable transmission, specifically the ring-package of a pushbelt. One can clearly see that beam oscillates like a pendulum around a minimal energy position.
ALE Beam Model: Full Deformation TUM Chair of Applied Mechanics 2017-03-15 | The ALE beam model consists of reference degrees and overlaid deformations. In this video all degrees of freedom are activated. It becomes clear that the deformation is not natural yet qualitatively okay and correct for small deformations (right at the beginning).
