- Showing results for:
- Reset all filters
Journal articleAlAttar A, Kormushev P, 2020,
Kinematic-model-free orientation control for robot manipulation using locally weighted dual quaternions, Robotics, Vol: 9, Pages: 1-12, ISSN: 2218-6581
Conventional control of robotic manipulators requires prior knowledge of their kinematic structure. Model-learning controllers have the advantage of being able to control robots without requiring a complete kinematic model and work well in less structured environments. Our recently proposed Encoderless controller has shown promising ability to control a manipulator without requiring any prior kinematic model whatsoever. However, this controller is only limited to position control, leaving orientation control unsolved. The research presented in this paper extends the state-of-the-art kinematic-model-free controller to handle orientation control to manipulate a robotic arm without requiring any prior model of the robot or any joint angle information during control. This paper presents a novel method to simultaneously control the position and orientation of a robot’s end effector using locally weighted dual quaternions. The proposed novel controller is also scaled up to control three-degrees-of-freedom robots.
Journal articleCursi F, Mylonas GP, Kormushev P, 2020,
Adaptive kinematic modelling for multiobjective control of a redundant surgical robotic tool, Robotics, Vol: 9, Pages: 68-68, ISSN: 2218-6581
Accurate kinematic models are essential for effective control of surgical robots. For tendon driven robots, which are common for minimally invasive surgery, the high nonlinearities in the transmission make modelling complex. Machine learning techniques are a preferred approach to tackle this problem. However, surgical environments are rarely structured, due to organs being very soft and deformable, and unpredictable, for instance, because of fluids in the system, wear and break of the tendons that lead to changes of the system’s behaviour. Therefore, the model needs to quickly adapt. In this work, we propose a method to learn the kinematic model of a redundant surgical robot and control it to perform surgical tasks both autonomously and in teleoperation. The approach employs Feedforward Artificial Neural Networks (ANN) for building the kinematic model of the robot offline, and an online adaptive strategy in order to allow the system to conform to the changing environment. To prove the capabilities of the method, a comparison with a simple feedback controller for autonomous tracking is carried out. Simulation results show that the proposed method is capable of achieving very small tracking errors, even when unpredicted changes in the system occur, such as broken joints. The method proved effective also in guaranteeing accurate tracking in teleoperation.
Journal articleFalck F, Doshi S, Tormento M, et al., 2020,
Robot DE NIRO: a human-centered, autonomous, mobile research platform for cognitively-enhanced manipulation, Frontiers in Robotics and AI, Vol: A17, ISSN: 2296-9144
We introduceRobot DE NIRO, an autonomous, collaborative, humanoid robot for mobilemanipulation. We built DE NIRO to perform a wide variety of manipulation behaviors, with afocus on pick-and-place tasks. DE NIRO is designed to be used in a domestic environment,especially in support of caregivers working with the elderly. Given this design focus, DE NIRO caninteract naturally, reliably, and safely with humans, autonomously navigate through environmentson command, intelligently retrieve or move target objects, and avoid collisions efficiently. Wedescribe DE NIRO’s hardware and software, including an extensive vision sensor suite of 2Dand 3D LIDARs, a depth camera, and a 360-degree camera rig; two types of custom grippers;and a custom-built exoskeleton called DE VITO. We demonstrate DE NIRO’s manipulationcapabilities in three illustrative challenges: First, we have DE NIRO perform a fetch-an-objectchallenge. Next, we add more cognition to DE NIRO’s object recognition and grasping abilities,confronting it with small objects of unknown shape. Finally, we extend DE NIRO’s capabilitiesinto dual-arm manipulation of larger objects. We put particular emphasis on the features thatenable DE NIRO to interact safely and naturally with humans. Our contribution is in sharinghow a humanoid robot with complex capabilities can be designed and built quickly with off-the-shelf hardware and open-source software. Supplementary material including our code, adocumentation, videos and the CAD models of several hardware parts are openly availableavailable athttps://www.imperial.ac.uk/robot-intelligence/software/
Journal articleChappell D, Wang K, Kormushev P, 2020,
Asynchronous Real-Time Optimization of Footstep Placement and Timing in Bipedal Walking Robots
Online footstep planning is essential for bipedal walking robots to be ableto walk in the presence of disturbances. Until recently this has been achievedby only optimizing the placement of the footstep, keeping the duration of thestep constant. In this paper we introduce a footstep planner capable ofoptimizing footstep placement and timing in real-time by asynchronouslycombining two optimizers, which we refer to as asynchronous real-timeoptimization (ARTO). The first optimizer which runs at approximately 25 Hz,utilizes a fourth-order Runge-Kutta (RK4) method to accurately approximate thedynamics of the linear inverted pendulum (LIP) model for bipedal walking, thenuses non-linear optimization to find optimal footsteps and duration at a lowerfrequency. The second optimizer that runs at approximately 250 Hz, usesanalytical gradients derived from the full dynamics of the LIP model andconstraint penalty terms to perform gradient descent, which finds approximatelyoptimal footstep placement and timing at a higher frequency. By combining thetwo optimizers asynchronously, ARTO has the benefits of fast reactions todisturbances from the gradient descent optimizer, accurate solutions that avoidlocal optima from the RK4 optimizer, and increases the probability that afeasible solution will be found from the two optimizers. Experimentally, weshow that ARTO is able to recover from considerably larger pushes and producesfeasible solutions to larger reference velocity changes than a standardfootstep location optimizer, and outperforms using just the RK4 optimizeralone.
Conference paperPardo F, Levdik V, Kormushev P, 2020,
Scaling all-goals updates in reinforcement learning using convolutional neural networks, 34th AAAI Conference on Artificial Intelligence (AAAI 2020), Publisher: Association for the Advancement of Artificial Intelligence, Pages: 5355-5362, ISSN: 2374-3468
Being able to reach any desired location in the environmentcan be a valuable asset for an agent. Learning a policy to nav-igate between all pairs of states individually is often not fea-sible. Anall-goals updatingalgorithm uses each transitionto learn Q-values towards all goals simultaneously and off-policy. However the expensive numerous updates in parallellimited the approach to small tabular cases so far. To tacklethis problem we propose to use convolutional network archi-tectures to generate Q-values and updates for a large numberof goals at once. We demonstrate the accuracy and generaliza-tion qualities of the proposed method on randomly generatedmazes and Sokoban puzzles. In the case of on-screen goalcoordinates the resulting mapping from frames todistance-mapsdirectly informs the agent about which places are reach-able and in how many steps. As an example of applicationwe show that replacing the random actions inε-greedy ex-ploration by several actions towards feasible goals generatesbetter exploratory trajectories on Montezuma’s Revenge andSuper Mario All-Stars games.
BookDeisenroth MP, Faisal AA, Ong CS, 2020,
Mathematics for Machine Learning, Publisher: Cambridge University Press, ISBN: 9781108455145
Conference paperSaputra RP, Rakicevic N, Kormushev P, 2020,
Sim-to-real learning for casualty detection from ground projected point cloud data, 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2019), Publisher: IEEE
This paper addresses the problem of human body detection-particularly a human body lying on the ground (a.k.a. casualty)-using point cloud data. This ability to detect a casualty is one of the most important features of mobile rescue robots, in order for them to be able to operate autonomously. We propose a deep-learning-based casualty detection method using a deep convolutional neural network (CNN). This network is trained to be able to detect a casualty using a point-cloud data input. In the method we propose, the point cloud input is pre-processed to generate a depth image-like ground-projected heightmap. This heightmap is generated based on the projected distance of each point onto the detected ground plane within the point cloud data. The generated heightmap-in image form-is then used as an input for the CNN to detect a human body lying on the ground. To train the neural network, we propose a novel sim-to-real approach, in which the network model is trained using synthetic data obtained in simulation and then tested on real sensor data. To make the model transferable to real data implementations, during the training we adopt specific data augmentation strategies with the synthetic training data. The experimental results show that data augmentation introduced during the training process is essential for improving the performance of the trained model on real data. More specifically, the results demonstrate that the data augmentations on raw point-cloud data have contributed to a considerable improvement of the trained model performance.
Journal articleRakicevic N, Kormushev P, 2019,
Active learning via informed search in movement parameter space for efficient robot task learning and transfer, Autonomous Robots, Vol: 43, Pages: 1917-1935, ISSN: 0929-5593
Learning complex physical tasks via trial-and-error is still challenging for high-degree-of-freedom robots. Greatest challenges are devising a suitable objective function that defines the task, and the high sample complexity of learning the task. We propose a novel active learning framework, consisting of decoupled task model and exploration components, which does not require an objective function. The task model is specific to a task and maps the parameter space, defining a trial, to the trial outcome space. The exploration component enables efficient search in the trial-parameter space to generate the subsequent most informative trials, by simultaneously exploiting all the information gained from previous trials and reducing the task model’s overall uncertainty. We analyse the performance of our framework in a simulation environment and further validate it on a challenging bimanual-robot puck-passing task. Results show that the robot successfully acquires the necessary skills after only 100 trials without any prior information about the task or target positions. Decoupling the framework’s components also enables efficient skill transfer to new environments which is validated experimentally.
Conference paperFalck F, Doshi S, Smuts N, et al., 2019,
Human-centered manipulation and navigation with robot DE NIRO
Social assistance robots in health and elderly care have the potential tosupport and ease human lives. Given the macrosocial trends of aging andlong-lived populations, robotics-based care research mainly focused on helpingthe elderly live independently. In this paper, we introduce Robot DE NIRO, aresearch platform that aims to support the supporter (the caregiver) and alsooffers direct human-robot interaction for the care recipient. Augmented byseveral sensors, DE NIRO is capable of complex manipulation tasks. It reliablyinteracts with humans and can autonomously and swiftly navigate throughdynamically changing environments. We describe preliminary experiments in ademonstrative scenario and discuss DE NIRO's design and capabilities. We putparticular emphases on safe, human-centered interaction procedures implementedin both hardware and software, including collision avoidance in manipulationand navigation as well as an intuitive perception stack through speech and facerecognition.
Conference paperTavakoli A, Levdik V, Islam R, et al., 2019,
Exploring Restart Distributions, Montréal, Canada, The Fourth Multidisciplinary Conference on Reinforcement Learning and Decision Making, Publisher: arXiv
We consider the generic approach of using an experience memory to help exploration by adapting a restart distribution. That is, given the capacity to reset the state with those corresponding to the agent's past observations, we help exploration by promoting faster state-space coverage via restarting the agent from a more diverse set of initial states, as well as allowing it to restart in states associated with significant past experiences. This approach is compatible with both on-policy and off-policy methods. However, a caveat is that altering the distribution of initial states could change the optimal policies when searching within a restricted class of policies. To reduce this unsought learning bias, we evaluate our approach in deep reinforcement learning which benefits from the high representational capacity of deep neural networks. We instantiate three variants of our approach, each inspired by an idea in the context of experience replay. Using these variants, we show that performance gains can be achieved, especially in hard exploration problems.
This data is extracted from the Web of Science and reproduced under a licence from Thomson Reuters. You may not copy or re-distribute this data in whole or in part without the written consent of the Science business of Thomson Reuters.