2018 IEEE International Conference on Simulation, Modeling, and Programming for Autonomous Robots (SIMPAR)最新文献

{"title":"Reinforcement learning for non-prehensile manipulation: Transfer from simulation to physical system","authors":"Kendall Lowrey, S. Kolev, Jeremy Dao, A. Rajeswaran, E. Todorov","doi":"10.1109/SIMPAR.2018.8376268","DOIUrl":"https://doi.org/10.1109/SIMPAR.2018.8376268","url":null,"abstract":"Reinforcement learning has emerged as a promising methodology for training robot controllers. However, most results have been limited to simulation due to the need for a large number of samples and the lack of automated-yet-safe data collection methods. Model-based reinforcement learning methods provide an avenue to circumvent these challenges, but the traditional concern has been the mismatch between the simulator and the real world. Here, we show that control policies learned in simulation can successfully transfer to a physical system, composed of three Phantom robots pushing an object to various desired target positions. We use a modified form of the natural policy gradient algorithm for learning, applied to a carefully identified simulation model. The resulting policies, trained entirely in simulation, work well on the physical system without additional training. In addition, we show that training with an ensemble of models makes the learned policies more robust to modeling errors, thus compensating for difficulties in system identification. The results are illustrated in the accompanying video.","PeriodicalId":156498,"journal":{"name":"2018 IEEE International Conference on Simulation, Modeling, and Programming for Autonomous Robots (SIMPAR)","volume":"310 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131895318","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The control toolbox — An open-source C++ library for robotics, optimal and model predictive control","authors":"Markus Giftthaler, Michael Neunert, M. Stäuble, J. Buchli","doi":"10.1109/SIMPAR.2018.8376281","DOIUrl":"https://doi.org/10.1109/SIMPAR.2018.8376281","url":null,"abstract":"We introduce the Control Toolbox (CT), an open-source C++ library for efficient modeling, control, estimation, trajectory optimization and Model Predictive Control. The CT is applicable to a broad class of dynamic systems but features interfaces to modeling tools specifically designed for robotic applications. This paper outlines the general concept of the toolbox, its main building blocks, and highlights selected application examples. The library contains several tools to design and evaluate controllers, model dynamical systems and solve optimal control problems. The CT was designed for intuitive modeling of systems governed by ordinary differential or difference equations. It supports rapid prototyping of cost functions and constraints and provides standard interfaces for different optimal control solvers. To date, we support Single Shooting, the iterative Linear-Quadratic Regulator, Gauss-Newton Multiple Shooting and classical Direct Multiple Shooting. We provide interfaces to general purpose NLP solvers and Riccati-based linear-quadratic optimal control solvers. The CT was designed to solve large-scale optimal control and estimation problems efficiently and allows for online control of dynamic systems. Some of the key features to enable fast run-time performance are full compatibility with Automatic Differentiation, derivative code generation, and multi-threading. Still, the CT is designed as a modular framework whose building blocks can also be used for other control and estimation applications such as inverse dynamics control, extended Kalman filters or kinematic planning. The CT is available as open-source software under the Apache v2 license and can be retrieved from https://bitbucket.org/adrlab/ct.","PeriodicalId":156498,"journal":{"name":"2018 IEEE International Conference on Simulation, Modeling, and Programming for Autonomous Robots (SIMPAR)","volume":"80 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-01-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127828852","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Conditional task and motion planning through an effort-based approach","authors":"Nicola Castaman, E. Tosello, E. Pagello","doi":"10.1109/SIMPAR.2018.8376270","DOIUrl":"https://doi.org/10.1109/SIMPAR.2018.8376270","url":null,"abstract":"This paper proposes a preliminary work on a Conditional Task and Motion Planning algorithm able to find a plan that minimizes robot efforts while solving assigned tasks. Unlike most of the existing approaches that replan a path only when it becomes unfeasible (e.g., no collision-free paths exist), the proposed algorithm takes into consideration a replanning procedure whenever an effort-saving is possible. The effort is here considered as the execution time, but it is extensible to the robot energy consumption. The computed plan is both conditional and dynamically adaptable to the unexpected environmental changes. Based on the theoretical analysis of the algorithm, authors expect their proposal to be complete and scalable. In progress experiments aim to prove this investigation.","PeriodicalId":156498,"journal":{"name":"2018 IEEE International Conference on Simulation, Modeling, and Programming for Autonomous Robots (SIMPAR)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131305681","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"DoShiCo challenge: Domain shift in control prediction","authors":"Klaas Kelchtermans, T. Tuytelaars","doi":"10.1109/SIMPAR.2018.8376264","DOIUrl":"https://doi.org/10.1109/SIMPAR.2018.8376264","url":null,"abstract":"Training deep neural network policies end-to-end for real-world applications so far requires big demonstration datasets in the real world or big sets consisting of a large variety of realistic and closely related 3D CAD models. These real or virtual data should, moreover, have very similar characteristics to the conditions expected at test time. These stringent requirements and the time consuming data collection processes that they entail, are currently the most important impediment that keeps deep reinforcement learning from being deployed in real-world applications. Therefore, in this work we advocate an alternative approach, where instead of avoiding any domain shift by carefully selecting the training data, the goal is to learn a policy that can cope with it. To this end, we propose the DoShiCo challenge: to train a model in very basic synthetic environments, far from realistic, in a way that it can be applied in more realistic environments as well as take the control decisions on real-world data. In particular, we focus on the task of collision avoidance for drones. We created a set of simulated environments that can be used as benchmark and implemented a baseline method, exploiting depth prediction as an auxiliary task to help overcome the domain shift. Even though the policy is trained in very basic environments, it can learn to fly without collisions in a very different realistic simulated environment. Of course several benchmarks for reinforcement learning already exist — but they never include a large domain shift. On the other hand, several benchmarks in computer vision focus on the domain shift, but they take the form of a static datasets instead of simulated environments. In this work we claim that it is crucial to take the two challenges together in one benchmark.","PeriodicalId":156498,"journal":{"name":"2018 IEEE International Conference on Simulation, Modeling, and Programming for Autonomous Robots (SIMPAR)","volume":"67 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114133529","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"ROS-health: An open-source framework for neurorobotics","authors":"Gloria Beraldo, Nicola Castaman, R. Bortoletto, E. Pagello, J. Millán, L. Tonin, E. Menegatti","doi":"10.1109/SIMPAR.2018.8376288","DOIUrl":"https://doi.org/10.1109/SIMPAR.2018.8376288","url":null,"abstract":"This paper aims at providing a preliminary description of ROS-Health, a novel framework for neurorobotics based on the middleware Robot Operating System (ROS). The increased interest in the neurorobotics field and the proliferation of several (neuro)physiological-based applications to control robotics devices made clear the importance to establish a standardized research platform in order to facilitate the distribution of the software, the replication of experimental results and the creation of an unified community to share and manage the code in the years. For this reason, we propose a common platform developed in the ROS ecosystem that takes advantage of its tools and capabilities. Furthermore, we describe the design guidelines that we are following in the preliminary definition of ROS-Health architecture. Finally, we present two illustrative use cases that highlight the advantages and benefits of the adoption of ROS-Health.","PeriodicalId":156498,"journal":{"name":"2018 IEEE International Conference on Simulation, Modeling, and Programming for Autonomous Robots (SIMPAR)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125527690","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Simulating differential games with improved fidelity to better inform cooperative & adversarial two vehicle UAV flight","authors":"Anne Redulla, Surya P. N. Singh","doi":"10.1109/SIMPAR.2018.8376282","DOIUrl":"https://doi.org/10.1109/SIMPAR.2018.8376282","url":null,"abstract":"Automatic determination of an areal vehicle's strategy rests on accurate simulation and forecasting to inform control decisions. As system dynamics affect the outcomes, when two or more vehicles interact finding the best strategy to take may be considered a differential game. While traditionally modelled using ideal kinematics, the effect on, and the variation of, the optimal strategy based on simulating realistic dynamics is investigated. A derivation for the optimal strategies of the players in ‘regular’ regions of the state space is completed. Two simulators were developed to compare game terminal results to the theoretical predictions. A MATLAB simulator with ideal player kinematics was the first simulator formed. Then, a high-fidelity simulator, using Microsoft's AirSim project was implemented. This involved configuring AirSim to run using software commands only, and extending the functionality to allow for the simulation of two separately-controlled drones. A trivial differential game with two agile players, termed Pedestrian Tag (PT), was used to identify the accuracy of time-to-capture predictions. The MATLAB simulator was found to match the model prediction, whereas the AirSim simulator required more gameplay time than predicted to achieve capture. For a Homicidal Chauffeur (HC) game, the MATLAB simulator results were consistent with the theoretical predictions. However, multiple outcomes of trials contrasted the predicted terminal results for the high-fidelity simulator. The results indicate that modelling the players to have ideal kinematics does not correctly predict the outcome of a pursuit-evasion game with full/realistic dynamics. Although some deviation from the model assumptions was introduced due to implementation constraints, the primary factor was concluded to be the realistic velocities of the drone agents due to unaccounted dynamics such as inertia and drag. Future research topics prompted by this work include applying the simulation to more differential games, and comparing against player strategies developed from other methods.","PeriodicalId":156498,"journal":{"name":"2018 IEEE International Conference on Simulation, Modeling, and Programming for Autonomous Robots (SIMPAR)","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128992048","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}