Mirko Ferrati

WALK‐MAN: A High‐Performance Humanoid Platform for Realistic Environments

In this work, we present WALK-MAN, a humanoid platform that has been developed to operate in realistic unstructured environment, and demonstrate new skills including powerful manipulation, robust balanced locomotion, high-strength capabilities, and physical sturdiness. To enable these capabilities, WALK-MAN design and actuation are based on the most recent advancements of series elastic actuator drives with unique performance features that differentiate the robot from previous state-of-the-art compliant actuated robots. Physical interaction performance is benefited by both active and passive adaptation, thanks to WALK-MAN actuation that combines customized high-performance modules with tuned torque/velocity curves and transmission elasticity for high-speed adaptation response and motion reactions to disturbances. WALK-MAN design also includes innovative design optimization features that consider the selection of kinematic structure and the placement of the actuators with the body structure to maximize the robot performance. Physical robustness is ensured with the integration of elastic transmission, proprioceptive sensing, and control. The WALK-MAN hardware was designed and built in 11 months, and the prototype of the robot was ready four months before DARPA Robotics Challenge (DRC) Finals. The motion generation of WALK-MAN is based on the unified motion-generation framework of whole-body locomotion and manipulation (termed loco-manipulation). WALK-MAN is able to execute simple loco-manipulation behaviors synthesized by combining different primitives defining the behavior of the center of gravity, the motion of the hands, legs, and head, the body attitude and posture, and the constrained body parts such as joint limits and contacts. The motion-generation framework including the specific motion modules and software architecture is discussed in detail. A rich perception system allows the robot to perceive and generate 3D representations of the environment as well as detect contacts and sense physical interaction force and moments. The operator station that pilots use to control the robot provides a rich pilot interface with different control modes and a number of teleoperated or semiautonomous command features. The capability of the robot and the performance of the individual motion control and perception modules were validated during the DRC in which the robot was able to demonstrate exceptional physical resilience and execute some of the tasks during the competition.

A manipulation framework for compliant humanoid COMAN: Application to a valve turning task

With the purpose of achieving a desired interaction performance for our compliant humanoid robot (COMAN), in this paper we propose a semi-autonomous control framework and evaluate it experimentally in a valve turning setup. The control structure consists of various modules and interfaces to identify the valve, locate the robot in front of it and perform the manipulation. The manipulation module implements four motion primitives (Reach, Grasp, Rotate and Disengage) and realizes the corresponding desired impedance profile for each phase to accomplish the task. In this direction, to establish a stable and compliant contact between the valve and the robot hands, while being able to generate the sufficient rotational torques depending on the valves friction, Rotate incorporates a novel dual-arm impedance control technique to plan and realize a task-appropriate impedance profile. Results of the implementation of the proposed control framework are firstly evaluated in simulation studies using Gazebo. Subsequent experimental results highlight the efficiency of the proposed impedance planning and control in generation of the required interaction forces to accomplish the task.

Yarp Based Plugins for Gazebo Simulator

This paper presents a set of plugins for the Gazebo simulator that enables the interoperability between a robot, controlled using the YARP framework, and Gazebo itself. Gazebo is an open-source simulator that can handle different Dynamic Engines developed by the Open Source Robotics Foundation. Since our plugins conform with the YARP layer used on the real robot, applications written for our robots, COMAN and iCub, can be run on the simulator with no changes. Our plugins have two main components: a YARP interface with the same API as the real robot interface, and a Gazebo plugin which handles simulated joints, encoders, IMUs, force/torque sensors and synchronization. Different modules and tasks for COMAN and iCub have been developed using Gazebo and our plugins as a testbed before moving to the real robots.

Upper-body impedance control with variable stiffness for a door opening task

The advent of humanoids has brought new challenges in the real-world application. As a part of ongoing efforts to foster functionality of the robot accommodating a real environment, this paper introduces a recent progress on a door opening task with our compliant humanoid, CoMan. We presents a task-prioritized impedance control framework for an upper body system that includes a dual-arm, a waist, two soft hands, and 3D camera. Aimed to create desired responses to open the door, a novel stiffness modulation method is proposed, incorporating a realtime optimization. As a preliminary experiment, a full door-opening scenario (approaching to the door and reaching, grasping, rotating and pulling the door handle) is demonstrated under a semi-autonomous operation with a pilot. The experimental result shows the effectiveness and efficacy of the proposed impedance control approach. Despite of uncertainties from sensory data, the door opening task is successfully achieved and safe and robust interaction is established without creating excessive forces.

On the Problem of Moving Objects With Autonomous Robots: A Unifying High-Level Planning Approach

Moving objects with autonomous robots is a wide topic that includes single-arm pick-and-place tasks, object regrasping, object passing between two or more arms in the air or using support surfaces such as tables and similar. Each task has been extensively studied and many planning solutions are already present in the literature. In this letter, we present a planning scheme which, based on the use of pre-defined elementary manipulation skills, aims to unify solutions which are usually obtained by means of different planning strategies rooted on hard-coded behaviors. Both robotic manipulators and environment fixed support surfaces are treated as end-effectors of movable and non-movable types, respectively. The task of the robot can thus be broken down into elementary building blocks, which are end-effector manipulation skills, that are then planned at the kinematic level. Feasibility is ensured by propagating unforeseen low-level failures at the higher level and by synthesizing different behaviors. The validity of the proposed solution is shown via experiments on a bimanual robot setup and in simulations involving a more complex setup similar to an assembly line.

A Modular Approach for Remote Operation of Humanoid Robots in Search and Rescue Scenarios

In this work we present a modular, robust and user-friendly Pilot Interface meant to control humanoid robots in rescue scenarios during dangerous missions.

A time expanded network based algorithm for safe and efficient distributed multi-agent coordination

In this paper we propose a novel approach for distributed traffic management of a group of mobile collaborative vehicles, moving within a shared environment. Our algorithm is based on a modified graph representation of the space-time where the robotic vehicles operate. The proposed graph representation augments standard path planning strategies by allowing multiple speeds. Robots can negotiate their route and set their speed to prevent and solve possible collisions while taking into account the energy consumption. Indeed, the algorithm will be formally proved to provide collision free paths.

The Walk-Man Robot Software Architecture

A software and control architecture for a humanoid robot is a complex and large project, that involves a team of developers/researchers to be coordinated and requires many hard design choices. If such project has to be done in a very limited time, i.e. less than one year, more constraints are added and concepts such as modular design, code reusability and API definition need to be used as much as possible. In this work we describe the software architecture developed for Walk-Man, a robot participant at the Darpa Robotics Challenge. The challenge required the robot to execute many different tasks such as walking, driving a car, and manipulating objects.These tasks need to be solved by robotics specialists in their corresponding research field, such as humanoid walking, motion planning or object manipulation. The proposed architecture was developed in 10 months, provided boilerplate code for most of the functionalities required to control a humanoid robot and allowed robotics researchers to produce their control modules for DRC tasks in a short time. Additional capabilities of the architecture include firmware and hardware management, mixing of different middlewares, unreliable network management,operator control station GUI. All the source code related to the architecture and some control modules have been released as open source projects.

Multi-object handling for robotic manufacturing

The purpose of this work is to move a step toward the automation of industrial plants through full exploitation of autonomous robots. A planning algorithm is proposed to move different objects in desired configurations with heterogeneous robots such as manipulators, mobile robots and conveyor belts. The proposed approach allows different objects to be handled by different robots simultaneously in an efficient way and avoiding collisions with the environment and self-collisions between robots. In particular, the integrated system will be capable of planning paths for a set of objects from various starting points in the environment (e.g. shelves) to their respective final destinations. The proposed approach unifies the active (e.g., grasping by a hand) and passive (e.g., holding by a table) steps involved in moving the objects in the environment by treating them as end-effectors with constraints and capabilities. Time varying graphs will be introduced to model the problem for simultaneous handling of objects by different end-effectors. Optimal exploration of such graphs will be used to determine paths for each object with time constraints. Results will be validated through simulations.

ASCARI: A Component Based Simulator for Distributed Mobile Robot Systems

ASCARI is a simulator dedicated to distributed and cooperative mobile robotics systems, designed as a framework for implementing and testing multi-agent collaborative algorithms, especially suited to evaluate algorithms performances with a non-perfect communication channel (e.g. delayed, limited bandwidth, limited range). Compared to state-of-art simulators, ASCARI meets a complex new requirement: inter-agent communication has to be integrated in the simulation loop.