Community Research and Development Information Service - CORDIS

H2020

SecondHands Report Summary

Project ID: 643950
Funded under: H2020-EU.2.1.1.5.

Periodic Reporting for period 1 - SecondHands (SecondHands: A Robot Assistant For Industrial Maintenance Tasks)

Reporting period: 2015-05-01 to 2016-04-30

Summary of the context and overall objectives of the project

SecondHands targets the development of a robot capable of providing proactive help to a human user. To this end, it pursues advances on robotics and computer vision in the following areas: a) grasping and manipulation of complex objects, b) physical human-robot interaction for handover and joint manipulation of heavy loads; c) real-time monitoring of human motion for intent detection; d) planning and reasoning with a focus on failure recovery and proactive robotic action.
The work is phrased in the context of a real-world industrial scenario at Ocado; Ocado uses a high degree of automation to deliver groceries in a predictable time-critical way and the reliable operation of this automation is critical to the business' success. Assisting the technicians whose job it is to provide this reliability is seen as a major potential contribution to the business' efficiency.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

This first year of work has seen progresses toward each of the scientific areas targeted by the project. Specifically: a) Mechanoinformatics: Design of a novel torque-controlled dual arm system with underactuated five-fingered hands, and release of ArmarX as a complete software development environment and functional architecture for integration in the project at a very early stage; b) Grasping and manipulation: Extension of the ARMAR grasping pipeline to deal with hand-over tasks, task specific grasping and partially unknown objects; c) Physical human-Robot Interaction: algorithms were developed to enable continuous adaptation of position and force in response to human motion and interaction force during handover and shown to offer a more efficient and smoother human-robot interaction.; d) Object tracking and Human Motion modelling: algorithms were developed for six degree of freedom object pose estimation, articulated body (human) pose estimation and dynamic SLAM amongst multiple moving objects; e) Reasoning: Fast planning search methods have been adopted, and new algorithms for the recognition and classification of human action primitives are being developed; f) Verbal human-robot interaction: low latency speech recognition and dialog system modules were developed to enable the technician to verbally ask the robot for tools.

Additionally, significant efforts were made toward building the demonstrator. A complete set of specification and CAD drawings have been created for the new Armar platform. Code from all partners was integrated in the ArmarX framework, which forms the software infrastructure of the Armar robot. This enabled a first demonstration of live object detection and human monitoring combined with novel algorithms to enable compliant grasping and smooth handover of objects from human to robot and vice-versa.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

SecondHands is still in its early stages, yet, it has already produced significant algorithmic development in robotics and computer vision for smooth and compliant object manipulation, fast and robust object tracking and complex reasoning in the face of uncertainty. This work has led to nine publications, with a further four currently under review.

The initial press releases about the project have generated significant interest, partly because of the ambitious nature of the project and partly because of the nature of the ambitions. There seems to be a growing realisation that robotic coworkers and assistants will have wide applicability in many scenarios where people still have a role to play but can benefit from automated support that enhances their performance.

Related information

Record Number: 190314 / Last updated on: 2016-11-14
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top