Skip to main content

Optimising Robot Performance while Dreaming

Periodic Reporting for period 2 - RobDREAM (Optimising Robot Performance while Dreaming)

Reporting period: 2016-02-01 to 2017-01-31

Our Project Vision in RobDREAM can be characterised by the following summary:
Sleep! For hominids and most other mammals sleep means more than regeneration. Sleep positively affects working memory, which in turn improves higher-level cognitive functions such as decision making and reasoning. This is the inspiration of RobDREAM!

What if robots could also improve their capabilities in their inactive phases – by processing experiences made during the working day and by exploring – or “dreaming” of – possible future situations and how to solve them best?

In RobDREAM we will improve industrial mobile manipulators’ perception, navigation and manipulation and grasping capabilities by automatic optimization of parameters, strategies and selection of tools within a portfolio of key algorithms for perception, navigation and manipulation and grasping, by means of learning and simulation, and through use case driven evaluation. As a result, mobile manipulation systems will adapt more quickly to new tasks, jobs, parts, areas of operation and various other constraints.

The two main project goals are:
- the acceleration of the setup process for new automation tasks through more structured and intuitive programming of the robot at task level. Allowing the system to exchange parts of the program with automatically generated and optimized routines.
- speeding up of planning and execution for general methods to solve navigation, manipulation and grasping tasks as well as general approaches for object recognition and object localisation through the use of context information and automated algorithm configuration techniques.

These two project goals and the project vision will be achieved by working on the scientific and technical objectives of the project:
(1) To perform a thorough use case and requirement analysis with a variety of end users to find out about the details on how future production assistants should assist shop floor workers with logistics, machine tending, and pre-assembly tasks.
(2) To shorten deployment and setup times of production assistants by enabling shop floor workers to intuitively tune operational parameters.
(3) To develop an architecture and test-bed that allows the robot to enhance its performance while “dreaming”.
(4) To enhance efficiency & performance of production assistants by automatically optimizing parameters depending on context information for the key technologies navigation, manipulation and grasping, and perception.
(5) To adapt a large portfolio of state-of-the-art algorithms in navigation, manipulation and grasping, and perception, in a uniform and open way, and make the portfolio grow by soliciting algorithms from third parties.
(6) To iteratively test and validate in realistic use cases during the whole duration of the project and transfer best practice solutions from academic to industrial partners and disseminate results beyond.
According to the project objectives described above the work briefly summarized below has been performed in the first reporting period:
(1) A use case and requirement analysis has been carried out in WP1 and this work will continue for the further design of the realistic manufacturing environment.
(2) A graphical programming environment has been empowered for the use within the RobDREAM project in WP6.
(3) A general structure for the DREAMing framework has been developed and development for optimising task performance for single and multiple contexts has been started successfully.
(4) Each key functionality for mobile manipulation (navigation, perception, mobile manipulation and grasping) has defined interfaces to the DREAMing architecture and defined individual DREAMBeds.
(5) Work towards the adaptation of state-of-the-art algorithms for the key functionalities for mobile manipulation has been started in WP3, WP4 and WP5.
(6) The relevant key features of a mobile manipulator have been tested already with the engineering framework for setting up mobile manipulation tasks.

According to the project objectives described above the work briefly summarized below has been performed in the second reporting period:
(1) Although the work on achieving this goal has been finalized during the first project period, the partners in RobDREAM have continued to monitor the requirements of mobile manipulation use cases.
(2) A method for simplified setup of mobile manipulation applicaitons has been developed and implemented in WP6. Also the programming framework has been used to set up the evaluation application and to integrate the input of the consortium partners.
(3) A method for simplified setup of mobile manipulation applicaitons has been developed and implemented in WP6.
(4) The DREAMBeds defined in the first reporting period of RobDREAM have been continuously adapted and used for optimizing parameters of the technologies in WP2-WP5 in local testbeds.
(5) The system integration efforts in WP6 led to a system which can integrate the algorithms developed by the RobDREAM consortium as well as third party algorithms.
(6) The evaluation efforts have been continued and greatly intensified with the availability of the target mobile manipulation system based on the KUKA KMR iiwa.
RobDREAM aims for progress beyond the state of the art in the following technological and scientific areas:
- Usability and applicability of mobile manipulators
o fully integrated mobile manipulator with perception, navigation, manipulation and planning facilities
o engineering framework for setting up mobile manipulation tasks
- Optimisation methods
o DREAM architecture and optimisation algorithms
- Navigation in changing environments
o Integration of the navigation portfolio in the DREAM architecture
o experience graph approach for mobile navigation tasks in industrial environments
o teach and replay for navigation with computed analytical representation for constrained optimisation problem
- Manipulation and grasping and motion planning
o Integration of the classical sampling based grasping in the DREAM architecture
o Regrasping approaches based on multisensory feedback (visual, tactile)
o Framework for grasp consolidation based on learning from similar examples
o Grasping approaches that consider task and context information
- Perception methods
o Depth-image based object detection integrated into DREAM architecture
o User interaction for perception with automatically derived object annotations

RobDREAM also makes major contributions to the following expected impacts defined in H2020-ICT-2014-1 (ICT 23-2014: Robotics, a. Research & Innovation Actions):
a) Increase Europe's market share in industrial robotics to one third of the market and maintain and strengthen Europe's market share of 50% in professional service robotics by 2020
d) Increase Industry-Academia cross-fertilisation and tighter connection between industrial needs and academic research via technology transfer, common projects, scientific progress on industry-driven challenges (major contribution by RobDREAM)
e) Deploy robotics technologies in new application domains
f) Improve Technology Readiness Levels of robotics technologies
g) Improve performance evaluation and certification of new robotic systems
h) Create and maintain world class research in Europe and achieve excellent standards of publications and research outputs
j) Ensure wide use of shared resources
Project Logo of the RobDREAM project