Three projects have been granted support in the first call within the new funding instrument called Industry Bridge.
The objective of the Industry Bridge instrument is to facilitate fast and flexible support for initial steps of research collaboration between WASP industries and WASP academic research groups. This is done by providing funding for university research engineers for up to one year, as well as possible support for necessary experimental equipment. The aim is also to explore the possibilities and form the foundations for longer and deeper collaborations in the other WASP initiatives, such as research arenas (WARA) and industrial/academic PhD projects.
The three successful projects in the first round of Industry Bridge funding are:
AD-EYE – an Open Modular Testbed
AD-EYE is a testbed for Automated Driving (AD) and Intelligent Transportation Systems (ITS), initiated by WASP PhD student Naveen Mohan. It is anchored at the Division of Mechatronics and Embedded Control Systems at KTH Royal Institute of Technology. This sub-project funded by the WASP industry bridge call is in conjunction with the Center for Trustworthy Edge Computing Systems (TECoSA), and industrial partners Ericsson and ZenseAct and aims to transform the KTH campus into an innovation arena for AD and ITS.
AD and ITS are very large and complex tasks, requiring a multidisciplinary approach. The goal of AD-EYE is to aid researchers to overcome hurdles in making the transition from lower to mid/high TRL levels, especially challenging for research in safety critical fields such as AD and ITS. Usual hurdles are, for example: regulatory requirements, lack of open data, approvals of Transportstyrelsen, or that multidisciplinary collaboration is required.
The AD-EYE testbed emphasizes architecture openness through modular, standardized interfaces, open-source base functionality and data. The modular reference implementation can be selectively used or improved by the individual researcher in order for them to keep their focus on their research interests, while still using the rest of AD-EYE testbed implementation and resources. The resources available are of different types such as the ability to drive on public roads on KTH campus, access to sensors, compute hardware, workstations, open data (e.g. https://www.adeye.se/open-kth), proprietary data (e.g. from Stockholm city) and communication (5G, WiFi, V2X).
In conclusion, AD-EYE aims to simplify the process of testing new WASP developments in realistic conditions on public roads easier (with a reference implementation) and safer (with an approved safety case and process).
Project members academia: Martin Törngren, Professor, KTH Royal Institute of Technology; Ahmed Terra, WASP industrial PhD student, Ericsson; Magnus Gyllenhammar, WASP Industrial PhD student, ZenseAct; Rusyadi Ramli, WASP affiliated PhD student, KTH Royal Institute of Technology; Lars Svensson, WASP affiliated PhD student, KTH Royal Institute of Technology; Naveen Mohan, Affiliated WASP PhD student, KTH Royal Institute of Technology.
Project members industry: Håkan Sivencrona, Senior safety expert, ZenzeAct; Elena Fersman, Director, AI Research, Ericsson; Mats Nordlund, Director Research, ZenseAct; Rafia Inam, Project Manager in AI, Ericsson.
DARMA_Bridge: Data-driven Foundations for Robust Deformable Object Manipulation
Countless manufacturing and every-day tasks require handling of non-rigid objects, for example, in food and clothing manipulation. Yet, the dynamics of deformable objects which vary widely, are complex and difficult to both model and simulate. This makes robotic manipulation of such objects using conventional control a challenge.
DARMA is a collaboration group formed under a WASP collaborative project grant which aims at developing learning-based control strategies for robotic manipulation of deformable objects. DARMA_bridge will help the transition of the DARMA project from simulation-based Reinforcement Learning for robotic manipulation to real-world experimental scenarios. This will be done both by investigating sim-to-real transfer methods, as well as by developing a custom robotic deformable object data collection system. In terms of software, our endeavor will be supported by our industrial partner Algoryx. To integrate our results on the YuMi robot, we will rely, in a great extent, on interaction with our other industrial partner, ABB. Further, our work can also lead to useful tools for the newly created WARA Robotics.
The added funding provided by the Industry Bridge project will greatly facilitate the development of both the software infrastructure needed for the sim-to-real research, as well as the considerable engineering time and hardware infrastructure which is needed for the completion of the robotic data collection system. In the end, the data collected will be used in conjunction with the simulation-to-real implementation to complement what is learned in simulation with knowledge of real-world measurements, potentially maximizing the success of our methods.
Project members academia: Yiannis Karayiannidis (co-PI), Associate Professor, Chalmers; Rita Laezza, PhD student, Chalmers; Florian Pokorny (co-PI), Associate Professor, KTH Royal Institute of Technology; Robert Gieselmann, PhD student, KTH Royal Institute of Technology.
Project members industry: ABB, Algoryx
WCIB: WARA Common Information Bridge for Enabling Research on Data-Driven Cloud Operations
Cloud and Edge data centers are growing in size, complexity, and heterogeneity to the point where manual management is becoming intractable. Autonomous management systems – which can identify trends and anomalies, efficiently allocate resources, and improve overall system and application performance – are becoming vital, but the “full stack” data necessary to develop innovative new approaches for these systems is very difficult (and often impossible) to obtain.
The WCIB project will create an online portal to allow WASP researchers to extract large volumes of operations data from a live Cloud facility – the Ericsson Research Data Centre (ER DC) in Lund. This data is truly “full stack”, ranging from hardware and facility systems to network traffic and software workload metrics. WCIB will expose over 1 TB of explorable operations data per day – making this the largest and richest source of full-stack data center operations data available anywhere in the world.
WCIB addresses several key engineering and scientific challenges related to correlating and integrating operational Cloud data. In addition to developing the online portal, scientific challenges addressed include: 1) automated identification and selection of sampling parameters; 2) dynamic integration of data; 3) automated “data cleaning” mechanisms to ensure user confidentiality without affecting semantics, through re-labeling etc. Funding from WASP will be used to employ a Research Engineer full-time for 12 months, working alongside Ericsson data center engineers to connect databases and metric stores, and develop the methods and tools necessary to create this state-of-the-art data resource. The project starts on 2021-09-01.
Project members academia: Paul Townend, Associate Professor; Erik Elmroth, Professor; P-O Östberg, Associate Professor; Johan Tordsson, Associate Professor; Monowar Bhuyan, Assistant Professor; Prashant Singh, Assistant Professor, Umeå University.
Project members industry: Johan Eker, Principal Researcher; Torgny Holmberg, Senior Researcher; Robert Marklund, Senior Researcher; Rodrigo Berg, Project Manager ER DC, Ericsson Research.
Published: March 31st, 2021