Booth CeBit 2018

Date: 12.06.2018
Venue: Hannover
Duration: 4 Days

TUM presented the Neurorobotics Platform and our Robot Mice at the CeBIT.
For one week you could see our robots live and learn how they and the NRP are used within the HBP research.
Thanks to the many people showing interest in the NRP and our Robots.



NRP Workshop at ERF2018

Date: 13.03.2018
Venue: Tampere
Duration: 3 Days

We organized a Workshop titled “Beyond conventional AI – the Neurorobotics Platform” that attracted significantly more people than last year during ERF2017 in Edinburgh.

Read more about this event on our blog:

Learn more about the ERF here:



24. Handelsblatt Jahrestagung Strategisches IT-Management 2018

Date: 16.01.2018
Venue: Sofitel Munich Bayerpost

Alois Knoll gave a keynote talk on neurorobotics and the Human Brain Project at the conference Handelsblatt Jahrestagung Strategisches IT-Management 2018. IT managers learned how HBP technology such as the Neurorobotics Platform can be used as tools for creating innovative brain-derived products and solutions. In the break, participants had the chance to meet our robot mouse.


HBP Summit 2017

Date: 17.10.2017
Venue: Glasgow

Duration: 4 days

The Human Brain Project (HBP) Summit is a unique forum for the HBP Consortium and its Partnering Projects to come together to present and learn about the latest scientific results and other project achievements, as well as to develop new ideas, plan next steps and network. The event consists of a 3-day interactive scientific programme, comprising plenary sessions, external keynote presentations, hackathons, demos, scientific highlight sessions and cross-disciplinary workshops.


User Workshop @ FZI Karlsruhe

Date: 24.07.2017
Venue: FZI, Karlsruhe, Germany
Duration: 3 Days

Thanks to all of the 17 participants for making this workshop a great time.

We held a successful Neurorobotics Platform (NRP) User Workshop in FZI, Karlsruhe.  We welcomed 17 attendants over three days, coming from various sub-projects (such as Martin Pearson, SP3) and HBP outsiders (Carmen Peláez-Moreno and  Francisco José Valverde Albacete). We focused on hands-on sessions so that users got comfortable using the NRP themselves.


Thanks to our live boot image with the NRP pre-installed, even users who did not follow the local installation steps beforehand could run the platform locally in no time. During the first day, we provided a tutorial experiment, exclusively developed for the event, which walked the users through the many features of the NRP. This tutorial experiment is inspired from the baby playing ping pong video, which is here simulated with an iCub robot. This tutorial experiment will soon get released with the official build of the platform.


On the second and third days, more freedom was given to the users so that they could implement their own experiments. We had short hands-on sessions on the Robot Designer as well as Virtual Coach, for offline optimization and analysis. Many new experiments were successfully integrated into the platform: the Miro robot from Consequential Robotics,  a snake-like robot moving with Central Patterns Generators (CPG), revival of the Lauron experiment, …


We received great feedback from the users. We are looking forward for the organization of the next NRP User Workshop!

3rd Japan-EU Workshop

Date: 15.06.2017
Venue: Biotech Campus Geneva
Duration: 1 day
In conjunction with the symposium „Building Bodies for Brains & Brains for Bodies“ on June 16 we had our 3rd Japan-EU Workshop on Neurorobotics in Geneva.
Speakers included:
Yoshihiko Nakamura (University of Tokyo)
Satoshi Oota (RIKEN)
Poramate Manoonpong (University of Southern Denmark),
Aaron Sloman (University of Birmingham)
It was the third in a hopefully now established series of workshops. For the past two workshops in this series in 2015 and 2016 see:

Symposium: Building Bodies for Brains & Brains for Bodies

Date: 16.06.2017
Venue: Biotech Campus Geneva
Duration: 1 day

It was a one-day symposium in the field of neurorobotics with the goal of improving robot behavior by exploiting ideas from neuroscience and investigating brain function using real physical robots or simulations thereof.
Contributions to this workshop focussed on (but not limited to) the relation between neural systems – artificial or biological – and soft-material robotic platforms, in particular the “control” of such systems by capitalizing on their intrinsic dynamical characteristics like stiffness, viscosity and compliance.


Developer Workshop 1.3 at Fortiss

Date: 06.04.2017
Venue: Fortiss
Duration: 2 days

As the NRP development team is in the process of releasing 1.2, it holds on 6th and 7th April its 6-monthly developer workshop, this time at fortiss in Munich. This workshop closes the 1.2 release cycle with technical presentations of the new features by developers, updates from our scientistific coordinator and a scrum release retrospective. It also opens the new 1.3 release cycle with new epics presentation by the product owner and scrum client followed by a full day of backlog grooming and road mapping for the next 6 months.
Moreover, this event is an occasion for developers to meet all in one place and exchange views and issues.



NRP Install Party

Date: 04.04.2017
Venue: Fortiss, Munich
Duration: 2 days

Selected SP10 scientists from all over Europe came to Fortiss for hands-on sessions with NRP developers and advanced help from scientific experts for their simulated experiments. With one expert dedicated to helping one researcher, we were able to adapt these user experiments to our platform while using it to its full potential during these two days in Munich.
In a continuing effort to use scientific requirements as the driving force behind our platform development, we were able to use this event to collect feature requests and problems to be addressed.


Developer Workshop in Pisa

Date: 17.10.2016
Venue: Pisa
Duration: 2 days

The developer workshop in Pontedera brought together the developers of our subproject. In a 2 day workshop, the recent Summit was discussed to draw conclusions from the feedback we had received.
Addressing the most pressing issues that had been discussed during the Summit was only part of that workshop. Another important aspect was to align the development with what users actually need by inviting a couple of researchers and making sure our plans were in alignment with theirs.
Then, working together, our roadmap was refined to ensure good progress for the months to come.hbp_small

NEST Workshop in Karlsruhe

Date: 03.11.2016
Venue: FZI, Karlsruhe
Duration: 2 days

The NEST workshop in Karlsruhe received a lot of attention with significantly more attendees than expected.

Presentations covered topics such as
– NEST: Current Developments (Hans Ekkehard Plesser & Susanne Kunkel)
– Neurorobotics and NEST in the HBP (Marc-Oliver Gewaltig)
– Modeling the Cerebellum (Egidio d’Angelo)
– Cerebellum Modelling with NEST (Alberto Antonietti)
– Interfacing the Neurorobotics Platform using MUSIC (Martin Schulze)
– The Potjans-Diesmann local microcircuit model using different neuron classes for excitatory and inhibitory neurons (Nilton Kamiji)
– Interactive visualization and steering of structural plasticity in NEST (Sandra Diaz)
– NestMC: A new multi-compartment neuron simulator (Alex Peyser)
– NEST, MUSIC, and ROS (Philipp Weidel)
– Porting WaveScales to NEST (Elena Pastorelli)


You can find more on the workshop website here:


Workshop at ERF2017

Date: 22.03.2017
Venue: Edinburgh
Duration: 3 Days

After the acceptance of our workshop proposal we presented the NRP during our session “Introduction to the HBP Neurorobotics Platform” at ERF2017 in Edinburgh and were able to engage in lively discussions about the platform, our roadmap and neurorobotics in general.


HBP Exhibition

date: 29.11.2016
venue: European Parliament
duration: 2 days
The first Human Brain Project exhibition took place at the European Parliament on 29 – 30 November 2016.


The HBP exhibition highlights the unique contributions HBP is making to brain research and how these contributions are benefiting European science, competitiveness and society and positioning Europe promanently among the growing number of large-scale brain initiatives worldwide. The exhibition was created as part of HBP’s participation in the workshop Understanding the Human Brain – A New Era of Big Neuroscience, organised by the European Parliament’s Science Technology Options Assessment group (STOA) on 29 November 2016.

The text and image above are taken from a booklet that you can find here: stoa_booklet-v-final_digital

First NRP User Workshop

Date: 11.01.2017
Venue: TU München
Duration: 1 Day

With the NRP running very smoothely, it was now time to invite actual external users to show our platform to them. A second part of this workshop was to understand their research requirements and get feedback from them. This helped us to understand what we need to implement in order for them to really use our product.

Below, you can see an impression from one of the presentations.


HBP Summit 2016

Date: 12.10.2016
Venue: Florence

Duration: 3 days

The HBP held its 4th annual Summit on 12 to 15 October in the vibrant city of Florence, Italy. The event was hosted by Laboratorio Europeo di Spettroscopie Non Lineari (European Laboratory for Non-Linear Spectroscopy; LENS), one of the key European institutions for light wave research, which has been a centre of excellence at the University of Florence since 1991.
The 2016 HBP was launched with ‘Open Day’ to enable dialogue and collaboration between the HBP, other scientific communities and the public. The Open Day took place on 12 October—for more information, please see the Open Day webpage.

Nearly 450 members of the HBP consortium, representatives of HBP’s first Partnering Projects, National Research Funding Agencies, and representatives from Italy’s national research council participated in the Summit. The programme featured a diverse mixture of plenary sessions, parallel working group meetings, a poster session with over 100 posters and networking opportunities. The sessions covered an array of scientific themes, a town hall meeting, an introduction to HBP’s first Partnering Projects, a young investigator’s plenary session, and a panel debate on “the Conscious Computer.” In a dedicated session on “the local context”, representatives of Italy’s National Research Council, the University of Florence and the Ministry of Research and Education provided insights about the role of the Flagship and benefits of a large science project for science and research in Italy. For more information on the sessions, please see the programme.

Florence, the capital of the Tuscany region of Italy, is considered one of the world’s most beautiful cities. It is particularly noted for its Renaissance art and architecture, and its historic centre was declared a World Heritage Site by UNESCO in 1982.

The HBP was proud to hold its Summit in such a magnificent location.


(Text and image from

Understanding the human brain

Date: 29.11.2016
Venue: European Parliament Brussels

At the event “Understanding the human brain – a new era of big neuroscience” hosted by the European Parliament in Brussels this week, the HBP sent delegates to present our work. The SP10 did not miss out on this opportunity either.

The picture shows Prof. Alois Knoll with Tetsuo Yamamori and Roboy, a cutting edge technology robot in the HBP.


Click here to learn more about Roboy:

Click this link to learn more about the event:

More for less – Energy Efficiency in Neural Locomotion – Workshop

Date: 21.09.2016
Venue: Berlin

During September 21-23, 2016 as part of the Bernstein Conference in Berlin, the workshop for “Energy Efficiency in Neural Locomotion” (Stratmann, Röhrbein) will take place.

The following links will give you more detail about the event.

From Rats to Robot Navigation and Beyond

Date: 04.07.2016
Venue: TUM Campus Garching

Lecture Series in Neuro-Robotics

Speaker: Prof. Michael Milford, Queensland University of Technology

Prof. Milford from Queensland University of Technology will give a talk on the topic of From Rats to Robot Navigation and Beyond here in Munich in our new lecture series in Neuro-Robotics. You are cordially invited to join his fascinating presentation.



Intelligence Emerging from Neural Activities

Date: 01.07.2016
Venue: IAS Auditorium

Lecture Series in Neuro-Robotics

Speaker: Prof. Hirokazu Takahashi, The University of Tokyo

Prof. Takahashi from The University of Tokyo will give a talk on the topic of Intelligence Emerging from Neural Activities here in Munich in our new lecture series in Neuro-Robotics. You are cordially invited to join his fascinating presentation.



2nd Japan-EU Workshop on Neurorobotics

Date: 29.04.2016
Venue: University of Tokyo

Building robots with brain-inspired control structures is one of the central goals of the members working in the Neurorobotics sub-project of the HBP. They are, however, not alone in this task and so it should come as no surprise that also other researchers have made great progress in recent years. Research teams in Japan have made astonishing discoveries and provided fascinating insights into complex matters concerning biological systems such as brain structures and musculoskeletal models.

So, with researchers both in Europe and in Japan making great steps towards a better understanding of biological systems, the Japan-EU Workshop on Neurorobotics held in Tokyo aimed at bringing those experts together. Presentations from both sides gave an overview of their work and highlighted their latest achievements. The presentations were followed by lively discussions, which allowed the participants to engage and fully comprehend the potential advantage arising from gaining access to both sides’ resources.

To give a practical impression of what the University of Tokyo brings to the table, the students presented their research and the hydraulic robot Hydra (see Figure below).

With researchers from many different backgrounds and with various fields of interest a wide foundation was laid for further ventures. And even though the research was overlapping in some aspects, the large number of areas of expertise made it apparent how both parties in this workshop might benefit from the opportunities offered by a cooperation.

A big step in the direction of working together in the future was made and it stands to reason that a joint research is not far now and that the two teams can join forces in the near future to ultimately unravel the mysteries of the human brain in all its complexity.

Pictures, presentations and links to publications relevant to the workshop can be found on the SP’s homepage



BioRob 2016

Date: 26.06.2016
Venue: Singapore

On June 26th, 2016 a Neurorobotics workshop will take place as part of the “6th IEEE RAS & EMBS – International Conference on Biomedical Robotics and Biomechatronics”, the BioRob 2016, in Singapore.

This years motto: New perspectives in the synergy between Neuroscience and Robotics”


The following link will lead you to more detailed information on the event.

Wo/Man, Mind, Machine

Date: 13.06.2016
Venue: Berlin-Brandenburgische Akademie der Wissenschaften (BBAW), Einstein-Saal, Jägerstr. 22/23,  10117 Berlin

In Cooperation with the “Israel Young Academy” the “Young Academy” hosts a two day conference, June 13-14, 2016 in Berlin, Germany.

The following is the timeline for the event:


MONDAY, 13 JUNE 2016

10:00 Welcome & Introduction

10:15 Keynote: Aldo Faisal (Neurotechnology, Imperial College London): “Perception, Action, and the Grammar of Behaviour”

11:15–11:45 Coffee Break

11:45–13:15 Session 1: The Social Brain – How We (Fail to) Understand Others. Philipp Kanske & Mona Khoury-Kassabri

13:15–15:00 Lunch Break

15:00–16:00 Session 1 (continued)

16:00–16:30 Coffee Break

16:30–18:00 Session 2: Artificial Minds in Film, Performance, and Fiction. Caspar Battegay & Nissim Otmazgin

18:15 Evening lecture: Jennifer Parker-Starbuck (Theatre & Performance Studies, University of Roehampton, London ): “Cyborg Encounters”

19:15 Reception

20:30 Dinner


09:30 Keynote: Alexis Kirke (Computer Music, University of Plymouth): “For entertainment purposes only? How arts can contribute directly to research into the science of computing and HCI”

10:30–11:00 Coffee Break

11:00 Session 3: Mind, Man, Machine, and Music. Miriam Akkermann & Sharon Aronson-Lehavi

13:00–14:30 Lunch Break

14:30 Session 4: Processing Information – Brains, Bodies, and Bacteria. Sibylle Baumbach & Ron Milo
With Florian Röhrbein (Informatics, Technical University of Munich) on Neurorobotics and The Human Brain Project

16:30–17:00 Coffee Break

17:00 Roundtable & Final Discussion

17:30 End of the Conference


For more information, you can also take a look at the flyer:


Kickoff Meeting of SGA1 2016

Date: 07.06.2016
Venue: Geneva, CBS Building B1, 9 chemin des Mines, 1202 GE

hbp_smallWe are excited to announce the following agenda for the Kickoff Meeting of SGA1 during the 7th and 8th of June 2016 in Geneva.




9:15 – 10:00

General introduction (Marc-Oliver Gewaltig, EPFL)
What is SP10? What is SGA1? What is a Co-Design project?
What are the different roles in the project?
Scientific Coordinator election
10:00 – 10:15

Coffee break
10:15 – 11:45

Round table (5 minutes per institution)
11:45 – 12:00

SP10 Platform demo (Igor Peric, Georg Hinkel, FZI)
Live demo and discussion
12:00 – 13:00

Lunch (CBS Cafeteria)
13:00 – 13:05

Election of the Scientific Coordinator
13:05 – 13:15

Software development road map and methodology (Axel von Arnim, Fortiss)
13:15 – 15:00

Scientific experiments: research planning (chaired by Scientific Coordinator)
Summary of proposed experiments
Small research groups draft scientific experiments that will drive the
development of the Neurorobotics platform
15:00 – 15:15

Coffee break
15:15 – 16:45

Summary of research planning (chaired by Scientific coordinator)
Presentation of collected ideas,
Consolidation, writing of roadmap
Questions and Answers
16:45 – 18:30

Second Periodic Review preparation
(review presenters only)
19:30 – 21:00

SP10 networking Diner (restaurant “La Bagatelle”)

9:15 – 10:15

HBP resources
Resources (Axel von Arnim, Fortiss), visualization (Cyrille Favreau, BBP)
10:15 – 10:30

Coffee break
10:30 – 12:00

HBP services
Collab (Olivier Amblet, BBP)
12:00 – 13:00

Lunch (CBS Caféteria)
13:00 – 15:00

Epic grooming (SP10 Development Team – no researchers involved)
15:00 – 15:15

Kickoff wrap-up and closing (Prof. Alois Knoll, TUM)


Neurorobotics Learning Event in Tokyo

Date: 18.04.2015
Venue: University of Tokyo

The HBP Neurorobotics Team in cooperation with the University of Tokyo organized a whole day workshop focusing on Neurorobotics. The workshop took place on the 18th of April at University of Tokyo. The workshop was sponsored by the HBP and the Global University Project, MEXT. Further details can be found on the webpages: and

Here some impressions from the workshop:



Symposium Neurorobotics: Chances and Challenges

Venue: TUM IAS, Munich


Place: TUM Institute for Advanced Study & TUM Department of Informatics, Garching (close to Munich)

 Topic (in german):

Unser Gehirn. Es ist ständig aktiv, egal ob wir wach sind oder schlafen. Es ist der Ort, an dem unsere Ideen entstehen, der uns Freude ebenso wie Angst empfinden lässt. Es vollbringt jede Sekunde kognitive Höchstleistungen und lässt uns dabei alltägliche Aufgaben ebenso erledigen wie komplexe Zusammenhänge verstehen. Doch wie funktioniert das Gehirn eigentlich?

Üblicherweise sind empirische Untersuchungen basierend auf der Analyse von Gewebeproben, bildgebenden Verfahren wie fMRT, Messungen mittels EEG oder kognitiven Experimenten die Basis moderner Neurowissenschaften. Dank der immer einfacheren und günstigeren Verfügbarkeit hoher Rechenleistung hat sich in den letzten Jahren eine neue Perspektive für die Gehirnforschung ergeben: Statt Experimenten kommen vermehrt Computersimulationen zum Einsatz, anhand derer Dynamik und Logik neuronaler Schaltkreise analysiert werden können.

Das resultierende Verhalten bei Interaktion mit der Umwelt wird in der Neurorobotik untersucht, indem die simulierten neuronalen Netze als „künstliche Gehirne“ zur Steuerung von Robotern verwendet werden.


Invited Speakers:

Priv.-Doz. Dr. Stefan Artmann
Nationale Akademie der Wissenschaften Leopoldina
Friedrich-Schiller-Universität Jena

Prof. Dr. Godehard Brüntrup SJ
Hochschule für Philosophie München

Priv.-Doz. Dr. Agnieszka Wykowska
LMU München



October 8, 2015 (1st day)

13:00 – 13:30 Welcome by the organizers
13:30 – 14:30 Talk by Florian Röhrbein:

Neurorobotics & the Human Brain Project

14:30 – 15:30 Talk by Godehard Brüntrup:

Was hieße es für einen Roboter, denken zu

15:30 – 16:00 Coffee break & discussion time
16:00 – 17:00 Talk by Florian Röhrbein & Florian Walter:

Introduction to Neuroscience

17:00 – 18:00 Discussion
18:00 – 19:00 Dinner

October 9, 2015 (2nd day)

09:00 – 10:30 Talk by Florian Walter:

Neurorobotics – State of the Art

10:30 – 11:30 Talk by Stefan Artmann:

Künstliches Bewusstsein – ein sinnvolles Ziel der
Künstlichen Intelligenz?

11:30 – 12:30 Lunch
12:30 – 13:30 Talk by Agnieszka Wykowska:

Linking social cognitive neuroscience with
robotics for better social robots

13:30 – 14:30 World Cafe – Discussion
14:30 – 15:00 Coffee break
15:00 – 16:00 Webinar:

Brain Computer interfaces and neuroprosthetics

16:00 – 17:00 Farewell


You can dowload the complete schedule and information about all speaker as PDF: Download

Advances in Biologically Inspired Brain-Like Cognition and Control for Learning Robots

Date: 28.09.2015
Venue: Hamburg

!!For interest!!

A summary of the workshop can be found on robohub:



Workshop for:

2015 IEEE/RSJ International Conference on Intelligent Robots and Systems
September 28 – October 03, 2015, Hamburg, Germany

The workshop will take place on Friday, October 2, 8-12 am.

To check the proceedings for our workshop, just click on the link.



Florian Walter (TUM)

Florian Röhrbein (TUM)

Stefan Ulbrich (FZI)

Rüdiger Dillmann (KIT)



Even today’s most advanced robots perform poorly in executing simple everyday tasks carried out routinely by humans and animals. This fact has very early motivated researchers to adopt neurobiological principles of cognition and control in robotics, that resulted in numerous approaches based on artificial neural networks and machine learning. However, many originally proposed methods employ neural networks solely for the purpose of approximating and replicating standard control architectures. At the same time, research focusing more on biological plausibility severely suffered from limitations on size and accuracy of the neural simulations imposed by hardware constraints.

In the recent years, advanced insights and the increasing availability of cheap processing power have brought new momentum to the field which has evolved into two tracks of research with different goals and methods. In the new and emerging field of neurorobotics, the focus is on a close correspondence to experimental findings from neurosciences. Detailed simulations of spiking neural networks and the use of biologically plausible neural learning rules are therefore more important than mathematical tractability or implementation efficiency and enable a seamless exchange of results between both disciplines. In contrast, new approaches extending the theory of classical artificial neural networks mostly rely on simpler neuron models but integrate them at much larger scales or embed them in novel network architectures.

This workshop seeks to provide a platform to present and discuss advances in biologically inspired brain-like cognition and control for robotic applications based on both biologically plausible and artificial neural networks. By bringing together experts from both fields, we intend to foster a fruitful exchange between the different communities. Furthermore, the global scope of the workshop furthermore makes participation also equally attractive for researchers new to the field. Interactive demonstrations of tools and implementations will help to get all attendants actively involved. A dedicated poster session additionally offers enough room for individual discussion.


Topics of interest:

  • Neurorobotics
  • Bio-inspired learning
  • Spiking neural networks for robotic applications
  • Self-organization
  • Embodiment
  • Applications of neuromorphic hardware designs in robotics
  • Reservoir computing
  • Deep learning

This workshop is supported by Wataru Takano (University of Tokyo, Japan), Co-chair of the IEEE RAS Technical Committee on Robot Learning and by Kin Huat Low (Nanyang Technological University, Singapore), Chair of the Technical Committee on Biorobotics


Invited Speakers:

  1. Auke Jan Ijspeert (confirmed)
    Biorobotics Laboratory, École Polytechnique Fédérale de Lausanne, Switzerland
  2. Herbert Jäger (confirmed)
    Jacobs University, Germany
  3. Joni Dambre (confirmed)
    Reservoir Computing Lab, Ghent University, Belgium
  4. Manfred Hild (confirmed)
    Neurorobotics Research Laboratory, Beuth Hochschule für Technik Berlin,Germany
  5. Jason Yosinski (confirmed)
    Cornell Creative Machines Lab, Cornell University, USA


Tentative schedule:

      8:30   –     8:45 Welcome
      8:45   –     9:15 Invited Talk 1:
Auke Jan Ijspeert
      9:15   –     9:45 Invited Talk 2:
Manfred Hild
      9:45   –   10:15 Invited Talk 3:
Joni Dambre
    10:15   –   10:40 Live Demo Session: The HBP Neurorobotics Platform
    10:40   –   11:00 Coffee Break
    11:30   –   12:00 Invited Talk 4:
Herbert Jäger
    12:00   –   12:30 Invited Talk 5:
Jason Yosinski



Learning about Animal Locomotion Control from Robots, Auke Ijspeert

The ability to efficiently move in complex environments is a fundamental property both for animals and for robots, and the problem of locomotion and movement control is an area in which neuroscience and robotics can fruitfully interact.

Animal locomotion control is in a large part based on spinal cord circuits that combine reflex loops and central pattern generators (CPGs), i.e. neural networks capable of producing complex rhythmic or discrete patterns while being activated and modulated by relatively simple control signals. These networks are located in the spinal cord for vertebrate animals and interact with the musculoskeletal system to provide “motor primitives” for higher parts of the brain, i.e. building blocks of motor control that can be activated and combined to generate rich movements.

In this talk, I will present how we model the spinal cord circuits of lower vertebrates (lamprey and salamander) using systems of coupled oscillators, and how we test these models on board of amphibious robots. The models and robots were instrumental in testing some novel hypotheses concerning the mechanisms of gait transition, sensory feedback integration, and generation of rich motor skills in vertebrate animals.

I will also discuss how the models can be extended to control biped locomotion, and how they can help deciphering the respective roles of pattern generation, reflex loops, and descending modulation in human locomotion.


Self-Exploration of Autonomous Robots Using Attractor-Based Behavior Control, Manfred Hild

An autonomous robot which is equipped with sensorimotor loops and situated within a specific environment can be regarded as a dynamical system. In the language of dynamical systems theory, behavioral body postures and repetitive body motions then correspond with fixed points and quasiperiodic orbits. Both of which can either be naturally stable, i.e., attractors of the situated physical body, or be stabilized by the whole system including the sensorimotor loops.

As is well-known, even most simple two-neuron networks may already exhibit many co-existing attractors, which, if properly chosen, may nicely go along with the overt behavior of an autonomous robot. Standing and walking is one example, obstacle avoidance another one. The question arises, how larger neural networks can be found (preferable by the robot itself), that explore behavioral options, starting from scratch and getting increasingly rich over time.

Attractor-Based Behavior Control (ABC) follows the afore-mentioned path and confidently finds attractors which correspond to energy-efficient behavioral body postures, either fully alone or in gently guiding physical human-machine interaction. The latter helps protecting the robot from harming itself as it would easily happen if using, e.g., motor-babbling or homeokinetic learning rules. In addition, ABC learning does not only find behavioral attractors, but also the corresponding attractor-connecting heteroclinic orbits which can be utilized to generate stable motion trajectories.

After briefly revisiting the necessary concepts, I will introduce ABC-Learning and demonstrate how it enables an autonomous robot to self-explore its behavioral capabilities from scratch and without any given body model.

Practical approaches to exploiting body dynamics in robot motor control, Joni Dambre

Motor control systems in the brain of humans and mammals are hierarchically organised, with each level controlling increasingly complex motor actions. Each level is controlled by the higher levels and also receives sensory and/or proprioceptive feedback. Through learning, this hierarchical structure adapts to its body, its sensors and the way these interact with the environment.

An even more integrated view is taken in morphological or embodied computation. On the one hand, there is both biological and mechanical (robotics) evidence that a properly chosen body morphology can drastically facilitate control when the body dynamics naturally generate low level motion primitives. On the other hand, several papers have used robot bodies as reservoirs in a reservoir computing setup. In some cases, reservoir computing was used as an easy way to obtain robust linear feedback controllers for locomotion.

In other cases, the body dynamics of soft robots were shown to perform general computations in response to some input stimulation. In general, very specific highly compliant bodies were used. At Ghent University’s Reservoir Lab, we have previously used reservoir computing to generate locomotion on quite different robot platforms: the highly compliant tensegrity robot Recter and the far less compliant quadruped robot Oncilla and a new low cost modular quadruped puppy robot. In all cases, we succeeded in generating stable gaits. However, not surprisingly, not all robot bodies are equally suitable to help generating their own motor actuations. As a result, the reservoir computing principle alone was not always sufficient.

We present an overview of our experience with these different robot platforms and give practical guidelines for applying physical reservoir computing to new robots. We finally discuss some perspectives on a more systematic evaluation between body morphology, compliance and the complexity of generating stable gaits for locomotion.

Generating and Modulating Complex Motion Patterns with Recurrent Neural Networks and Conceptors, Herbert Jaeger

In biological brains “higher” cognitive control modules regulate “lower” brain layers in many ways. Examples for such top-down processing pathways include triggering motion commands (“reach for that cup”), modulating ongoing lower-level motor pattern generation (“wider steps, slooow down!”), setting attentional focus (“look closer… there!”), or predicting the next sensory impressions (“oops – that will hit me”). Not much is known about computational mechanisms which would implement such top-down governance functions on the neural level. As a consequence, in machine learning systems which are based on artificial neural networks, top-down regulation is rarely implemented.

Specifically, today’s top-performing pattern recognition systems (“deep learning” architectures) do not exploit top-down regulation pathways. This talk gives an introduction to a novel neural control mechanism which addresses such top-down governance mechanisms in modular, neural learning architectures. This computational principle, called conceptors, allows higher neural modules to control lower ones in a dynamical, online-adaptive fashion. The conceptor mechanism lends itself to numerous purposes:

  • A single neural network can learn a large number of different dynamical patterns (e.g. words, or motions).
    • After some patterns have been learnt by a neural network, it can re-generate not only the learnt “prototypes” but a large collection of morphed, combined, or abstracted patterns.
    • Patterns learnt by a neural network can become logically combined with operations AND, OR, NOT subject to rules of Boolean logic. This reveals a fundamental link between the worlds of “subsymbolic” neural dynamics and of “symbolic” cognitive operations.
    • This intimate connection between the worlds of neural dynamics and logical-symbolic operations yields novel algorithms and architectures for lifelong learning, signal filtering, attending to particular signal sources (“party talk” effect), and more.

Expressed in a nutshell, conceptors enable “top-down logico-conceptual control” of the nonlinear, pattern-generating dynamics of recurrent neural networks.

Training and Understanding Deep Neural Networks for Robotics, Design, and Perception, Jason Yosinski

Artificial Neural Networks (ANNs) form a powerful class of models with both theoretical and practical advantages. Networks with more than one hidden layer (deep neural networks) compute multiple functions on later layers that share the use of intermediate results computed on earlier layers. This compositional, hierarchical structure provides a strong bias, or regularization, toward solutions that seem to work well on a large variety of real-world problems.

In this talk we will examine this bias in action via several vignettes. First we will look at a method for using ANNs to learn fast gaits for walking robots. Second, we will see how the same method can be applied to design three dimensional solid objects. Finally we will discuss a few simple experiments that shed light on the inner workings of neural nets trained to classify images. The studies shed light on the computation performed by each layer of a network and by the network as a whole. The experiments taken together reveal some surprising behaviors of large networks and lead to a greater understanding and intuition for the computation performed by deep neural nets.

The Neurorobotics Platform of the Human Brain Project, Stefan Ulbrich

The Neurorobotics Platform (NRP) is a web-based simulation environment for neuroscientists for performing neurorobotic experiments. It is developed in the sub-project 10 “Neurorobotics” of the Future and Emerging Technologies (FET) Flagship project “The Human Brain Project” funded by the European Commission. The software grants neuroscientists painless access to sophisticated brain, robot and physical simulators and provides the necessary tools for designing brain-body interfaces, virtual world and robot modeling as well as the definition of complex experiments. The NRP is still under development and this talk presents its current state and capabilities as well as an outlook on future development.



1. Neurocognitive Architecture for Autonomous Task Recognition, Learning, and Execution (NARLE)

Yulia Sandamirskaya (1), Mikhail Burtsev (2)
(1) Institute of Neuroinformatics, University of Zurich and ETH Zurich, Switzerland
(2) NBICS Centre, NRC “Kurchatov Institute”, Moscow, Russian Federation

Humans and higher animals are able to solve new problems and reuse the acquired solutions later in their behaviour. This capacity of the biological neuronal systems for constant adaptation enables biological agents to refine their skills and expand their behavioural repertoire. Understanding how neuronal dynamics may be organised to enable learning new skills will enable us to build cognitive robots that are adaptive and can autonomously acquire new behaviors. Recently, we have introduced a neurocognitive architecture that addresses the problem of learning and execution of hierarchical behaviors and complex skills.

The proposed architecture uses Dynamic Neural Fields (DNFs) to implement low-level motor and perceptual behaviours and a Functional System Network (FSN) to tie these behaviors in goal-directed sequences. The DNFs enable a continuous, dynamical representation of perceptual features and motor parameters, which may be coupled to the robot’s sensors and effectors. Attractor states and instabilities of the DNFs account for segregation of discrete cognitive states and mark behaviorally relevant events in the continuous flow of the sensorimotor dynamics. The FSN, in its turn, comprises dynamical elements that can be arranged in a multilayered network by a learning process, in which new layers and elementary behaviors are added on demand.

Combination of the DNF and FSN frameworks in a neurocognitive architecture NARLE enables pervasive learning and adaptation both on the level of individual behaviors and goal-directed sequence, leading to more adaptive intelligent robotic systems, capable to learn new tasks and extend their behavioral repertoire in stochastic real-world environments. In our robotic scenario, the systems for object and scene representation, action parsing, and sequence learning are detailed and their interplay in a table-top assembly task is described. We discuss the potential of this new biologically-inspired cognitive architecture in the fields of modelling human cognition and autonomous robotics.

2. Learning Techniques for Neurorobotics – A Survey on the Role of the Factor Time

Florian Walter, Florian Röhrbein, Alois Knoll
Chair for Robotics and Embedded Systems, Technische Universität München, Germany

Neurorobotics enables the interaction of simulated biological neural networks with both virtual and real environments in closed perception-action loops. Controlling robotic actuators, processing sensor readings and implementing goal-directed behavior requires the adjustment of synaptic weights by means of learning. The temporal dynamics of the detailed neural models employed in neurorobotics enable the use of learning techniques which incorporate a notion of time. This poster provides an overview of concepts and methods from this field with a special focus on prospective applications in neurorobotics.

3. Cognitive Walking: Recruitment of an Internal Body Model

Malte Schilling, Holk Cruse
Center of Excellence Cognitive Interaction Technology, University of Bielefeld, Germany

The overall goal of our approach is to understand general cognitive mechanism and to implement those in a functional minimal cognitive system. We are starting from a biological inspired hexapod walking control system and we extended this system towards a minimal cognitive system. In our understanding Cognition is understood following the notion of McFarland and Bösser (1993) as the ability to plan ahead by means of an internal simulation (Hesslow, 2002) relying on internal representation (Steels, 2003; Glenberg, 1997) which are grounded in embodied experiences (Gallese & Lakoff 2005). The extension of our system leads towards a control system for the six-legged walking robot Hector which allows the robot to deal with novel situation and to plan its behavior in advance.

Crucial to this approach is, on the one hand, the grounding of internal representation in behavior and, on the other hand, the flexible reuse of such embodied internal models in new contexts which is realized as a form of internal simulation. On the poster, we will present, first, the underlying behavior-based control system from which adaptive and well-cordinated walking behavior emerges. Second, how functional internal models are grounded in lower level behavior. As a first internal representation an internal model of the body will be introduced realized as a simple recurrent neural network. Third, we will show how this functional internal model of the body can be exploited in internal simulation for planning ahead exploiting the predictive forward function of the model.

4. Real-time Cerebellar Control of a Compliant Robotic Arm

Christoph Richter (1), Sören Jentzsch (2), Florian Röhrbein (3), Patrick van der Smagt (3), Jörg Conradt (1)
(1) Electrical and Computer Engineering, Technische Universität München, Germany
(2) fortiss GmbH, München, Germany
(3) Chair for Robotics and Embedded Systems, Technische Universität München, Germany

Flexible and compliant real-time control of artificial limbs is a challenging endeavor for conventional control algorithms. Conversely, the neuro-control of biological limbs is usually highly stable despite the apparent complex nonlinearities and flexibilities. Part of this is caused by cerebellar fast learning of motor actions embedded in a complex sensory feedback system.

In an attempt to exploit this flexibility in a real-time robotic setting, we apply cerebellar control mechanisms to operate a compliant, anthropomimetic robotic arm built using the Myorobotics framework. The spiking neural network we use is simulated in real-time on a SpiNNaker computer. A custom interface translates between SpiNNaker’s digital synaptic spikes and the robot’s sensors and actuators on a CAN bus. We demonstrate initial results of the implementation with a one-degree-of-freedom robotic joint as proof of principle. Here, we benefit from the modular and extensible design of both the Myorobotics framework and the SpiNNaker platform.


Florian Walter
Institut für Informatik VI
Technische Universität München
Boltzmannstr. 3
85748 Garching bei München

4th Neurorobotics Performance Show

Date: 01.07.2015
Venue: Pisa, Italy

On July 1-2, 2015 the 4th Neurorobotics Performance Show took place in Pisa, Italy. This time the meeting was hosted by the Scuola Superiore Sant’Anna. The participants discussed organizational aspects as well as the implementation progress. You can download the complete schedule for both days here. Afterwards the SP10 Developer Workshop took place, where all software developers were able to discuss implementation issues in more detail. If you’re interested take a look at the agenda.

Here some impressions from the workshop: