INRC2

INRC2

AI embodiment over 3D simulation and neuromorphic hardware

INRC2

The simulated embodiment of AI control algorithms is key to further adoptance of neuromorphic computing in real world applications. The Neurorobotics Platform and Intel‘s Loihi research chip combine their strengths to enable it.

Project description

The Neurorobotics Platform (NRP) is a web-based software co-developed by fortiss in the frame of the Human Brain Project, that enables neuroscientists and AI experts to connect spiking neural networks with simulated bodies, to bridge the gap between AI algorithms and real-world applications.

Here we work on integrating Intel‘s research neuromorphic chip Loihi into the NRP over a well established spiking neural network simulator called Nengo and to use it as a benchmarking tool to compare the execution and learning of spiking neural networks over different simulators, including Nest, SpiNNaker, Nengo. The benchmarking experiment is featuring a quadruped mobile device controlled with a spiking Central Pattern Generator (CPG). This bases on the work done in completed project INRC1.

Research contribution

This project is both an engineering and research project. Software engineering is needed for the actual integration of Loihi into the NRP. Research is needed for the implementation of the spiking locomotion of a simulated quadruped device with spiking CPG.

The former is mostly a synchronisation effort with Intel on one side, that deliver the Loihi and its drivers and the Human Brain Project on the other side, delivering the Neurorobotics Platform (NRP). fortiss, formerly leading the development of the NRP, has a key expertise to implement the middleware to enable running spiking neural networks in Loihi over Nengo in realistic 3D simulated experiments.

The latter aspect of this project, namely the research in quadruped locomotion, is even more important because little has been done so far in spiking neural networks driven locomotion with bio-inspired approaches, namely Central Pattern Generators. We have succesfully shown that this is possible in simple embodiments, like snakes and their oscillatory locomotion. The focus of this project on quadrupeds takes this to another level of complexity, where four independant legs have to be synchronized to achieve a realistic movement for various gaits, in particular within fluids.

Project duration

Start: 01.11.2019

Intel and the Intel logo are trademarks of Intel Corporation in the U.S. and other countries.
 Emmanouil Angelidis

Your contact

Emmanouil Angelidis

+49 89 3603522 554
angelidis@fortiss.org

More information

Project partner

Intel Neuromorphic ComputingEPFL École Polytechnique Fédérale de Lausanne Biorobotics Lab