From insect navigation to neuromorphic intelligence
Lead Research Organisation:
University of Edinburgh
Department Name: Sch of Informatics
Abstract
Italy
People |
ORCID iD |
Barbara Webb (Principal Investigator) |
Publications
Zhu L
(2023)
Neuromorphic sequence learning with an event camera on routes through vegetation.
in Science robotics
Description | This small award was for attendance at the CapoCaccia Workshop toward Neuromorphic Intelligence, originally for 2020, postponed due to Covid till 2022 (and shortened in attendance due to a Covid outbreak at the event). The main outcome was enhanced collaboration to establish an insect-inspired algorithm on a neuromorphic computing platform; there is a paper under review. It also seeded a new collaboration as described in the outcomes. |
Exploitation Route | Several representatives of companies developing neuromorphic hardware were at the workshop, including for Intel (developing Loihi). Strong interest was shown in the potential for taking additional algorithms I have developed through to implementation on this hardware. |
Sectors | Digital/Communication/Information Technologies (including Software),Electronics |
Title | Motion compensator for ant |
Description | We have developed a novel system that allows ant to walk freely and untethered on top of a large sphere which compensates for its motion (using fast camera tracking and feedback) to keep it in the same position. |
Type Of Material | Improvements to research infrastructure |
Year Produced | 2019 |
Provided To Others? | No |
Impact | This system can be used with 'virtual reality' feedback to the ant to greatly enhance the scope and throughput of behavioural experiments. |
Description | EPFL |
Organisation | Swiss Federal Institute of Technology in Lausanne (EPFL) |
Country | Switzerland |
Sector | Public |
PI Contribution | We introduce the target of understand, through detailed observations, how ants handle objects. |
Collaborator Contribution | My partner, Pavan Ramdya, has hosted my post-doctoral researcher for a week to gain detailed insight into methods for filming and tracking the movements of small insects in high resolution. He has offered ongoing advice as we develop our system. |
Impact | So far, this has contributed to development of a tracking system in Edinburgh. This collaboration has also resulted in plans to apply for HFSP funding. |
Start Year | 2022 |
Description | Neuromorphic hardware on a robot |
Organisation | Royal Institute of Technology |
Country | Sweden |
Sector | Academic/University |
PI Contribution | We have designed a spiking neural circuit that uses event camera input and learns to recognise familiar input sequences. We have collected data from a robot in outdoor scenarios and shown that the neural circuit can be implemented on SpiNNaker hardware. We are now building a closed-loop system so that the robot can react to the output of the circuit. |
Collaborator Contribution | Our partner has provided access and support to an interface system allowing the event camera to provide real-time input to the SpiNNaker hardware. |
Impact | Multidisciplinary: engineering and biology |
Start Year | 2022 |