Frangipani
Top 20
On 28 January, Jörg Conradt, who leads the Neuro Computing Systems Lab at KTH Royal Institute of Technology in Stockholm, gave a presentation titled "Energy-smart Neuromorphic Sensing and Computation for Future Space Applications” at the AI for Space Applications Workshop that took place at KTH’s Digital Futures hub.
The workshop was held as part of the ASAP project:
View attachment 77025
View attachment 77021
While the first presentation slide didn’t look too exciting from a BRN shareholder’s perspective…
View attachment 77022
… we actually did get a mention later on, when Jörg Conradt touched upon a number of neuromorphic startups…
View attachment 77023
…although he didn’t get it quite right. Not only did he mix up the names of our company and its neuromorphic processor - he also misspelled and pronounced the latter as “Aikida”. And since when does BrainChip sound like a place name?
To be fair, though, he actually gave us an “out of this world” mention by pointing to what he mistook to be our company’s name and telling his audience (from 14:50 min):
“In fact, this company develops IP, and a very big Sweden-based space company has recently teamed up with Aikida [sic] to develop hardware - computing hardware - that they can send into space, that is potentially very robust against radiation and other effects.
So we have a common project starting on what to do with that hardware, and that’s my point of contact into space.”
It sounds as if Jörg Conradt’s lab at KTH will somehow be involved with whatever Frontgrade Gaisler has planned for Akida…
All the more reason he ought to familiarise himself with A & B ASAP.
Encouragingly, Akida made another appearance shortly after (this time with impeccable orthography) - if only for a split second, before Jörg Conradt moved on to the next slide. It was in the context of a research project titled Neuromorphic Edge Computing for Urban Traffic Monitoring in the city of Stockholm, funded by Digital Futures, a cross-disciplinary research centre established by KTH Royal Institute of Technology, Stockholm University and RISE Research Institutes of Sweden. Akida was listed alongside SpiNNaker 2 and Loihi 2 under “neuromorphic chips” and even had the honour of providing an exemplary image for that category.
View attachment 77027
View attachment 77030
View attachment 77031
Jörg Conradt didn’t specify whether or not the different neuromorphic processors were going to get benchmarked against each other, but I assume that’s the plan.
Here is some further information I found about said project that runs from January 2024 to December 2025 and “estimates a 100x reduction in power and a 20x reduction in installation cost.”
![]()
Edge computing for urban traffic monitoring — Digital Futures
Embedding low-power event cameras with low-power neuromorphic local computation provides a uniquely scalable solution for the growing need to monitor urban traffic and resource flows in real time...www.digitalfutures.kth.se
View attachment 77028
View attachment 77029
View attachment 77032
I wonder whether this urban traffic monitoring project with the help of event-based cameras was somehow inspired by the 2023 tinyML Pedestrian Detection Hackathon submission (utilising Akida) from Cristian Axenie’s SPICES lab at TH Nürnberg. It doesn’t seem that far fetched to make a connection - after all, they know each other well: Jörg Conradt was Cristian Axenie’s PhD supervisor at TUM (Technical University of Munich).
In the light of my recent post about Jörg Conradt (KTH Stockholm), his upcoming talk at the University of Cambridge (and online via Zoom) titled “Low-power embedded event-based vision processing for low-latency robotics” could be rather interesting…