GUANGZHOU UNIVERSITY

Their core work

Guangzhou University contributes specialized expertise in bio-inspired visual computing — specifically, engineering neural systems that mimic how insects detect and respond to approaching objects. Their researchers model the visual pathways of flying insects (particularly motion-sensitive and looming-sensitive neurons) to build collision detection algorithms suitable for real-time, embedded hardware. In H2020 projects, they participated as third-party knowledge exchange partners under MSCA-RISE, sending and receiving researchers to collaborate with European teams on both the theoretical modelling and VLSI hardware realization of these systems. Their work sits at the intersection of computational neuroscience, computer vision, and embedded systems design.

What they specialise in

How they've shifted over time

In the early period (STEP2DYNA, 2016–2021), the focus was foundational: building and validating bio-inspired neural models of spatial-temporal processing for collision detection in general dynamic environments. The emphasis was on modelling and realization — getting the biology-to-algorithm translation right. By the later project (ULTRACEPT, 2018–2024), the work matured into a specific application domain — vehicle collision avoidance — and expanded to multi-modal perception including thermal imaging, alongside VLSI hardware implementation. The trend is a clear move from theoretical neural modelling toward deployable, hardware-realized systems for autonomous transport safety.

Guangzhou University is moving from academic neural modelling toward applied autonomous vehicle safety systems — making them increasingly relevant to industry partners in automotive, robotics, and embedded AI.

How they like to work

Guangzhou University has participated exclusively as a third party in MSCA-RISE staff exchange projects, meaning their engagement model is researcher mobility and knowledge transfer rather than formal project leadership or funded participation. They join established European research consortia as external knowledge nodes — contributing specialized expertise through visiting researchers rather than through deliverables or budgets. This suggests they are best approached as a deep-expertise collaborator for specific technical problems, particularly from European teams seeking access to their neural computing group, rather than as a consortium coordinator.

Guangzhou University has connected with 19 unique consortium partners across 6 countries through just two MSCA-RISE projects, indicating relatively dense networks per project. Their partners are primarily European academic and research institutions, with their Chinese base giving them a distinctive non-European perspective within these EU-funded consortia.

What sets them apart

Guangzhou University brings a rare combination: deep knowledge of insect neuroscience translated directly into engineering-ready collision detection algorithms, validated in both simulation and VLSI hardware. Few European groups combine this biological fidelity with embedded systems realization at the same depth. For a consortium working on autonomous systems or robotics that needs biologically plausible, computationally lightweight object detection, they fill a niche that standard computer vision groups do not cover.

Highlights from their portfolio

• ULTRACEPT
  A six-year project (2018–2024) that applied multi-modal bio-inspired neural perception — including thermal imaging and VLSI implementation — directly to the vehicle collision avoidance problem, marking the group's most applied and industrially relevant work.
• STEP2DYNA
  The foundational project establishing the spatial-temporal neural processing framework for collision detection, running from 2016 to 2021 and laying the theoretical groundwork that ULTRACEPT then industrialized.