Neuromorphic & Secure Hardware (NSH)

Page content

The Neuromorphic & Secure Hardware (NSH) research team develops brain-inspired computing technologies that reproduce key principles of biological neural systems in efficient, scalable and resilient hardware and software.

Our research combines computational neuroscience, artificial intelligence, electronic and computer engineering to create new computing architectures inspired by the way biological neural systems process information. We investigate how principles such as sparse connectivity, event-driven computation, neural plasticity, parallel processing and self-repair can be translated into practical computing systems.

Research

The team investigates the development of neural architectures and computational systems inspired by the structure and function of biological brains.

A particular focus is the development of large-scale neural systems capable of emulating biological sensory and cognitive functions, including visual, auditory and haptic processing. These systems are designed for implementation on reconfigurable and specialised computing platforms, providing opportunities for high-performance, low-power and adaptive computation.

Our research includes:

Neuromorphic computing architectures
Spiking neural networks
Brain-inspired neural architectures
Neural and neural-glial computation
Neural plasticity and learning
Event-driven and sparse computation
Reconfigurable and FPGA-based neural systems
Neuromorphic hardware acceleration
Hardware/software co-design
Bio-inspired sensory processing
Self-repairing and resilient computing
Neural systems for robotics
AI acceleration and energy-efficient computing

Spiking Neural Networks

Spiking neural networks (SNNs) provide an important foundation for our neuromorphic research. We investigate neural models, network topologies, learning mechanisms and encoding strategies that capture important characteristics of biological information processing.

Our research explores how spiking architectures can be efficiently implemented on computational hardware and how event-driven processing can reduce unnecessary computation while retaining the adaptive characteristics of neural systems.

Neuromorphic Hardware

We develop and investigate hardware platforms that can accelerate or emulate neural computation.

Research includes the implementation of large-scale neural architectures on reconfigurable hardware, with particular interest in exploiting parallelism, event-based processing and efficient arithmetic approaches. These architectures provide a route towards computing systems capable of processing sensory information in real time while reducing computational and energy requirements.

The team also investigates specialised hardware for modelling neural and neural-glial interactions. This includes the development of astro-centric computing platforms, such as the EMBRACE platform, which provides a hardware environment for investigating complex interactions between neurons and glial cells.

Brain-Inspired Sensory Processing

Neuromorphic systems can provide efficient approaches to processing complex sensory information.

Our research investigates neural architectures for emulating biological sensory capabilities such as vision, sound and haptics. Particular emphasis is placed on visual processing and the development of architectures capable of extracting and processing meaningful information from high-dimensional sensory streams.

These technologies can support applications in robotics, intelligent sensing, medical technologies and industrial process monitoring.

Learning and Adaptation

Biological neural systems are highly adaptive. We investigate learning algorithms and computational mechanisms that enable neuromorphic systems to adapt to changing environments and workloads.

This includes research into neural plasticity, learning rules, network topologies, information encoding and efficient training approaches. The objective is to develop systems that can learn from data and respond dynamically without relying solely on conventional, computationally intensive processing architectures.

Resilient and Self-Repairing Computing

A further research direction is the development of resilient neural systems inspired by the biological capacity for adaptation and recovery.

We investigate methods for incorporating self-repair and fault tolerance into neural implementations, particularly on FPGA and other reconfigurable platforms. Such approaches aim to enable large-scale intelligent systems to continue operating in the presence of hardware faults or degradation.

Applications

Neuromorphic Engineering research is aimed at applications where conventional computing architectures can be limited by energy consumption, latency, scalability or adaptability.

Potential applications include:

Intelligent robotics and autonomous systems
Sensory fusion
Computer vision
Medical signal and image processing
Industrial process control
Edge AI and low-power intelligent systems
Real-time pattern recognition
Adaptive and resilient computing
Brain-inspired artificial intelligence

Through collaboration with neuroscience and AI researchers, the team translates principles discovered in biological neural systems into new computational architectures and technologies.

Research Facilities and Platforms

The team works across software and hardware research environments for developing, modelling and evaluating neuromorphic systems. This includes high-performance computing resources, reconfigurable computing platforms and specialist neural modelling and hardware systems.

The team also collaborates closely with the NeuroAI and Neurotechnology (NAINT) research team, enabling a continuous pathway from neuroscience and neural data through computational models to neuromorphic implementation.

Research Vision

Our long-term vision is to develop a new generation of intelligent computing systems that are more efficient, adaptive, scalable and resilient by drawing directly from the principles of biological neural computation.

By integrating neuroscience with computer and electronic engineering, the Neuromorphic Engineering team seeks to move beyond conventional computing architectures towards systems that can sense, learn, adapt and respond in ways inspired by biological intelligence.

Team Members

The Neuromorphic Engineering team brings together expertise in computational neuroscience, neural engineering, neuromorphic computing, artificial intelligence, electronic engineering, reconfigurable computing and hardware/software co-design.