SARA ALZAABI (ABU DHABI)
In an interview with Aletihad, Dr. Baker Mohammad, Professor and Chair of Computer & Communication Engineering, Group Lead of the System on Chip Lab, from Khalifa University, discussed his research on a self-powered memory device that mimics brain functions to advance smart electronics.
Drawing inspiration from and comparisons with other memory technologies, Dr. Mohammad said: "Imagine a memory device that powers itself and works a bit like the human brain! The brain is amazing at strengthening or weakening connections based on experience - a concept called 'synaptic plasticity'. This device mimics that behaviour, allowing it to learn and adapt like we do."
He explained that the device allows systems to function independently, using innovations like a self-rectifying memristor and piezoelectric nanogenerator (PENG) for real-time applications. It mimics the brain's ability to adjust connections based on activity.
Dr. Mohammad said: "This device adjusts its electrical properties in response to signals: positive voltage pulses make the connection stronger (potentiation), while negative pulses make it weaker (depression)."
The device also "forgets" weaker signals, similar to the brain, and uses oxygen vacancies in the zinc oxide (ZnO) layer to replicate synaptic behaviour. It even powers itself with a PENG, converting mechanical energy into electricity.
Dr. Mohammad noted the difficulties of integrating the PENG with the memory device while avoiding issues like sneak currents.
"We overcame this by designing a structure that mimics the unidirectional signal flow of biological synapses," he said.
The team also optimised the dynamic adaptability of the device and fine-tuned the PENG to match the memory device's requirements.
Dr. Mohammad stressed that the PENG is key to the device's self-powered functionality: The PENG converts mechanical energy from physical stimuli, like pressure or motion, into electricity, which powers the memory device."
This self-powered design enables autonomy, eliminating the need for external power, making it ideal for wearables and biomedical devices.
Dr. Mohammad explained: "The PENG converts energy from movements like walking or heartbeats into electricity, powering the device without batteries."
He also discussed how the self-powered memory device improves energy efficiency and real-time data processing by integrating neuromorphic resistive memory with a piezoelectric nanogenerator (PENG).
He said: "This technology represents a breakthrough in energy efficiency and real-time data processing by integrating neuromorphic resistive memory with a piezoelectric nanogenerator (PENG). By harvesting energy from mechanical stimuli, mimicking brain-like processing functions, and incorporating a unidirectional memory structure, the device establishes a new standard for energy-efficient, responsive computing."
Noting the benefits of the technology, Dr. Mohammad said: "The self-powered memory device holds transformative potential across several industries and fields."
These include biomedical devices, consumer electronics, aerospace and defense, where its ability to operate autonomously and process data locally offers significant advantages, such as reducing the need for battery replacements, enabling real-time feedback, and improving energy management in remote or energy-constrained environments.
Dr. Mohammad shared insights on the next steps for advancing the self-powered memory device and expanding its applications, emphasising the importance of refining the device's design, scaling production and integrating it into practical systems.
He mentioned expanding the device's use to precision farming and smart cities, stressing the need for reliability, sustainability, and industry-academia partnerships to drive development.
Regarding AI and adaptive systems, he said: "The self-powered neuromorphic memory device, integrating piezoelectric nanogenerators (PENGs), can significantly impact AI and adaptive systems.
"It overcomes traditional computing bottlenecks by combining memory and processing, enabling efficient real-time data processing and reducing energy consumption."
The technology's adaptability and energy independence suits robotics, healthcare, and space exploration.
Dr. Mohammad further stressed the need for partnerships across sectors to scale and integrate the device.