Virginia Tech
Design and SDR Prototyping of Online Real-Time Machine Learning for NextG MIMO and ISAC
Abstract
dc:description.abstractgeneralAs we move toward a more connected future, our wireless networks need to be faster and more reliable than ever. Machine Learning (ML), the technology behind AI, is the key to making this happen. However, using ML in real-time is difficult; not only do wireless signals change dramatically in mobile scenarios, but the configurations of the network interface can change in less than a millisecond. Using existing ML methods to process these changes in real-time is usually too complex or power-hungry. To solve this, we focus on a technique called Reservoir Computing. This is a lightweight type of ML that can learn using very little data, making it perfect for the fast-paced world of wireless signals. Think of it as a specialized, high-speed shortcut that allows devices to learn and adapt to their surroundings instantly. To ensure this technology works in the real world, we also developed two efficient new designs to run on small, low-power devices: the ButterflyNet and symmetric hybrid neuron processing architectures. To prove these concepts work, we didn't just run computer simulations; we built a specialized radio testbed and hardware prototypes. We successfully tested these in real-world conditions using both simulated signals and actual over-the-air transmissions. Our results show that these ``smart'' radios can handle complex tasks, such as filtering out interference and managing signals for moving devices, more effectively than today's standard technology. Ultimately, this work bridges the gap between theoretical AI and practical hardware, providing a blueprint for a future where our devices don't just transmit data, but actively learn and optimize themselves in real-time.
Degree
thesis:*- Name thesis:degree_name
- Doctor of Philosophy
- Level thesis:degree_level
- doctoral
- Discipline thesis:degree_discipline
- Electrical Engineering
- Department dc:contributor.department
- Electrical Engineering
- Grantor dc:publisher
- Virginia Tech
- Year dc:date.issued
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Liang, Yibin
- Chairs dc:contributor.committeechair
-
- Liu, Lingjia
- Yi, Yang
- Committee members dc:contributor.committeemember
-
- Midkiff, Scott F.
- Buehrer, R. Michael
- Ji, Bo
Subjects
dc:subject × 7Rights
dc:rights- Statement dc:rights
-
- In Copyright
- Licence dc:rights.uri
- Language dc:language.iso
- en
Identifiers
dc:identifier.*- Dc Identifier Other
- vt_gsexam:47098