Research Topic | Research Area | Details |
Data-Driven Decision Making | AI/ML, NLP, Network Monitoring | Many of the challenges in open ecosystems typically come in the area of integration of systems and solutions. Being able to quickly identify and fault sectionalize issues will require new tools. How can we use natural language processing of code and OpenAI to generate solutions (integration, monitoring, troubleshooting, visualizing, etc) and enable easier adoption for partners. Purpose: Using OpenAI to create NLP driven contextual system integration framework. Primarily intended as an aid to system integrators, to assist with software delivery into pipelines and producing automated quality assessments driven by SRE principles. |
Efficient, low latency, and pipeline FFT architecture for hardware implementation | DSP, Algorithms, wireless communications, FPGA, Hardware | Pipelined FFT with low resource usage/design complexity/latency is a key technology for wireless communications. One widely used architecture is the Radix-22 SDF (single-path delay feedback), "A New Approach to Pipeline FFT Processor", He and Torkelson, 1996. This is used in Ericsson's product portfolio. One issue with Radix-22 SDF architecture is that DIF (decimation in frequency) version has input in natural order and out in bit-reversed order while DIT (decimation in time) version has input in bit-reversed order and output in natural order. It's highly desirable to explore alternate FFT architectures that are similar to Radix-22 SDF in resources usage/design complexity/latency, but with both input and output in natural order. Purpose: Pipelined FFT architecture with both input and output in natural order, suitable for efficient FPGA implementation. |
Applicable Semantic Communications for Metaverse | Communications, Metaverse, resource allocation, algorithms, wireless systems, sensors | The operation of Metaverse heavily depends on the data collection and processing that accurately reflects human movements and changes in surroundings. Semantic communications become an essential component for Metaverse, as there is a growing need to extract the semantics from the raw data collected from the end devices such as head movement, arm swings, gestures, and speeches. With the semantics extraction, the end devices can interpret and filter out the information concerned by the Metaverse server and only transmit the needed information to save communication resources and reduce computation latency at the Metaverse server. Meanwhile, at the Metaverse server side, semantic information from videos can be extracted based on the user’s preference with irrelevant details ignored to reduce downlink pressure. Existing evaluation tools are designed to study wireless communications transmitting information bits over a fading and noisy communication channel. There is a lack of essential analytical and evaluation tools to study semantic communications in the context of Metaverse. This project aims at developing such an analytical framework to perform analysis and optimization of semantic communications for Metaverse.
Purpose: Developing sematic communications architecture for Metaverse, analysis and optimization of semantic-aware data generation and resource allocation for sematic communications, developing evaluation framework for sematic-aware algorithms. |
Multi-task and Meta Reinforcement Learning for Network Management | ML (RL), algorithms, wireless communication | This project aims at developing reinforcement learning (RL)methods that are suited for deployment in wireless networks. The main challenge that the work will overcome is the inability of traditional RL method to generalize. New methods will be explored to develop robust RL techniques that can improve the convergence and stability of RL under two practical situations: 1) Network Deployments and/or Network Conditions. How can previously learned RL policies quickly adapt to different deployments and changes in the wireless network environment? 2) Operator Preferences. How can previously learned RL policies quickly adapt to operators re-configuring their network and algorithm preferences. Purpose : Develop a robust RL solution to control a network function of interest to Ericsson. Techniques from multi-task and meta RL will be investigated to achieve this. |
Integrated Sensing and Communication Systems | Sensors, Algorithms, Resource allocation, modeling, Communication Systems architecture, Wireless Transmitter/Receiver architecture | Future networks are envisaged to have a sense of the physical world through measuring and imaging the surrounding environment to enable advanced location-aware services, such as drone communications, vehicular communications, environmental monitoring, indoor positioning services and human motion recognition. In principle, a radio emission can simultaneously deliver information from the transmitter to the receiver and extract environmental knowledge based on the scattered echoes. Hence, it is possible to design integrated sensing and communication (ISC) functionalities integrated into one system, sharing the same resources in time, frequency and space domains. Existing evaluation tools are designed to study wireless systems dedicated for communications. There is a lack of some essential analytical and evaluation tools to study ISC in the context of 5G NR. This project aims at developing such analytical framework to perform analysis and optimization of ISC. Purpose : Developing an analytical and evaluation framework for ISC, analysis and optimization of resource allocation for ISC, developing information theoretical models for ISC, designing novel beamformers, waveforms and receiver architectures. |