Morteza Hashemi

My research interests span the broadly defined areas of communication systems, real-time data networking, control, learning, and cyber-physical systems. Specifically, my research focuses on improving the performance of communications and computing systems through innovative techniques in data transfer and learning. In this path, I strive to work on emerg­ing interdisciplinary topics such as the application of machine learning in communication problems. Differentiating theme of my research is based on exploiting side information and out-of-band resources across various problems and system applications including: (1) en­abling low latency and high energy efficiency in millimeter wave systems such as 5G and Internet-of-Things (IoT) (2) optimizing the complex interplay among encoding, decoding, and feedback information in coded communication and computation systems, while most work in the area typically optimize one or two of these parameters simultaneously, (3) pro­viding reliable and robust wireless communications for Machine-to-Machine (M2M) and ve­hicular networks, (4) applying learning techniques to optimize policies within cooperative communications settings. The significance of my research contributions has been demon­strated in 1 book chapter, 6 accepted and published journal papers (1 additional journal papers are under review), and 8 accepted and published proceedings papers in peer-reviewed conferences (2 additional conference papers under submission). The following sections pro­vide further details of my research, organized by the research topic across the network stack, in a bottom-up fashion: • Physical layer: Novel mmWave communications architectures (2.1); feedback-based network coding schemes (2.2); • Network layer: Robust data networking within vehicular networks (2.3); • Application layer: Applications of online learning models in data networking: investi­gating efficient beam alignment in mm Wave systems 2 Research Projects 2.1 Millimeter Wave Communications with Low Latency and High Energy Efficiency: In order to address the problem of spectrum scarcity that cellular providers are already experiencing, mm Wave is likely to be one of the key defining characteristics of 5G systems and IoT applications. However, due to variable channels, intermittent connectivity, high delay and energy usage, it is most likely that mm Wave systems will have a sub-6 GHz companion interface. To fill this gap in the previous work, this research thrust is aimed to dramatically enhance the performance of standalone mm Wave systems by fully integrating the mm Wave and sub-6 GHz technologies. Our proposed architecture leverages the sub-6 GHz band to assist in beamforming and data transfer. In particular, we design a proactive queue control and scheduling policy to fully exploit the abundant while intermittent mm Wave capacity. In addition, a joint resource allocation framework is obtained to maximize the energy efficiency and achieve a high data rate. This research will take an important step towards the development of mobile mm Wave technology, which will significantly alleviate the enormous spectrum crunch that we are encountering right now due to the proliferation of new applications and services on smart devices. 2.2 Graph-Based Network Coding with Feedback: The ability to transfer data reliably over an unreliable service is intrinsic to a number of emerging technologies such as digital video broadcasting and cloud computing. Classical coding schemes such as the Rateless codes can often provide reliable communications for sufficiently long block lengths. However, one key requirement for coded communications and computation is to achieve a fast decoding rate. This feature is vital in a wide spectrum of applications, including streaming applications and distributed computing across delay or fault prone processors. For this purpose, my work on Rateless codes leverages a parsimonious use of the feedback channel to significantly enhance the real-time decoding of data. The developed coding scheme focuses on dynamically adjusting the encoder operation based on a low overhead use of the feedback channel. Feedback information is obtained based on the structural inference of the decoding graph. In addition to reliability and real-time decoding, my work exploits Rateless codes to achieve some rudimentary level of security and privacy almost for free. 2.3 Reliable Wireless Communications in Vehicular Environments: The goal of this project is to establish a reliable and robust wireless communications inside vehicles, mitigating the issues with physical wires such as cost, weight and maintenance. The outcomes of this research pave the path toward a cheaper, more reliable, and "greener" transportation system. In particular, modern vehicles incorporate tens of sensors to provide vital sensor information to electronic control units (ECUs). In the current architecture, vehicle sensors are connected to ECUs via physical wires, which increase the cost, weight and maintenance effort of the car, especially as the number of electronic components keeps increasing. To address this issue, wireless networks have been contemplated for replacing the current wires with wireless links. However, the ability to reliably aggregate data in one or several processing centers is critical to the monitoring capabilities of the sensors, which are typically constrained in both energy and computational power. For intra-vehicle networks, this aggregation is further complicated by the highly dynamic channel properties that vehicles may experience as they travel through areas with different radio interference patterns or road quality that can physically perturb the sensors. To date, several single-hop communication models based on Zigbee, RFID, and ultra-wideband technologies, have been examined for intra-car wireless networking. In contrast to the previous works, we show that multi-hop networking provides clear benefits within cars: it reduces radio energy consumption while simultaneously improves the reliability performance from %80 (single-hop) to %95 (multi-hop) for all sensor nodes distributed across the vehicle. 2.4 Applications of Bandit Models in Data Networking: The goal of this research thrust is to exploit bandit models in data networking problem. Specifically, we investigate the problem of beam alignment in mm Wave systems, and design an optimal algorithm to reduce the overhead associated with the classical beam alignment schemes. Specifically, due to directional communications in mm Wave systems, the transmitter and receiver beams need to be aligned that will incur a high delay overhead since without a priori knowledge of the transmitter/receiver location, the search space spans the entire angular domain. This is fur­ther exacerbated under dynamic conditions (e.g., moving vehicles) where the access to the base station (access point) is highly dynamic with intermittent on-off periods, requiring more frequent beam alignment and signal training. To mitigate this issue, we consider an online stochastic optimization formulation where the goal is to maximize the directivity gain (i.e., received energy) of the beam alignment policy within a time period. We exploit the inher­ent correlation and unimodality properties of the model, and demonstrate that contextual information improves the performance. To this end, we propose an equivalent structured Multi-Armed Bandit model to optimally exploit the exploration-exploitation tradeoff. In contrast to the classical MAB models, the contextual information makes the lower bound on regret (i.e., performance loss compared with an oracle policy) independent of the number of beams. This is a crucial property since the number of all combinations of beam patterns can be large in transceiver antenna arrays, especially in massive MIMO systems. 3. Conclusion In summation, the application of communications, control, networking, and learning tech­niques in emerging technologies, when studied from the theory and system points of view, provides timely and well-motivated research problems. The intersection of these research areas provides substantial exploration opportunities, and my research interests sit squarely at their nexus. As demonstrated by my past and present projects, I have the ability to for­mulate original research ideas, perform theoretical and empirical research to support these ideas, publish my results, and initiate research collaboration with researchers. I look forward to continuing and expanding these activities as a future faculty.