REU 2021 Has Been Postponed Until 2022 – Learn More Here
Please Review the REUs and Select Your Choice Below for an REU to Apply For
Machine Learning and Control Algorithm design for Connected Autonomous Vehicles
Primary PI: Fei Miao
Vehicle-to-vehicle (V2V) and Vehicle-to-Infrastructure (V2I) wireless connectivity is the next frontier in road transportation, which will greatly benefit the safety and reliability of autonomous cars. Information shared among autonomous vehicles provides opportunities to better coordination schemes and also raises novel challenges. In the future, connected autonomous vehicles (CAVs) equipped with both self-driving technology and V2V connectivity, will lead to vastly improved road safety and efficient traffic flow. However, existing frameworks for connected vehicles or platoons have not tackle with the high-level reasoning about under what conditions coordination among vehicles can be built. Only when neighbors are trustworthy and drive well, sharing information and coordinate with other vehicles will improve safety of the connected vehicles and efficiency of the traffic flow. Hence, it is critical for future connected autonomous vehicles systems to operate based on hierarchical decision-making theories, techniques and coordination protocols with other vehicles under complicated environments. This project aims to build theories and implement experiments for a hierarchical decision-making process of autonomous vehicles under the connected vehicles environment. The benefits of sharing different types of information is analyzed at different scales, from system level efficiency to safety guarantee of each individual autonomous vehicle.
Mobile Sensing, Learning and Data Mining for Smart Cities
Primary PI: Suining He
Due to facilitating urbanization and exploding urban population, urban computing has attracted much attention across industry and academia. Urban computing involves the computing applications of networks, sensors, internet of things, computational power, data and interdisciplinary knowledge domains to improve the living quality of densely populated areas. Topics in Dr. He’s research project include designing and implementing general framework of urban computing, vehicle computing applications, smart cities, vehicle sensing and data collection, urban transportation data management for applications like ride/bike sharing, vehicle data mining and machine/deep learning techniques for urban data processing, spatial-temporal data analytics, and knowledge domain fusion. Implementation backgrounds with deep learning, Android/iOS application development or spatial-temporal data mining is preferred.
Efficient deep neural networks for image recognition
Primary PI: Caiwen Ding
Machine learning has been experiencing a phenomenal resurgence thanks to the big data and the significant advances in processing speeds. Deep learning or deep neural networks (DNNs) has been able to deliver impressive results in many challenging problems such as visual and recognition tasks, machine translations, and drug discovery. Despite the advantage of improved overall accuracy, the deep-layered structure and large model sizes increase the computational complexity and memory requirements. To achieve higher scalability, performance, and energy efficiency, in our lab, we build efficient machine learning & deep neural network systems using advanced algorithms and optimization techniques. Students can learn many hands-on coding techniques such as Tensor-flow, PyTorch, Python, as well as embedded systems such as Jetson TX2, Jetson Nano, Cell phones, FPGAs
Development of a Configurable Real-time High-speed Wireless Communication Platform for Industrial Internet of Things (IIoT) Systems.
Primary PI: Song Han
As a major component of the U.S. economy, the process industries in recent years are fully embracing the concept of Industry 4.0 and renovating their manufacturing plants towards smart factories, where intelligent sensing and field devices are pervasively deployed and wirelessly connected in the field to provide real-time and reliable process monitoring, diagnosis and control. This paves the way for better understanding of the manufacturing process, thereby enabling efficient and sustainable production. During this paradigm shift, the underlying wireless communication fabric plays a key role in interconnecting those intelligent devices to perform safety- and mission-critical sensing and control tasks. Most of the existing industrial wireless technologies, however can only support a narrow class of applications. They neither can provide deterministic nor high-speed communication for mechanical and robotic control, and thus cannot serve as an ideal network fabric for advanced process control systems. To address this technical gap, this project aims to base on the research team’s previous work on RT-WiFi protocol design to develop a full-blown RT-WiFi communication system. The proposed work is essential for establishing the viability of the RT-WiFi technology in the real-world market place. Whereas our previous work has laid the foundation for the technology, real-world deployment requires both extensibility to multi-clusters and better assurance on reliability. We will address these issues in this project to transform a demonstration of concept into an implementation that can meet the demands of the real-world industrial environment. In this project, the research team will partner with Emerson Process Automation, a leading company in the process control industry, to pursue the proposed research and development tasks. This partnership will bring the research team insights on the communication fabric design needs from the customers perspectives and provide a unique real-life industrial facility to deploy, evaluate and demonstrate the RT-WiFi technologies.
Apply for REU