Robomaster ICRA AI Challenge is a competition held by DJI and IEEE International Conference on Robotics and Automation. It emphasizes on writing AI algorithms and tuning and choosing and training machine learning models to power pure automatic robots to accomplish complex goals like building maps, avoiding collision and tracking. Our team aims to build AI Robot tuples that can work purely automatically. Implementing and design decision policies according to competition requirements. Training AI networks to recognize and track objects. Implement localization and motion planning on robots under uncertainty.
Robomaster is a national robotics competition for university students, hosted by DJI. The project is about forming a team to design various types of smart robots which can engage in face-to-face, videogame-style battles. The preparation process involves the knowledge in the following disciplines: mechanical and hardware design, control theory application, computer vision and smart algorithm, technical management. The ultimate purpose of the project is to develop the skills, quality and team spirit of engineering students that can be beneficial in their future career.
AI-Driving is an interdisciplinary endeavor in experiential learning, in which students apply technical knowledges in A.I. and robotics. In this project, students will learn knowledge about computer vision, artificial intelligence and robotics and implement AI algorithms to work on auto-driving vehicles running in a real miniature environment. The objective of the project is to allow students to gain insights about the development framework for autonomous vehicles, and to attempt problem such as lane following and object recognition. Students will also form a team representing HKU to participate in International competitions such as AI-Driving Olympics held by the DuckieTown Foundation and other contests related to AI-Driving, e.g., AWS DeepRacer.
This project aims to develop the set of technologies to achieve convenient-to-use mobility support for daily use of the elderly. In this project, we are developing a smart elderly walker which is intended to play an active role in an elderly person’s daily life, with three fundamental functionalities that do not exist or not well supported by (smart) walkers in the market: smart walking assistance; falling prevention and support; autonomous mobility.
A set of mechanical, control, sensory, and AI technologies is being developed including:
(1) novel walker mechanical structure with omnidirectional mobility and outrigger mechanisms;
(2) dual-mode actuation and control for walking/standing support and fall prevention/ recovery;
(3) multimodal sensory data collection through soft sensory skin, and data processing on device and in the cloud, for event detection and control such as user front following and fall detection;
(4) sound-source localization for elderly localisation and auto-navigation of walker.
In this project, students will learn AI and robot related technical disciplines (such as machine vision, embedded system design, mechanical control, inertial navigation, human-computer interaction, etc.) through designing and building intelligent robots according to the rules of the RoboMaster Robotics Competition. Different types of robots are required to cooperate with each other and work together to attack the base of the opponent and at the same time protect their own base.
Robomaster is a national robotic competition for university students, hosted by DJI during the summer vacation on a yearly basis. It combines technology and e-gaming competition style, thus being exciting and unconventional. HKU Robomaster has been recruiting enthusiastic engineering students since 2017. We aim to cultivate the quality of members that will benefit their career with the practical experience gained during the preparation and participation process. Our work is divided into technical management, machinal & hardware management, and software & algorithm development.
Conventionally, information suffers from exponential decay in long distance transmission. However, in exciting projects like building the quantum internet, an information carrier has to pass through many intermediate quantum servers before reaching the receiver. Such a goal cannot be achieved without a scheme of reliably transmitting data over a long sequence of noisy channels. Chiribella et al. formulated the second level of quantization of quantum information theory by considering the superposition of quantum trajectories and quantum channels and demonstrated communication advantage of this model. We apply the model further and show the theoretical feasibility of communication through asymptotically many noisy channels.
Microbially Induced Calcite Precipitation (MICP) is an emerging bio-geotechnical soil improvement technique to enhance the engineering performance of soil. Application of biotechnology has been suggested as an environmental-friendly alternative to conventional approaches. MICP involves using bacteria to hydrolyze urea into carbonate ions through bacterial metabolism. The carbonate ions combine with calcium ions to produce calcite crystal precipitates, which fill the soil pores and bind the soil grains together.
In this project, the effects of several influencing factors on the engineering properties of the MICP treated silica sand are investigated. Permeability and unconfined compressive strength tests will be carried out to characterize the engineering properties of the treated sand. Precipitation patterns of calcite crystals within the sand after MICP treatment will be investigated using Scanning Electron Microscopy (SEM).