This project presents advancements in utilizing artificial intelligence (AI) for precise control of robotic systems in surgical procedures, focusing on close-loop control strategies. Close-loop control, employing data-driven feedback mechanisms such as deep reinforcement learning (DRL) and imitation learning (IL), offers adaptability to unstructured environments but requires significant data for training, posing challenges like data scarcity and privacy concerns. We address these challenges by proposing end-to-end training strategies for motion planning, mimicking human behavior through DRL for autonomous exploration and interaction with the environment. A key focus lies in navigating flexible endoscopes within the gastrointestinal tract, vital for diagnostic and therapeutic procedures. By modeling deformations using finite element methods (FEM) and training control policies through on-policy DRL, we enable autonomous navigation in dynamic soft tissue environments. Our contributions include employing model-free DRL for controlling tendon-driven flexible robots in contact scenarios and integrating intuitive hand gesture recognition for intraoperative intervention, enhancing safety and efficiency. This research underscores the potential of AI-driven control systems in revolutionizing robotic surgery, promising precise and adaptable robotic interventions with improved patient outcomes.
Category: ypec-2024
Project Background The College Entrance Examination poses significant challenges for young students. Aiming to provide assistance and guidance, we build our website (www.gaokao985.online), holding the motto “Assist every tiny dream”. Introduction to Functions Our website offers seven key functions. Users can filter universities and majors by attributes like province, university types, major types, etc. to find ideal choices. Besides, the major recommendation interface helps users select majors aligned with their future career goals. Additionally, the mark analysis tool aids in evaluating subject scores, helping users draw up a study plan. Users can share messages on our message board to express opinions and support peers. Furthermore, our real-time chat room allows users to exchange opinions and consult experts. Technical Details Firstly, we used HTML and CSS for website structure and design. Besides, we adopted JavaScript to handle form submissions and facilitate data transfer to the server. What’s more, we employed AJAX to achieve asynchronous data exchange on web pages, which allows web pages to exchange and update data with the server without reloading the entire page. We also utilized PHP and SQL to connect to the server and perform database operations. Session management helps tracked user login state and relevant messages.
Categories
UG20 – Swipe and Travel
Swipe and Travel revolutionizes group travel planning with AI-powered, personalized recommendations. Current travel planning is frustrating. The internet is filled with generic information, and group consensus is time-consuming. Therefore, we aim to enhance this process by leveraging AI to analyze the user’s personality and provide a platform that optimizes group itinerary planning by balancing everyone’s preferences. The core technology is a sophisticated AI and LLM model that uncovers the relationship between personality and travel preferences. Through quizzes and big data analysis, we construct a comprehensive profile of users. Coupled with extensive datasets on worldwide travel spots and their defining attributes, our personality-based recommendation system offers tailored suggestions by matching user profiles with attraction features. Through optimization algorithms, we create itineraries that strike a balance between the preferences of all group members while mapping out the simplest route. Swipe and Travel is also intuitive and fun. User-friendly interface allows travelers to effortlessly swipe through options, with easy editing functions to tailor their plans. Creating a vibrant community, users will be inspired to share discovered hidden spots, while continually expanding our database. Swipe and Travel has a competitive edge in personalization and collaboration that revolutionizes the way people explore the world together.
Smart homes are the embodiment of the Internet of Things (IoT) in our daily lives. With the continuous progress of AI technology, the prospects for the development of smart homes are vast. Future smart homes will become more intelligent and personalized, better fulfilling users’ diverse needs. Six newcomers in the field of microelectronics, inspired by their intuitive understanding of home life, have decided to seize this opportunity to create a smart home system that reflects their vision. This project serves as their initial exploration of IoT, while also applying skills in chip design, communication, and other technologies to empower artificial intelligence. The project selects an STM32 series microcontroller as the system’s core and integrates environmental sensing devices like temperature sensors, humidity sensors, and light sensors onto a single system board through PCB design. This allows for real-time monitoring of temperature, humidity, light, and intensity, enabling functions such as automatic alarms and automatic light switching. Additionally, the project incorporates 2.4G communication technology to transmit real-time data to the cloud, enabling remote control and data viewing of the smart home system through a WeChat mini-program designed by us independently.
Our Solartronic Autonomous Mini (SAM) Robot is an autonomous weight-loading robot programmed to navigate intelligently, avoiding obstacles and following human targets. This Arduino-based robot can utilize a solar energy collection system by integrating a photovoltaic panel into the surface. It can operate in an environmentally friendly and sustainable manner with a carrying capacity of up to 10 kg to assist in various tasks. SAM Robot is designed to intelligently and independently explore its surroundings, dodging obstacles and tracking human targets. This Arduino-based robot is an eco-friendly technical solution since it can be fully powered by the collected solar energy. It uses a variety of sensors and sophisticated algorithms to identify objects in front of it while reducing its dependency on conventional power sources. The SAM Robot’s intelligence is notable for its capacity to identify obstacles and follow human targets. Its control algorithm is designed to monitor the robot’s trajectory and detect any impending obstructions, whether in motion or stationary. Based on this feature, the SAM Robot is a very useful tool for a range of applications, including the transport of small objects as well as the handling of heavy or bulky loads.
As the traffic environment becomes increasingly complex, visually impaired individuals face numerous challenges when traveling, such as: inability to recognize traffic lights; struggling to pass when tactile paving is occupied; having trouble to use visually-based interactive devices like smartphones; difficulty in timely obstacle avoidance; inability to determine their current location in unfamiliar environments; and difficulty in seeking help during emergencies. Therefore, developing an intelligent travel assistance system based on voice interaction holds significant practical importance. This design aims to create a smart travel assistance system based on voice interaction with the following technical features: 1) Sidewalk detection using computer vision algorithms to detect deviations from the centerline, alerting with a buzzer; 2) Obstacle and traffic signal recognition using deep learning models and radar; 3) Intelligent voice interaction using AI speech recognition and NLP models; 4) GPS positioning to broadcast current location and nearby points of interest; 5) Fall detection using a gyroscope to measure angular velocity and acceleration, sending an emergency alert when necessary. This design not only provides visually impaired individuals with low-cost, high-safety travel assistance services but also enhances their ability and confidence to travel independently, improves their travel experience, and increases their enjoyment of traveling.
Newly constructed building units require inspection and maintenance. As the number of new units is increasingly large due to urban expansions, and the varying designs also exists to meet the increasingly sophisticated demand from the market, this raises the need and workload of building inspection. The building management industry needs to adapt to the diverse inspection work. Currently, property management relies mainly on experienced inspectors for training the newcomers, but this traditional approach has a limited capacity and leads to a shortage of manpower, unable to meet the demands of property maintenance. Our team plans to use Mixed Reality (MR) smart glasses technology and collaborate with the property management industry to provide high-quality and efficient training courses for newcomers to the industry. The courses provide comprehensive and intuitive, interactive demonstrations of various tasks for the trainee with the guidance of the content from the veterans. The courses can also be customized to adapt to the progress of the trainee, as well as the needs of the company and on the buildings assigned. The team also aim to create an MR-powered inspection process, to achieve digitization, standardization, and intelligent building inspection and verification processes, increasing inspection throughput without sacrificing quality.
Smart Library Management System is an innovative project that integrates advanced technology and AI elements to revolutionize library seat management. The system aims to optimize seat resource utilization, prevent seat hogging, and enhance space efficiency in libraries. Through the use of IoT platforms and intelligent real-time management, the project effectively detects seat occupancy, provides real-time seat availability situations to students, and alerts staff to promptly handle any issues. Additionally, the system incorporates features such as student ID verification for personal belongings, convenient online seat availability checks, and efficient staff management. The project also showcases the use of 3D modeling and printing, as well as the utilization of powerful development boards like STM32 and ESP8266. By automating seat monitoring and preventing seat hogging, the system contributes to a more efficient and equitable use of library facilities, enhancing the overall experience for students. The integration of cutting-edge technologies, such as pressure sensors, NFC card readers, and vibration motors, showcases the potential of IoT and AI-driven solutions to revolutionize the way libraries manage their physical spaces and serve their communities.
Categories
UG17 – Dokodemo Call
An AR assistance platform revolutionising remote guidance. Imagine your family member struggling with a new laundry machine. With Dokodemo Call, you seamlessly connect via video call and remotely insert AR objects into their environment, offering step-by-step instructions. Through their phone screen, you witness their surroundings and place virtual guides directly within their world in real-time. This innovative approach transcends traditional remote support, providing unparalleled clarity and interactivity. Dokodemo Call’s intuitive interface ensures effortless navigation for users of all ages, eliminating technical barriers. Whether troubleshooting household appliances or guiding industrial processes, Dokodemo Call empowers users to conquer challenges with precision and efficiency. As we envision widespread adoption, Dokodemo Call stands poised to transform how we connect and assist remotely, bridging distances with seamless AR integration.
Generative AI has the potential to revolutionize various industries by generating text, data, or other forms of media based on user-provided prompts. However, ensuring data security and privacy remains a challenge, especially when dealing with sensitive personal information. This project explores the combination of prompt engineering and cryptography as a solution to optimize and personalize prompts while maintaining user privacy. We propose an intermediary prompt engineering application, which can be embedded as a browser extension or a chatbot, that stores encrypted personal details and processes prompt engineering. This method allows users to grant our product temporary access to their data, and optimize the prompts for the underlying AI model, ensuring a personalized and secure experience. Using our product, individuals or businesses who are not proficient in prompt engineering, yet concerned about user privacy, can enjoy seamless, continuous, and personalized services employing generative AI. However, challenges and limitations exist, including the potential information leak to generative AI in the stage of prompt engineering and final submission, as well as similarity to existing context options in terms of user experience. Future research should focus on prompt engineering at the local level or utilizing homomorphic encryption of data to further enhance data security and privacy.