ABOUT
My name is Bohao Chu, graduated from Duisburg-Essen University of Germany with a Master's degree in Computer Engineering, and graduated from Zhengzhou University of China with a Bachelor's degree in Computer Science and Technology.
I have worked for a long time as a research assistant in the Internet of Things Research Institute at Zhengzhou University and HiWi in the department of Transport Systems and Logistics at Duisburg-Essen University.
The areas I have explored were mainly focused on Interactive Sensing and Embedded Applied Machine Learning, with involvement in WEB, PCB, FPGA, and Embedded Operating Systems (Linux, RTOS).
ABOUT
My name is Bohao Chu, graduated from Duisburg-Essen University of Germany with a Master's degree in Computer Engineering, and graduated from Zhengzhou University of China with a Bachelor's degree in Computer Science and Technology.
I have worked for a long time as a research assistant in the Internet of Things Research Institute at Zhengzhou University and HiWi in the department of Transport Systems and Logistics at Duisburg-Essen University.
The areas I have explored were mainly focused on Interactive Sensing and Embedded Applied Machine Learning, with involvement in WEB, PCB, FPGA, and Embedded Operating Systems (Linux, RTOS).
RESEARCH OVERVIEW
Since I entered university in 2016, I started my computer learning journey. Computer technology is always full of attraction for me, at first I was interested in any field, from Web technology (e.g., I built my first website and registered a domain name in 2016), to Microprocessor Programming and Embedded Systems, to Circuit Board Design, and later Machine Learning. In the process of studying and working, I have accumulated a lot of useful basic knowledge and met many like-minded people. Later, I tried to find my favorite field, which is Interactive Sensing and Embedded Machine Learning, which is of great significance to the realization of intelligent environment and intelligent life. Currently, I am looking for opportunities to do further research in this field. The picture below shows my research field and the corresponding projects (parts).
DOING
In the past six months, I have focused on the following two edge computing projects, which mainly involve embedded development and machine learning. These two projects are comprehensive projects for which we have designed dedicated boards and run embedded machine learning classification models on them, aiming to implement edge computing.
MACHINE LEARNING
Over the past few years, I have been involved in several machine learning projects. As shown in RESEARCH OVERVIEW, I divide these projects into generative models and categorical models. These models are implemented based on the Tensorflow or PyTorch framework, and more deep learning networks such as CNN, CVAE, LSTM, and MLP are applied. At the same time, I also tried quantization and pruning of the models to run these models on embedded devices.
Edge Computing Sensor Kit
General Activity Recognition Based on Integrated Multi-sensor Information Fusion
Multimodal Human Trajectory Prediction
Multimodal Human Trajectory Prediction Based on Collision-Free CVAE
Supply Chain Demand Forecast
Fusion ARIMA-LSTM model for Supply Chain demand forecasting based on Machine Learning
EMBEDDED SYSTEM
The embedded field was the first field I was involved in, when I was a rookie we always liked to do interesting things with sensors and microprocessors, but many things we forgot to record. Later these became the basis for our exploration of interactive sensing, and edge computing. At the same time, we are exploring deeper areas, such as embedded Linux, and various IoT communication technologies.
Arvato Implant Kit Based on RFID and LoRa
The medicine box is tracked through the RFID tag, and the data transmission is completed through LoRa.
Automatic Sorting Logistics Trolley
Logistics Sorting Robot System Based on Automatic Path Planning
Linux Gateway with LoRa and 4G
Embedded Linux Gateway for LoRaWAN Networking and Data Transfer over 4G
Water Service Machine
Voice and RFID controlled Water Service Machine
Voice Controlled Smart Wardrobe
Automatic Clothes Pickup and Storage Wardrobe Based on Speech Recognition Control
Bluetooth Low Energy Lock
Bluetooth Low Energy Lock based on nRF51822
NPU Based Vision Computing
General Activity Recognition Based on Integrated Multi-sensor Information Fusion
Indoor Haze Detector
Dual Purpose Indoor Haze Detection and Pen Holder
WEB SOFTWARE
In the IoT scenarios we design, we always want to be able to show our results in a beautiful way and at the same time be easily accessible to others, and WEB software can help us achieve that. For example, we realize the linkage between real world and virtual world through digital twin technology, and visualize the sensor data transmission in real time through data visualization, instead of an abstract process. At the same time, we designed an IoT data platform where our IoT devices can easily access our platform and transmit data for our future data analysis.
GPT SIER
Personal Chatbot based on GPT 3.5 Turbo Natural Language Large Model
Modular Data Analysis Platform
Modular data platform for self-service data analytics
Digital Twins
Digital Twin based on Sensor Data and Machine Learning
Personal Blog
Static Personal Website based on Bootstrap 5
Sensor Data Stream
Real-time Visualization of Sensor Data Streams
FPGA
It has always been our goal to run complex machine learning models on embedded devices, and we want to implement model inference as fast as possible, so we have tried various methods to speed up computation, such as model quantization, pruning, and using ASIC gas pedals. Based on this pursuit, we explored using FPGAs to customize our gas pedals, so we tried to port some open source projects, which are still in the exploration stage.
Customized Tiny WiFi
Exploring SDR Based Linux mac80211 Compatible Full Stack IEEE802.11/Wi-Fi Design
Open Small RISC-V
Exploring FPGA-based RISC-V core for three-level pipeline CPU