Pinned Repositories
Coursework_ELEN6889_large_data_stream_processing
Course assigment of ELEN6889: large data stream processing
CSEE4119_CustomedTcpOverUdp
This program implements a TCP-like protocol on top of UDP. It supports variable window size, and can recover from data loss, data corruption, and data delay. Basically, I implement the TCP 20 bytes header as follow: [ Source port ][ Dest port ] [ Sequence Number ] [ Acknowledgement Number ] [ Flags ][Receive window] [ Checksum ][ Urgent ] [ Payload ... ] I run my sender, receiver and the link emulator on one machine. The link emulator tool we will use is called newudpl. The link emulator acts like a "proxy", i.e., the sender is configured to send packets there and the proxy is configured to send packets to the receiver and it can drop, corrupt, reorder and delay packets. The receiver process sends its acknowledgements directly to the data sender. I use timeout mechanism for retransmission as per TCP (without fast retransmission), such as RTT estimation and update of timeout interval, to handle packet loss and use checksum in UDP/TCP to check if the packet is corrupted, and use ack number & sequence number to check if the packet is delivered in order or not. I use a FIN request to signal the end of the transmission. For supporting the variable window size, I implement my own mechanism in this program.
EECS4764_IoT_EntranceGuard
A safeguard door access control unit with mask recognition, temperature monitoring, and hand sanitizing models.
ELEN6690_PhishingURLDetection
Phishing and malicious URL detection using different ML methods.
Mitigating-High-Costs-via-Risk-Aware-Value-Oriented-Net-Demand-Forecasting
The paper ''Mitigating High Costs via Risk-Aware Value-Oriented Net Demand Forecasting'' submitted to NAPS 2024
RL-IoT
Leveraging reinforcement learning to recover the semantics of protocol messages and to take control of the device to reach a given goal, while minimizing the number of interactions. Assume we know only a database of possible IoT protocol messages, whose semantics are however unknown.
hjj1998's Repositories
hjj1998/CSEE4119_CustomedTcpOverUdp
This program implements a TCP-like protocol on top of UDP. It supports variable window size, and can recover from data loss, data corruption, and data delay. Basically, I implement the TCP 20 bytes header as follow: [ Source port ][ Dest port ] [ Sequence Number ] [ Acknowledgement Number ] [ Flags ][Receive window] [ Checksum ][ Urgent ] [ Payload ... ] I run my sender, receiver and the link emulator on one machine. The link emulator tool we will use is called newudpl. The link emulator acts like a "proxy", i.e., the sender is configured to send packets there and the proxy is configured to send packets to the receiver and it can drop, corrupt, reorder and delay packets. The receiver process sends its acknowledgements directly to the data sender. I use timeout mechanism for retransmission as per TCP (without fast retransmission), such as RTT estimation and update of timeout interval, to handle packet loss and use checksum in UDP/TCP to check if the packet is corrupted, and use ack number & sequence number to check if the packet is delivered in order or not. I use a FIN request to signal the end of the transmission. For supporting the variable window size, I implement my own mechanism in this program.
hjj1998/Coursework_ELEN6889_large_data_stream_processing
Course assigment of ELEN6889: large data stream processing
hjj1998/EECS4764_IoT_EntranceGuard
A safeguard door access control unit with mask recognition, temperature monitoring, and hand sanitizing models.
hjj1998/ELEN6690_PhishingURLDetection
Phishing and malicious URL detection using different ML methods.
hjj1998/Mitigating-High-Costs-via-Risk-Aware-Value-Oriented-Net-Demand-Forecasting
The paper ''Mitigating High Costs via Risk-Aware Value-Oriented Net Demand Forecasting'' submitted to NAPS 2024
hjj1998/RL-IoT
Leveraging reinforcement learning to recover the semantics of protocol messages and to take control of the device to reach a given goal, while minimizing the number of interactions. Assume we know only a database of possible IoT protocol messages, whose semantics are however unknown.