Shiva Rajagopal

I created a real-time system that can use an NTSC video signal to simulate the lighting effects of a TV in a low-power fashion. This enables such applications as real-time ambient TV lighting, as well as simulating a TV's light in an unoccupied home to give the appearance of an occupied home. For this to be believable, the system must be able to keep up with the flickering and update speed of a TV signal, and therefore requires a dedicated, real-time processor. The goal of this project was to… Show more

I created a real-time system that can use an NTSC video signal to simulate the lighting effects of a TV in a low-power fashion. This enables such applications as real-time ambient TV lighting, as well as simulating a TV's light in an unoccupied home to give the appearance of an occupied home. For this to be believable, the system must be able to keep up with the flickering and update speed of a TV signal, and therefore requires a dedicated, real-time processor. The goal of this project was to create a general purpose algorithm that can accomplish this task, and use PWM outputs to drive a larger circuit that controls an RGB LED for verification. The user can send in any NTSC video signal and see the immediate changes on the LED as the video's dominant color changes. Show less

Deep Packet Inspection is a critical component in Network Intrusion Detection Systems, but it is a problem that grows in complexity very quickly as internet infrastructure improves and bandwidth increases. The problem is further exacerbated by the frequent emergence of new threats on the internet. This project explores using High Level Synthesis (HLS) to speed up the development time of new DPI systems and implementing new rules, using two common string matching algorithms, Aho-Corasick and… Show more

Deep Packet Inspection is a critical component in Network Intrusion Detection Systems, but it is a problem that grows in complexity very quickly as internet infrastructure improves and bandwidth increases. The problem is further exacerbated by the frequent emergence of new threats on the internet. This project explores using High Level Synthesis (HLS) to speed up the development time of new DPI systems and implementing new rules, using two common string matching algorithms, Aho-Corasick and Boyer-Moore. The results of our experiments show that using HLS to create DPI systems is feasible, but still requires heavy optimization and specialized hardware to make the system as efficient as it needs to be. Additionally, using the Synopsis ASIC toolflow, we found that using Aho-Corasick along with off-chip RAM is the preferred choice for string matching in terms of cycle time and required area on the chip. The implementation of the state machines for both algorithms is also critical for creating an efficient implementation, as HLS cannot optimize this very well.
(Message me for GitHub access) Show less

System's objective is to have a fast, reliable measurement of biological tissue impedance. Contributions include designing a new data acquisition component for the larger system, including use of an external analog-to-digital converter (ADC) and a PIC32 microcontroller to make an open-source, smaller, and less expensive component for this part of the system. Additionally, consultation and support for the electronic components of the system was provided to the rest of the team throughout the… Show more

System's objective is to have a fast, reliable measurement of biological tissue impedance. Contributions include designing a new data acquisition component for the larger system, including use of an external analog-to-digital converter (ADC) and a PIC32 microcontroller to make an open-source, smaller, and less expensive component for this part of the system. Additionally, consultation and support for the electronic components of the system was provided to the rest of the team throughout the year. This project involves applying electrical and computer engineering to biological concepts. Show less

The objective of this project was use Embedded Linux on a Raspberry Pi to replace the old, proprietary, expensive logging device in a bird tracking system. The system included UART connections combined with a Python backbone. Accomplishments of the project include:
- Decreased price of base station without decreasing quality and staying within system's form factor.
- Integrating internet database connection into system, enabling logging from remote locations.
- Embedded Linux system… Show more

The objective of this project was use Embedded Linux on a Raspberry Pi to replace the old, proprietary, expensive logging device in a bird tracking system. The system included UART connections combined with a Python backbone. Accomplishments of the project include:
- Decreased price of base station without decreasing quality and staying within system's form factor.
- Integrating internet database connection into system, enabling logging from remote locations.
- Embedded Linux system resilience. Radio scanner and GPS could be disconnected during system operation without loss of operation, and internet could be disconnected at any point without losing data that would be sent to the internet database. Show less

Worked with two other students to implement a fully embedded limited-vocabulary voice recognition system on a Xilinx Zedboard.
- Signal processing component implemented in C++ and compressed input data by a factor of 20.
- Neural network implemented in C++ using 3 layers and offline training to save space on FPGA.
- Signal processing and neural network code modified to be hardware-friendly and synthesized to HDL using Xilinx Vivado HLS. SoC microphone interface implemented in C using… Show more

Worked with two other students to implement a fully embedded limited-vocabulary voice recognition system on a Xilinx Zedboard.
- Signal processing component implemented in C++ and compressed input data by a factor of 20.
- Neural network implemented in C++ using 3 layers and offline training to save space on FPGA.
- Signal processing and neural network code modified to be hardware-friendly and synthesized to HDL using Xilinx Vivado HLS. SoC microphone interface implemented in C using Xilinx Vivado and SDK to properly connect SoC hardware.
- System achieved 90% accuracy and sub-second processing time, and used less than 50% of FPGA resources. Show less

Worked with a partner to envision, design, build, and test our own product using an ATMega1284P Microcontroller and analog hardware. Implemented system using:
-CAD-designed and 3D-printed enclosure for sensors
-Analog circuitry to process and clean signals
-Microcontroller/Embedded programming to output sound at appropriate times
See web page for more detail

Worked with two other students in various roles including architect, verification lead, and design lead. Implemented quad-core multicore processor in Verilog RTL composed of three subsystems:
-Fully-bypassed 5-stage pipelined processor with variable latency multiplier
-Two-way set-associative blocking cache with finite-state-machine controller
-Ring interconnection network with adaptive routing algorithm
Verified each subsystem and final design using extensive unit and integration… Show more

Worked with two other students in various roles including architect, verification lead, and design lead. Implemented quad-core multicore processor in Verilog RTL composed of three subsystems:
-Fully-bypassed 5-stage pipelined processor with variable latency multiplier
-Two-way set-associative blocking cache with finite-state-machine controller
-Ring interconnection network with adaptive routing algorithm
Verified each subsystem and final design using extensive unit and integration testing. Evaluated each subsystem and final design using various micro-benchmarks. Show less