Tanvir Singh — Portfolio

Custom 360° LiDAR device setup

Example LiDAR scan output

3D LiDAR Scanner

Custom ToF + stepper platform • C/C++ · Python

Built a 360° LiDAR device using a VL53L1X ToF sensor and 28BYJ-48 stepper motor. Programmed motion control, data capture, and UART serial streaming in C/C++. Processed scans and rendered 2D/3D views in Python to visualize room geometry and obstacles.

1. Device Overview

(a) Features

Enclosed single-package device (cardboard), approx. 30 cm × 15 cm.
Main capability: scans a 3D area to output a spatial reconstruction.
One push button (breadboard) to start/finish; on-board MCU button to reset.
ToF sensor VL53L1X, full 360° with 16 measurements per revolution; range up to 400 cm, up to 50 Hz; I²C + UART.
Generates navigable 3D models (Python/Open3D).
Open source: firmware in C/C++ (Keil), host in Python (Jupyter).
MCU: MSP432E401Y @ 80 MHz (from 120 MHz); 3–5 V; 1024 KB Flash, 256 KB SRAM; UART ~7.5 Mbps.
Approx. system cost (without MCU): $60 CAD.

(b) General Description

Integrated embedded system: ToF sensor provides distances; stepper performs 360° sweep; Python renders 3D model.
Digital I/O: push-button start, LED activity, measurement logging; coordinates derived (YZ/XYZ) from ToF.
Data path: transduction → conditioning/ADC → MCU → UART (byte-wise) to PC → Python/Open3D visualisation; host runs in polling mode and records to file.

Figure 1 — LiDAR system block / data flow diagram

Figure 2 — Device characteristics table

2. Detailed Description

(a) Distance Measurement

ToF sensor: 3415-POLOLU VL53L1X; one emitter + one receiver measure time-of-flight → distance.
Wiring: VIN=3.3 V, GND=0 V; SDA/SCL → PB2/PB3 for I²C; vendor API used for setup, ranging, and data access.
Stepper: MOT-28BYJ48, IN1..IN4 → PH0..PH3; powered 5 V/GND. Full-step sequence drives rotation.
Motion cadence: 250 ms per state; 16 samples per 360° ⇒ 22.5°/sample. Data sent over COM6.
Host visualisation: Python Open3D builds point cloud, connects neighbours/sets, adds per-set bounding boxes; fully rotatable 3D view.
Firmware: PLL 80 MHz; I²C (sensor), UART (PC), GPIO (stepper+button), polling every 10 ms.
Loop: read distance → store → rotate 22.5° → UART transmit → repeat for 16 samples.

3. Circuit Schematic

Figure 3 — LiDAR circuit schematic

4. Programming Logic Flowcharts

Figure 4 — Programming logic flowchart 1

Figure 5 — Programming logic flowchart 2

Screenshot of the Quizler educational web app

Quizler App

DeltaHacks project • React · HTML/CSS

Built a collaborative education app at DeltaHacks focused on ease of use. Designed a clean React UI that lets students join a shared room and quiz each other in real time.

Hardware architecture for upsampling & colourspace conversion

RGB to YUV colourspace coversion & horizontal downsampling

Hardware Image Decompressor

Altera DE2 FPGA • Verilog · VGA

Implemented a YUV image decompressor in Verilog and displayed frames in real-time over VGA. Designed memory buffers and a streaming datapath on the DE2 board, ensuring reliable timing with no visual artifacts.

1. Introduction

Implemented McMaster Image Compression Rev. 17 on an Altera DE2 for 320×240 images.
Pipeline: RGB → YUV, DCT, quantization, SRAM, UART, VGA, FSM-based control.

2. Implementation Details

(a) Upsampling and Colour Space Conversion

Figure 1 — Custom state table for flow/timing

Figure 2 — Utilised registers & signals table

Timing: common-case loop = 12 cycles per half; 240 rows.
Multiplier roles: M1/M2 for U,V interpolation; M3/M4 for colour conversion.
Accumulators: RGB ↔ YUV matrix multiply.
Throughput: ~978 cycles/row, 234,963 total.
Utilization: ~85% multipliers.

(b) Resource Usage and Critical Path

Resources: 2,435 / 114,480 LEs (~2%).
Optimization: centralize constants/operands; drive repeats from control.
Critical path: 14.787 ns due to combinational depth and placement.

3. Conclusion

4-week build combining FSM, VGA/SRAM/UART, and DSP blocks.
Emphasis on task division, communication, systematic debugging.
Robust, scalable base for future features.

Pacemaker UART processing block diagram

Rate/Duty cycle conversion block

Pacing modes block and outputs

Python Tkinter GUI for pacemaker

Custom Pacemaker Prototype

Embedded control • Simulink · Python (Tkinter)

Prototyped a pacemaker system that regulates patient heart rate via bi-directional telemetry. Modeled control logic in Simulink and built a real-time Python GUI to monitor heart-rate data and adjust pacing modes.

Hardware setup for blink-controlled car

Blink-Controlled Car

Signal processing • Arduino · Python

Engineered a small vehicle that responds to users’ blink gestures. Refined EEG signal processing to extract blink patterns and translated them into motor commands for responsive navigation.

Projects

3D LiDAR Scanner

Quizler App

Hardware Image Decompressor

Custom Pacemaker Prototype

Blink-Controlled Car