DriveSight

A mobile robot project with a camera for visual tracking, based on a Raspberry Pi Pico 2W and ESP32-S3.

info

Author: Neagu Andrei-Cristian
GitHub Project Link: https://github.com/UPB-PMRust-Students/project-mohgTheOmen.git

Description

DriveSight is a manually controlled robotic platform equipped with real-time video streaming. It integrates a camera module for live feed monitoring and a wireless controller interface for direct motor control, providing precise, remote navigation and situational awareness.

Motivation

I started this project to challenge myself and explore how far I could push the capabilities of embedded hardware, combining real-time control, wireless communication, and camera streaming on constrained systems.

Architecture

Main Components:

• Raspberry Pi Pico 2W (Motor Controller): Controls motors, ultrasonic sensors, and receives movement commands via UART.
• ESP32-S3: Captures a live camera feed and sends control commands from a connected Bluetooth controller.
• Raspberry Pi Pico 2W (Display Host): Hosts a Wi-Fi access point and receives data from the motor controller to display real-time sensor values or system status on a small screen.

Feedback Mechanisms:

• Current sensors (INA219): Monitor motor current to help maintain consistent speed across varying surface conditions.
• Ultrasonic sensors: Provide obstacle detection and distance feedback for navigation and collision avoidance.

User Interface:

• Web-based interface: Hosted on the ESP32-S3 for viewing the live camera feed.
• Bluetooth controller: Used for manual driving input.
• On-board display: Connected to Pico no.2, shows live telemetry from the motor controller over Wi-Fi.

Connections Between Components:

• UART: Used for communication between the ESP32-S3 and motor controller Pico.
• Wi-Fi: Motor controller Pico connects to the display host Pico (AP mode) for data transmission.
• I2C: Used by the motor controller Pico for reading data from INA219 current and ultrasonic sensors.
• GPIO: Used to control the L298N motor driver from the motor controller Pico.

Log

Week 21 - 27 April

• Tested components individually.

Week 28 April - 4 May

• Configured the ESP32-S3 to connect to a Wi-Fi network and stream the live camera feed. This involved setting up the esp_camera library for capturing frames and integrating it with the ESP-IDF networking stack.
• Configured the second Raspberry Pi Pico to display telemetry data received via TCP from the central Pico.

Week 5 - 11 May

• Tested UART communication between the ESP32-S3 and the motor controller Pico 2W to ensure reliable data exchange.
• Developed the motor control functionality to process commands received from the Bluetooth controller.

Week 12 - 18 May

• Integrated motor movement logic to dynamically adjust based on proximity feedback from the ultrasonic sensors.
• Began designing the chassis for the robot, focusing on structural stability and accommodating all hardware components.

Week 19 - 25 May

• Completed hardware assembly for the robot, including wiring and mounting all components.
• Developed and integrated software for the distance sensors, enabling them to function as parking sensors with real-time feedback on the remote display.
• Implemented motor control logic to respond to Bluetooth controller input, allowing for precise manual driving.
• Most core features are functional; only finishing touches and minor bug fixes remain.

Week 26 - 30 May

• Finalized integration tests for all hardware modules to ensure stable operation during the PM Fair.
• Verified live camera streaming, Bluetooth controller input, and telemetry display on the remote unit.
• Performed a full system demo run to check for any last-minute issues and confirmed all features are presentation-ready.
• Recorded the demo video and assembled all components into the 3D-printed housings.

Project Demo

Project Demo Video Link

Hardware

• Raspberry Pi Pico 2W: Main controller handling sensors and motor control.
• ESP32-S3: Provides camera feed for object tracking and sends control commands.
• Raspberry Pi Pico 2W (Display Unit): Hosts an access point and displays telemetry data.
• ST7735s: Display the telemetry data.
• L298N Motor Driver: Drives the two DC motors.
• 2× DC Motors: Provide movement for the robot.
• 2× INA219 Current Sensors: Monitor motor currents to detect stalls or resistance changes.
• 3× HC-SR04 Ultrasonic Sensors: Detect obstacles in front of the robot.
• Power Bank and Li-ion Batteries: Powers the system.
• Wires, Breadboard, and Connectors: For interconnecting all the components.

Pictures

Overview:

ESP32-S3:

AP Pico 2W:

Schematics

ESP32-S3:

Central Pico 2W:

AP Pico 2W:

Final Project

Main Unit

Remote

Bill of Materials

Device	Usage	Price
2x Raspberry Pi Pico 2W	Main control board	39.66 lei
ESP32-S3	Camera and Wi-Fi	56.71 lei
L298N Motor Driver	Controls motor power	10.99 lei
ST7735s Display	Displays sensor data	27.99 lei
2x JGA25-370 Motors	Robot movement	49.00 lei
2x INA219 Current Sensor	Current monitoring	29.99 lei
2x TLP281 Optocoupler	Signal isolation	8.49 lei
3x HC-SR04 Ultrasonic Sensor	Obstacle detection	14.99 lei
Li-Ion 3x Battery Holder	Power source	20.13 lei
Li-Ion 1x 18650 Battery Powerbank	Power source	19.93 lei
Li-Ion 2x 18650 Battery Powerbank	Power source	30.35 lei
Wires, Connectors, Breadboard, Resistors	Wiring	Already had them

Software

ESP32-S3 (C)

Library	Description	Usage
ESP-IDF	Official Espressif IoT Development Framework	Base SDK for Wi-Fi, networking, etc.
esp_camera	Camera driver for OV3660	Captures and streams MJPEG frames
Bluepad32	Bluetooth game controller library	Receives input from Bluetooth controller

Raspberry Pi Pico 2W (Rust)

Library	Description	Usage
embassy	Async embedded runtime for embedded systems	Task scheduling and async infrastructure
embassy-executor	Asynchronous task executor	Schedules and runs tasks in a no_std async environment
embassy-rp	RP2040-specific HAL for Embassy	GPIO, SPI, PWM, ADC, and other peripherals
embassy-time	Timer and delay handling	Non-blocking frame-rate control, delays, timeouts
embassy-sync	Synchronization primitives	Used for async channels and mutexes
embassy-gpio	GPIO abstraction for Embassy	Manages input/output pins for buttons, sensors, and motor control
embassy-futures	Futures and async utilities for Embassy	Enables combining and managing multiple async tasks and events
cyw43	Wi-Fi driver for the CYW43 chip	Connecting the Pico 2W to Wi-FI
cyw43-pio	PIO-based driver for the CYW43 chip	Enables Wi-Fi using the RP2350's PIO peripheral
embassy-net	TCP/IP networking stack	Hosts server or client for sending/receiving data
embedded-hal	Hardware abstraction layer for embedded systems	Standardized traits for peripherals (GPIO, I2C, SPI, etc.)
hcsr04-async	Async driver for ultrasonic distance sensor (if used)	Obstacle detection
defmt	Lightweight logging framework for embedded systems	Enables efficient, structured logging
defmt-rtt	RTT (Real-Time Transfer) backend for defmt	Outputs logs via RTT for real-time debugging
embedded-graphics	2D graphics library for embedded systems	Drawing text, shapes, and UI elements on the display
mipidsi	Display driver for MIPI-compatible SPI LCDs	Driving the ST7735s display via SPI
rand	Random number generation	Used internally by the TCP stack for port/sequence numbers
smoltcp	Embedded TCP/IP networking stack	Provides low-level TCP/IP networking for the Pico 2W
static-cell	Safe static memory cell abstraction	Manages static data for async tasks and peripherals
alloc-cortex-m	Heap allocator for Cortex-M microcontrollers	Enables dynamic memory allocation in embedded Rust
libm	Math library for embedded systems	Provides floating-point math functions in no_std environments
heapless	Fixed-capacity data structures without heap	Used for queues, buffers, and collections in constrained systems
panic-probe	Minimal panic handler for embedded Rust	Reports panics via RTT or serial for debugging

Links

1. Personal Project Repo