2. Hardware and Assembly

The JetRacer ROS 2 platform is built atop the mechanical chassis provided by the Waveshare JetRacer AI Kit, but significantly upgraded with external telemetry hardware.

Bill of Materials (BOM)

Component	Purpose	Interface
NVIDIA Jetson Nano (4GB)	Core Brain. Handles CUDA accelerated YOLO inference and Nav2 spatial math.	Host
Waveshare JetRacer Pro Chassis	The Ackermann steering frame and DC motors.	Mechanical
RP2040 Microcontroller	Acts as the real-time embedded motor driver handling I2C signals.	I2C
Sony IMX219 (8MP)	The physical eye of the car tracking lines and objects.	CSI Ribbon Cable
RPLidar A1M8	360° laser scanner providing depth and maps up to 12 meters away.	USB (Serial)
MPU9250 AHRS IMU	9-DOF Absolute Orientation Sensor. Interfaced via RP2040 Serial Bridge.	Serial (via RP2040)
INA219 Monitor	Precise Voltage/Current sensing for power health.	I2C (via RP2040)
SSD1306 OLED	Real-time status display (IP, Battery, Faults).	I2C (Direct to Jetson)
Wheel Encoders	Hall-effect sensors determining raw physical forward momentum (Odometry).	GPIO -> RP2040
USB Microphone	Allows voice-activated routing commands.	USB
Standard Wireless Gamepad	Teleoperation, Deadman-switch, and Mission Control Sentinel.	Bluetooth / USB

Physical Architecture and Mounting

graph TD
    subgraph Compute
        NANO[Jetson Nano]
    end

    subgraph Sensory Array
        CSI(IMX219 Camera) -- Ribbon --> NANO
        LID(RPLidar A1) -- USB Serial --> NANO
        OLED(SSD1306 OLED) -- I2C --> NANO
        MIC(USB Microphone) -- USB --> NANO
    end

    subgraph Actuation & Embedded Sensing
        NANO -- USB Serial --> RP[Raspberry Pi RP2040 Driver]
        IMU(MPU9250 IMU) -- I2C --> RP
        INA(INA219 Power) -- I2C --> RP
        ENC(Wheel Encoders) -- GPIO --> RP
        RP -- PWM --> MOTORS[Rear DC Motors]
        RP -- PWM --> SERVO[Front Steering Servo]
    end

Important Hardware Quirks

[!WARNING]
The Encoder-IMU Divide: Because cheap IMUs drift horribly when mathematically double-integrating Acceleration into Linear Speed, our Sensor Fusion explicitly ignores the IMU's Linear Acceleration. Linear speed is derived EXCLUSIVELY from your Wheel Encoders. The IMU is mathematically isolated to only provide precise Yaw-Rate (Rotation).

[!CAUTION] Ackermann Kinematics: Because this is a car-like robot, the motors cannot turn the robot 'in place' like a vacuum cleaner. All Nav2 calculations must pass through our cmd_vel_to_steering.py inverse-kinematic converter, which measures the hardware wheelbase ($0.255m$) to calculate the precise angular curvature of the steering servo.

---

[!TIP] Next Step: Proceed to 03. Deployment and Docker to set up the software environment.