SLAM Car & Maze Test

In this lab, we worked with the MBot, a differential drive robot with 2 parallel wheels and a rear caster. Each motor is equipped with a magnetic wheel encoder. The robot also features a scanning 2D Lidar and a 3-axis IMU, and communicates via Wi-Fi with its onboard Jetson Nano compute module. Low-level motor control is implemented on the MBot Control Board using a RaspberryPi RP2040-based microcontroller.

Overview

Action

  • Cascade Control & PID
  • 3-DOF rigid-body coordinate transforms
  • Planar kinematics of a differential-drive ground robot
  • Motion models with uncertainty

Perception

  • Quadrature Encoders
  • MEMS Inertial Measurement Unit
  • 2D LIDAR Rangefinders
  • Camera and Fiducial Marker Detection

Reasoning

  • Monte Carlo Localization
  • Simultaneous Localization and Mapping
  • A* Search
  • Path Planning

Result

The car successfully navigated through a maze


Fig 1. Testing the Mbot during lab time


Fig 2. Hardware and Software Deployment


Fig 3. Result of the SLAM System