ZAS ROBOTICSBUILD THE FUTURE

Robotics

Robot Odometry

Build kinematics, wheel odometry and sensor fusion on ZAS Car-3 through live bench experiments and a robot navigation capstone.

University / undergraduate
ZAS Car-3 robot with encoder motors and onboard sensors

Course Overview

A university-level, hands-on robotics course built around ZAS Car-3. Each module connects a concept to a demonstration on the same robot: measure wheel motion, observe odometry drift, calibrate inertial and magnetic sensors, tune filters and compare dead reckoning with a fused pose estimate. Six core bench sessions culminate in a waypoint-following race, with a bonus particle-filter localisation exercise.

Hands-on training on ZAS Car-3

Six core modules plus one bonus module. Each session includes a physical robot demonstration. Training dates and enrolment details will be confirmed separately.

Learning Outcome

  • Relate left and right wheel speeds to the linear and angular velocity of a differential-drive robot.
  • Estimate pose (x, y and heading) from encoder ticks and recognise accumulated dead-reckoning error.
  • Use MPU6050 gyro and accelerometer readings to explore heading, tilt and drift.
  • Calibrate QMC5883L compass readings and account for magnetic distortion and tilt.
  • Tune a complementary filter and combine encoder, IMU and compass data with Kalman / EKF pose estimation.
  • Apply PID waypoint following and compare a dead-reckoning robot with a sensor-fused robot.
  • Explore Monte Carlo localisation with ultrasonic and IR range measurements in the bonus module.

Industry Skills

  • Differential-drive kinematics and wheel-base calibration
  • Encoder-based pose estimation and trajectory logging
  • IMU and magnetometer calibration
  • Complementary filtering and Kalman / EKF sensor fusion
  • PID waypoint control and experimental validation
  • Particle-filter localisation using range sensors (bonus)

Industry Roles

Related role examples inferred from the skills covered; these are not job-placement claims.

  • Mobile robotics engineer
  • Embedded robotics developer
  • Robotics controls engineer
  • Sensor-fusion / localisation engineer

Detailed Syllabus

Each stage combines concepts with a hands-on demonstration on the robot.

01Differential-Drive Kinematics
  • Wheel speed and linear / angular velocity
  • Two-wheel-drive and caster chassis motion
  • Wheel-base calibration
Hands-on demonstration

Command a 1 m straight drive and a 90-degree pivot; compare measured and commanded motion and tune the wheel-base constant.

02Wheel Odometry
  • Encoder ticks to wheel distance
  • Dead-reckoning pose estimation: x, y and heading
  • Odometry trail visualisation on OLED / serial
Hands-on demonstration

Drive a taped 1 m square open-loop, plot the odometry trail and observe the failure to close.

03Inertial Sensing (MPU6050)
  • Gyro integration for heading
  • Accelerometer-based tilt
  • Gyro drift
Hands-on demonstration

Spin the robot in place for 60 seconds and observe gyro-only heading drift in real time.

04Absolute Heading (QMC5883L)
  • Magnetometer heading
  • Hard-iron and soft-iron distortion
  • Tilt compensation and calibration
Hands-on demonstration

Perform a live figure-eight calibration and compare raw and calibrated heading on the OLED.

05Sensor Fusion I - Complementary Filter
  • Fast but drifting gyro measurements
  • Slower, noisy compass measurements
  • Live tuning of the blend weight
Hands-on demonstration

Tune the blend weight through serial input and compare the fused trace with both raw inputs.

06Sensor Fusion II - Kalman / EKF Pose and Capstone Race
  • Fusion of encoder odometry, IMU and compass
  • A single robot pose estimate
  • PID waypoint following
Hands-on demonstration

Race a dead-reckoning-only robot against an EKF-fused robot on the same square course from the same start line.

07Particle Filter (Monte Carlo Localisation)Bonus
  • Ultrasonic and left / right IR sensors as a three-beam rangefinder
  • Correcting a particle cloud against a known map
  • Localisation from an unknown start
Hands-on demonstration

Place the robot at an unknown start in a taped rectangular pen and observe the particle cloud converge as it drives and takes range readings.

Inside the course

Explore the robot, course introduction and demonstrations below.

ZAS Car-3 robot with encoder motors and onboard sensors
ZAS Car-3: the hands-on platform used throughout the odometry and sensor-fusion course.

Training Hardware

  • ESP32-S3-Nano controller with WiFi / BLE telemetry
  • 2 x JGA25 DC370 encoder motors
  • TB6612FNG motor driver
  • MPU6050 IMU
  • QMC5883L compass
  • 0.96-inch OLED display
  • Onboard ultrasonic sensor and 2 x IR sensors for the bonus localisation module

What You Will Build

  • Tuned filter code
  • A logged trajectory per team
  • A capstone comparison of dead reckoning and EKF-fused waypoint following

Course Documents

Car-3 Robotics Journey - course brochurePDF · Opens in a new tab

KEEP EXPLORING

More in Robotics