Quanskill Apply
NEWRobotics Systems · Level 2 (Advanced)

Build one autonomous robot. Upgrade it every week.

A 12-week advanced ROS 2 programme in robotics software engineering. Take a mobile manipulator from its first launch to a complete warehouse mission, prove every capability with benchmark data, and finish with hands-on sim-to-real deployment.

Apply for the November 2026 batch Explore the curriculum
Duration
12 weeks
Core stack
ROS 2 Jazzy
First batch
Late Nov 2026
Career track
Paid internship*
capstone_mission · warehouse worldsimulation
The Week 12 capstone: the robot localises at the dock, navigates the warehouse, avoids a moving person, picks an item from a shelf and delivers it to the drop zone. SHELF B-04 RACK A RACK C DROP ZONE DOCK
LOCATE
NAVIGATE
AVOID
PICK
DELIVER
ROS 2 JazzyGazebo HarmonicNav2MoveIt 2slam_toolboxros2_controlOpenCVPyTorchONNX Runtimerobot_localizationDockerRViz2Isaac Sim
12weeks of hands-on robotics engineering
12capability upgrades on one robot
7layers of the autonomy stack, sense to act
3mopaid internship pathway (conditions apply)
The course robot: a differential-drive mobile base with a 6-DOF arm, 2D LiDAR, camera and IMU. CameraRGB · detection 6-DOF armMoveIt 2 · grasping 2D LiDARSLAM · costmaps IMU· EKF fusion Diff-drive baseros2_control · odometry

Meet your robot

One mobile manipulator, built up over 12 weeks

You don't start from scratch each week. You build one complete robot software stack on a simulated mobile manipulator in a Gazebo warehouse, and every module makes it more capable.

Platform
Mobile base with a 6-DOF robotic arm
Sensors
2D LiDAR, cameras and IMU
World
Warehouse environment in Gazebo Harmonic
Week 12 mission
Locate, navigate, avoid, pick and deliver

The autonomy stack

Seven questions every autonomous robot must answer

  1. 01

    Sense

    What are my sensors reporting?

    W1 · W3
  2. 02

    Estimate

    Where am I?

    W5 · W7
  3. 03

    Perceive

    What is around me?

    W6
  4. 04

    Predict

    Where will it move next?

    W9
  5. 05

    Plan

    What should I do?

    W8 · W10
  6. 06

    Control

    How do I do it accurately?

    W2 · W4
  7. 07

    Act

    Move, pick and deliver.

    W10 – W12

How you learn

Engineered the way real robotics teams work

The method repeats every week. The robot's capability does not: each week adds something new to the same machine.

One robot, 12 upgrades

Every module adds one production capability to the same evolving robot.

Problem first

Each week opens with a real robot problem. Theory, maths and tools are introduced to solve it.

Failure-first debugging

Systems break on purpose: sensor noise, bad frames, blocked paths, poor tuning. You diagnose and fix them.

Benchmark-gated progress

A capability is accepted only with evidence: error, latency, stability, accuracy or success rate.

Build Break Measure Integrate EVERY WEEK

A typical week

01Concept

Problem, maths and system thinking

02Guided lab

Build the new capability

03Failure lab

Break it deliberately and diagnose it

04Benchmark

Measure output quality

05Integration

Connect it to the evolving robot

06Review

Demo, explain and defend design choices

Curriculum

12 weeks. 12 capability upgrades.

From ROS 2 fundamentals to SLAM, Nav2 autonomous navigation, MoveIt 2 manipulation and Physical AI. Select a week to see what you build and how it is measured.

Before Week 1 · Setup & readinessUbuntu 24.04, ROS 2 Jazzy, Gazebo Harmonic, Git and VS Code, plus short refreshers so everyone starts ready.

Robot capability: 1 / 12 upgrades

Week 01 · Foundations

Robot runtime

Nodes, topics, services, actions, parameters and launch files.

Build
Bring up the course robot with a one-command launch.
Measure & test
Topic rate, latency and runtime health; test bad QoS and wrong frames.
Outcome
Robot launches correctly and you can inspect and use every ROS 2 interface.

Week 02 · Foundations

Kinematics & math model

Transforms, quaternions, forward and inverse kinematics, Jacobians and base kinematics.

Build
Write arm and base kinematics and visualise them in RViz2.
Measure & test
Singular poses, joint limits and unreachable targets; record IK error.
Outcome
A working kinematics library with a short benchmark report.

Week 03 · Foundations

Digital twin

URDF/xacro, links, joints, collisions, sensors and Gazebo world setup.

Build
Model the robot and its warehouse world with LiDAR, camera and IMU.
Measure & test
TF tree, sensor rates and real-time factor; test bad frames and inertia.
Outcome
Robot model launches cleanly and all core sensors publish.

Week 04 · Control & sensing

Controlled motion

PID, trajectory generation and closed-loop base and arm control with ros2_control.

Build
Tune motion controllers for stable movement.
Measure & test
Overshoot, settling time and tracking error; test saturation and delay.
Outcome
Controllers are tuned and repeatable.

Week 05 · Control & sensing

State estimation

Wheel odometry, IMU, covariance, Extended Kalman Filter and robot_localization.

Build
Fuse odometry and IMU into one usable pose estimate.
Measure & test
Drift and RMSE; test wheel slip and IMU bias.
Outcome
Fused state beats raw odometry, with its limits explained.

Week 06 · Control & sensing

Perception

Camera models, calibration, OpenCV, pretrained object detection and object localisation.

Build
Detect target objects and publish detections in the robot frame.
Measure & test
Precision, recall, FPS and latency; test blur, lighting and clutter.
Outcome
A perception node with usable detections and a benchmark report.

Week 07 · Autonomy

SLAM

Occupancy grids, slam_toolbox, map saving and localisation on a known map.

Build
Create a map and relocalise the robot in it.
Measure & test
Map quality and localisation uncertainty; test corridors and drift.
Outcome
A saved map with repeatable localisation.

Week 08 · Autonomy

Autonomous navigation

Nav2 costmaps, planners, controllers, recovery behaviours and waypoint missions.

Build
Send the robot safely to goals across the map.
Measure & test
Success rate, path length and time; test blocked paths and narrow spaces.
Outcome
Robot passes the defined navigation evaluation.

Week 09 · Autonomy

Tracking & prediction

Multi-object tracking, data association and short-horizon motion prediction.

Build
Track moving people and objects and publish their near-future motion.
Measure & test
Tracking and prediction error; test occlusion and ID switches.
Outcome
A stable tracker with a reported evaluation.

Week 10 · Manipulation & deployment

Manipulation

MoveIt 2, collision scenes, grasp pose generation and a pick-and-place pipeline.

Build
Pick a detected object and place it at a target location.
Measure & test
Grasp success and planning time; test pose error and clutter.
Outcome
Repeated pick-and-place trials succeed and failures are documented.

Week 11 · Manipulation & deployment

Physical AI

Robot demonstrations, dataset preparation and one imitation-learning skill.

Build
Record demonstrations, train a small policy and compare it with a classical baseline.
Measure & test
Policy success and failure categories under distribution shift.
Outcome
Dataset, trained policy and comparison report.

Week 12 · Manipulation & deployment

Integration, deployment & sim-to-real

Mission orchestration, testing, rosbag2, Docker packaging and hardware interfaces.

Build
Run the full locate → navigate → predict → pick → deliver capstone and connect it to the physical deployment lab.
Measure & test
Mission success and end-to-end latency; test restarts and sensor faults.
Outcome
Capstone demo, repository, report and deployment checklist.

Core topics are built, run, tested and explained hands-on. Advanced tools such as NVIDIA Isaac Sim and Isaac Lab are covered as guided overviews.

Sim-to-real

From simulation to a real robot

The capstone runs in simulation. A separate guided deployment lab shows how ROS 2 commands a low-cost physical robot through a clear hardware stack, ending in a station-based warehouse delivery demo.

Data path: a delivery request goes to the ROS 2 mission manager on a laptop, through a serial bridge to an Arduino Uno, which drives the sensors, motor driver and 2WD robot. Status feedback returns the other way. ROS 2 laptopmission managerdelivery requests Serial bridgeROS 2 nodecommand protocol Arduino UnofirmwareIR · ultrasonic 2WD robotL298N motor driverline-following base → commands: GOTO HOME · A · B · C ← status: station reached · moving · stopped · obstacle
  1. Verify the hardwareChassis, motors, sensors, power and wiring pass the checklist.
  2. Test motors and sensorsForward, stop and turn commands with live sensor data.
  3. Line following with obstacle stopThe robot follows the track and stops safely for obstacles.
  4. Define the message protocolCommand and status messages, visible on the serial monitor.
  5. Build the ROS 2 serial bridgeROS 2 can now command the robot.
  6. Run delivery missionsHOME, A, B and C station deliveries succeed.

Tools & technology

The industry-standard robotics stack

The same open-source tools used by robotics teams in production, in Python and modern C++.

ROS 2 core

ROS 2 Jazzyrclcpprclpytf2colconlaunch

Simulation

Gazebo HarmonicIsaac Sim / Isaac Lab (overview)

Languages & OS

Python 3C++17Ubuntu 24.04 LTS

Model & control

URDFxacroros2_control

Estimation

robot_localizationEKF

Perception & AI

OpenCVPyTorchONNX Runtime

Mapping & navigation

slam_toolboxNav2

Manipulation

MoveIt 2OMPL

Engineering

Git/GitHubDockerVS Coderosbag2RViz2
Your laptop
CPU
64-bit, 4+ cores
Memory
16 GB (8 GB min.)
Storage
50 GB free
OS
Ubuntu 24.04, native or dual boot
GPU
Optional

Outcomes

What you leave with

  • A complete, integrated robot software stack running in simulation
  • Weekly benchmark reports that document your engineering evidence
  • Failure-analysis experience, beyond happy-path demos
  • Hands-on sim-to-real experience with a physical ROS 2 deployment
  • A Git repository and capstone report you can present to employers

Assessment

Measured, reviewed, defended

  • Weekly output: a working capability, one failure analysis, one benchmark and a short README
  • Version control: all code submitted through Git with clear commits
  • Code review on every weekly submission
  • Capstone: full mission demo, repository, report and live review

Career pathway

3month paid
internship

* Conditions apply

Eligible graduates can move straight into a 3-month paid internship on live robotics projects, turning programme work into real engineering contributions.

3+meaningful pull requests submitted for review
1reproducible technical demo
1documentation or tutorial contribution
1testing, dataset or benchmark contribution

Roles this programme prepares you for

Robotics Software EngineerROS 2 infrastructure & integration
Autonomy / Navigation EngineerEstimation, SLAM & path planning
Perception EngineerCalibration, detection & tracking
Manipulation EngineerArm planning, grasping, pick-and-place
Simulation EngineerDigital twins & evaluation worlds
Robot Learning / Physical AI EngineerDatasets & learned policies
Validation / Deployment EngineerBenchmarking, testing & sim-to-real

FAQ

Common questions

Is a physical robot kit included?

The main programme is simulation-based. A guided, low-cost physical deployment lab is included. You don't need to own a robot unless a specific batch states otherwise.

Does every learner get the internship?

No. The 3-month paid internship is offered to eligible graduates based on weekly work and capstone performance. Conditions apply.

Do I need a high-end computer?

No. A GPU is optional, and all graded work can be completed in Gazebo. See the laptop guidance above.

Will I master every advanced topic in one week?

Core topics are hands-on and assessed. Some advanced tools are introduced as overviews and are not promised as deep mastery within a single week.

Does the full mobile manipulator move to a real robot?

No. The programme covers integration, packaging, testing and sim-to-real practice. The physical hardware bridge is taught through the deployment lab.

Is job placement guaranteed?

No. The programme builds the skills, engineering evidence and portfolio quality that robotics roles look for.

Admissions · November 2026 batch

Reserve your seat in the first batch

  1. Apply online at quanskill.com
  2. Readiness call: a short fit conversation with our team
  3. Confirm your seat for the batch starting late November 2026
Apply at quanskill.com
Phone+84 904289555
Emailinfo@quanskill.com
Websitewww.quanskill.com