Day 04 ยท Report
๐ธ๐ฌ Singapore Trip Report โ Day 4
- 09 September 2026
- Singapore
- 6 min read
- 14 photos

Timeline ยท 7 stops
Back at ITE College East, but today the robots leave the floor and take to the air. Drone workshop day โ programming the HULA DRONE with block coding to take off, fly patterns, and land on target. Yesterday was 2D, today is 3D, and the difference is bigger than I expected.
08:00 ยท Singapore Time๐ Coach to ITE College East (Again)
Same routine as yesterday. Breakfast, formal outfit, out the door. Same chartered coach picking us up outside the hotel, same 45-minute ride east through Bugis, past the CBD, out towards Simei. Bus AC still cranked. Group energy noticeably higher though โ everyone's excited because drones.
๐ก What I learned
- Sticking to the same daily routine on a school trip actually helps. You stop worrying about logistics and just focus on the training.
- Even a familiar route feels new when you know the day's going to be different. Same road, totally different mindset.
- Formal outfit day 2 = wrinkles becoming visible. Note for next trip: pack more shirts.
09:00 ยท Singapore Time๐ซ Back at ITE College East
Different room this time. Instead of the robotics lab we used yesterday, we got a bigger open space with soft mats on the floor and safety netting around the flight zone โ because drones flying into people or walls is a real thing. Instructors did a proper safety briefing before anyone touched a drone. No hoodies, no loose sleeves near propellers, phones out of the flight zone.
๐ก What I learned
- Drones are not toys. Even small educational ones have spinning plastic propellers that can slice a finger or eye. Rules exist for a reason.
- Flight zones need soft floors (mats) and boundaries (nets) because drones drift, batteries die, and code fails.
- Preparing the environment is half of drone work. Real pilots do it too โ check the space, check the wind, check the exits.
09:30 ยท Singapore Time๐ Meet the HULA DRONE


Got introduced to the HULA DRONE โ a small educational quadcopter (4 rotors) with propeller guards for indoor safety. Lightweight, palm-sized, connects wirelessly to the laptop. Instructor broke down the parts: 4 motors, 4 propellers (2 spin clockwise, 2 counter-clockwise to cancel torque), flight controller board with a built-in gyroscope + accelerometer (IMU) for stability, a barometer for altitude, and the battery โ good for about 7 minutes per charge.
Then we opened the block-coding app. Same idea as mBlock from yesterday โ drag blocks, snap them together, click run. First blocks we tried: takeoff, wait 3 seconds, land. That's it. A full working program in 3 blocks.
๐ก What I learned
- Quadcopter physics: 2 propellers spin CW, 2 spin CCW so their torque cancels out. Otherwise the drone body would just spin in place.
- The IMU (gyro + accelerometer) runs stability calculations hundreds of times per second โ that's why the drone hovers instead of tipping over.
- Educational drone battery life is short (5โ10 min). You spend most of the class coding, not flying. That's actually good โ forces you to plan.
10:30 ยท Singapore Time๐ฎ First Flight

Nerves. Real nerves. Clicked "run" and my drone lifted off, hovered for 3 seconds at about waist height, then landed. It justโฆ worked. I know it's basic but seeing your own code make a physical object leave the ground is a completely different feeling from watching a wheel spin.
Second attempt added a bit of movement โ takeoff, forward 50 cm, wait, land. Went slightly further than expected because I didn't account for drift. Still landed inside the net. Small win.
๐ก What I learned
- Takeoff and landing are the two hardest phases of any flight โ real aviation and drones both. Most crashes happen in these two moments.
- Drones drift because of air currents, uneven motor calibration, and battery voltage dropping over time. Your code has to allow for it.
- The trust curve is real. First flight: terrified. Fifth flight: casual. Same code, same drone โ you just get used to it.
13:30 ยท Singapore Time๐งญ Coding Movement Patterns

After lunch, we went deeper. Instructor introduced the three axes of drone movement:
- Pitch โ tilt forward/back โ moves forward/back
- Roll โ tilt left/right โ strafes sideways
- Yaw โ rotate around vertical axis โ spins in place
Also altitude (up/down) which is separate. So a drone has 4 controls at once, vs the mBot which just had 2 wheel speeds. Way more to think about.
Coded a sequence: takeoff โ forward 1 m โ turn right 90ยฐ โ forward 1 m โ turn right 90ยฐ โ forward 1 m โ turn right 90ยฐ โ forward 1 m โ land. If the drone flew a clean square, the code worked. Mine did most of a square โ the last turn was under-rotated so it landed slightly off. Adjusted the turn time by 200 ms and got it right on the third try.
๐ก What I learned
- 6 degrees of freedom โ a drone can move in 3 directions (x, y, z) AND rotate around 3 axes (pitch, roll, yaw). Way more than a ground robot.
- Timing-based movement is inaccurate because motor speed varies with battery level. Real drones use positioning systems (GPS outdoors, optical flow indoors) instead of timing.
- Coding in 3D is harder than 2D. Every mistake happens in the air, and gravity is not forgiving.
15:00 ยท Singapore Time๐ Mission: Pattern Challenge

Final task: program a full autonomous mission โ takeoff, fly a set pattern through checkpoints (marked on the floor with cones and small hoops), and land on a target square. No manual control allowed. Just press run and hope.
My mission plan: takeoff โ forward โ up โ forward through the hoop โ right โ down โ land on target. Took 4 attempts. Failed 1 (landed off-target), failed 2 (hit the hoop edge), failed 3 (drifted too far left). Attempt 4 โ clean run. Landed dead centre on the target square. First time all week I actually cheered out loud.
๐ก What I learned
- Autonomous missions require planning on paper first. You can't debug a 30-second flight in real time โ you have to know exactly what each block should do before you run it.
- Iteration is everything. Each failed attempt teaches you one specific thing to fix. Attempts 1โ3 weren't wasted, they were data.
- The satisfaction of code + hardware working together is unmatched. This is the feeling I came here for.
16:30 ยท Singapore Time๐ Wrap-up
Instructors ran through everyone's mission runs, pointed out cool solutions, and gave feedback. Reminded us tomorrow is AI Robotics โ the last hands-on module before certification. Packed up, formal shirt now visibly wrinkled but I don't even care, boarded the coach back to the hotel.
๐ก What I learned
- 2D robotics (mBot) vs 3D robotics (drone) are worlds apart in difficulty. Adding one dimension (altitude) triples the complexity.
- Block coding scales. The same drag-and-drop concepts run drones, robots, apps, and games. It's not a "kid" tool โ it's a real one.
- Watching a drone execute your code autonomously hits different from watching a script run in a terminal. Physical output changes how you feel about programming.
Wrap-up
Flew something today. That's a sentence I couldn't have written a week ago. Also confirmed what I felt after Day 3 โ I really like this code + physical hardware combo. When it works, you see it work. When it fails, you see that too, usually spectacularly. Tomorrow: AI Robotics. Final training day before the cert exam. Bringing it. Goodnight.
Photo journal







