Pitch to PyBullet

Think of learning drone simulation exactly like starting your cricket training. You don't walk out onto a stadium pitch and immediately try to hit sixes off a fast bowler. If you don't know your stance, your balance, or your footwork, you will get knocked over instantly.

Drones are unique because they are underactuated. In plain English, that means they can move around in 6 different ways (Up/Down, Left/Right, Forward/Backward, Roll, Pitch, Yaw), but they only have 4 basic controls (the speeds of their 4 motors). To move forward, a drone actually has to tilt its nose downward so its motors push it across the room. This step-by-step guide is your basic net practice to master these physics inside the PyBullet simulator.

Phase 1: Pitch Inspection (The Basics)

Before an open match, the captain walks out to inspect the pitch. Is the grass damp? How will the ball bounce? In drone terms, this means looking at the drone's State Vector to see exactly where it is and how fast it is moving.

PyBullet tracks everything about the drone using a simple array of 20 numbers. We break that down to find our current coordinates, tilt angles, and speed:

# Reading the pitch metrics
state = obs["0"]["state"]
pos = state[0:3]     # Where is the drone in 3D space? [X, Y, Z]
quat = state[3:7]    # Raw orientation numbers (Quaternions)
rpy = state[7:10]    # Simple tilt angles: [Roll, Pitch, Yaw]
vel = state[10:13]   # How fast is it traveling linearly?
ang_v = state[13:16] # How fast is it spinning/twisting?

Phase 2: Stance & Balance (Takeoff & Hover)

Once you are on the field, you take your batting stance. You keep your feet firm, your head steady, and your weight perfectly balanced so you do not tip over. For a drone, taking a stance means taking off and locking into a perfectly still hover at 1 meter high (Z = 1.0).

To do this, we use a PID Controller. Imagine it like a continuous feedback brain: it calculates how far the drone is from its target height and automatically fine-tunes the RPM speed of all 4 propellers to fight gravity without shaking or crashing.

Phase 3: Footwork (Waypoint Navigation)

When a spinner tosses the ball wide, you change your footwork. You step cleanly out of your crease, move directly to the pitch of the ball, and hit a beautiful shot. Phase 3 teaches our drone proper footwork by making it fly to a sequence of coordinates forming an exact square pattern in mid-air.

The code keeps tabs on the drone's position. The moment it gets close enough to its current target coordinate, it cleanly switches its focus to the next corner of the square.

Phase 4: Timing vs. Power (Velocity vs. Thrust Control)

In batting, you have two styles. You can use pure timing—gently opening your bat face to let a fast ball guide itself past slip for a boundary. Or you can use brute power—using your muscles to swing hard and smash a pull shot over the boundary rope.

Drones control movement using those exact same two philosophies:

• Timing (Kinematic Control): You act like a director. You tell the simulator: "Hey, make the drone move forward at exactly 0.5 meters per second." The simulator handles the background math automatically.
• Power (Dynamic Control): You act like the engine itself. You bypass the smart helpers and tell the drone directly: "Spin Motor 1 at 15,000 RPM, and Motor 2 at 14,800 RPM!" You are controlling raw physical thrust.

Phase 5: Returning to the Crease (Safe Landing)

You step down the pitch to smash a shot, but you miss. What is your immediate priority? You scramble your body back into the crease instantly before the wicketkeeper stums you out! You must get back to safe ground cleanly.

Phase 5 handles our landing sequence. Instead of cutting the power instantly (which would make the drone drop like a stone and smash into pieces), the code slowly lowers the target altitude step-by-step. Once the sensors detect that the drone is safely touching the grass, it completely cuts the motor power to rest.