Gimbal lock — problem with Euler angles at θ = ±90°

3.5.5 · Physics › Guidance, Navigation & Control (GNC)

KYA ho raha hai

Middle angle kyun matter karta hai? Middle rotation hi last axis ko first axis ke relative tilt karti hai. Use $90°$ tak push karo aur last axis exactly first axis ke along aa jaati hai.

Lock ko first principles se KAISE derive karein

Hum "lock at 90°" memorize nahi karte — hum ise matrices mein hote hua dekhte hain.

Elemental rotations (right-handed, active):

R_{z} (ψ) = c_{ψ} s_{ψ} 0 - s_{ψ} c_{ψ} 0 001, R_{y} (θ) = c_{θ} 0 - s_{θ} 010 s_{θ} 0 c_{θ}, R_{x} (ϕ) = 100 0 c_{ϕ} s_{ϕ} 0 - s_{ϕ} c_{ϕ}

jahan $c_{α} = cos α, s_{α} = sin α$ .

Step — pitch ko singular value $θ = + 90°$ par set karo. Yeh step kyun? Kyunki hum suspect karte hain ki singularity wahin hai; hum ise directly test karte hain. Tab $c_{θ} = 0, s_{θ} = 1$ : $R_{y} (90°) = 00 - 1 010100$

Step — $R_{z} (ψ) R_{y} (90°) R_{x} (ϕ)$ ko multiply karo. Yeh step kyun? Hum dekhna chahte hain ki $ψ$ aur $ϕ$ abhi bhi independently kaam karte hain ya nahi.

Algebra karne par, poori matrix collapse hokar yeh ban jaati hai:

R = 00 - 1 sin (ϕ - ψ) cos (ϕ - ψ) 0 cos (ϕ - ψ) - sin (ϕ - ψ) 0

Step — result padho. Yeh step kyun? Yahi punchline hai. Har entry $ψ$ aur $ϕ$ par sirf combination $(ϕ - ψ)$ ke through depend karti hai. Toh $ψ$ ko $+ δ$ se aur $ϕ$ ko $+ δ$ se badalne par $R$ unchanged rehta hai. Do knobs, ek effect.

Yeh rates ko KAISE barbad karta hai (practical disaster)

Orientation static nahi hoti — GNC body angular rates $(p, q, r)$ ko Euler angles mein integrate karta hai. Kinematic relation yeh hai:

\dot{ϕ} \dot{θ} \dot{ψ} = 100 sin ϕ tan θ cos ϕ \frac{sin ϕ}{cos θ} cos ϕ tan θ - sin ϕ \frac{cos ϕ}{cos θ} p q r

Blow-up kyun hota hai: $tan θ$ aur $1/ cos θ$ dekho. Jab $θ \to \pm 90°$ , $cos θ \to 0$ , toh $\dot{ϕ}$ aur $\dot{ψ}$ infinity tak diverge ho jaate hain. Math ek perfectly finite physical rotation represent karne ke liye infinite angle rate maangta hai — coordinates singular hain, physics nahi.

Figure — Gimbal lock — problem with Euler angles at θ = ±90°

Worked examples

Steel-manned mistakes

Engineer ka fix

Orientation ko internally quaternions (ya rotation matrices / axis-angle) se represent karo.
Euler angles sirf human display ke liye convert karo, aur $θ = \pm 90°$ special case ko explicitly handle karo.
80/20 takeaway: physics kabhi singular nahi hoti; sirf Euler chart hoti hai. Chart badlo.

Recall Feynman: ek 12-saal ke bacche ko samjhao

Socho tum ek toy plane teen dials se steer kar rahe ho: left/right turn, nose-up/down, aur barrel-roll. Ab naak ko seedha upar ceiling ki taraf point karo. Ab "turn left/right" dial aur "barrel-roll" dial dono sirf plane ko same up-down line ke around top ki tarah ghuma rahe hain — dono bilkul same kaam karte hain! Tumne steering ka ek tarika kho diya. Plane abhi bhi kisi bhi taraf tumble kar sakta hai; tumhare teen dials confuse ho gaye hain. Astronauts ne isse ek smarter set of "dials" use karke fix kiya jishe quaternions kehte hain jo kabhi stuck nahi hote.

Flashcards

Gimbal lock kya hai ek sentence mein?

Ek Euler-angle representation mein ek rotational degree of freedom ka loss jab middle (pitch) angle

\pm 90°

tak pahunche, jisse pehla aur teesra rotation axis align ho jaata hai.

ZYX gimbal lock kis Euler angle par hota hai?

Jab pitch

θ = \pm 90°

ho.

Lock ke paas Euler rates kyun blow up hote hain?

Kinematic map mein

tan θ

aur

1/ cos θ

hote hain; jab

θ \to \pm 90°

cos θ \to 0

toh

\dot{ϕ}, \dot{ψ} \to \infty

θ = + 90°

par rotation matrix yaw/roll par sirf kis combination ke through depend karti hai?

Sirf

(ϕ - ψ)

ke through (aur

θ = - 90°

par

(ϕ + ψ)

ke through).

Kya gimbal lock rigid body ki physical limitation hai?

Nahi — body freely rotate karti hai; sirf coordinate chart (aur ise implement karne wale analog gimbals) ek DOF kho deta hai.

Gimbal lock ka standard engineering fix kya hai?

Internally quaternions (ya rotation matrices) use karo; Euler angles sirf display ke liye.

Kya quaternions mein gimbal lock hota hai?

Nahi; yeh

S O (3)

ka ek smooth global cover form karte hain jisme koi singular orientations nahi hain (sirf ek harmless

q \equiv - q

sign ambiguity hai).

Exactly

θ = 90°

se bachna kyun insufficient hai?

1/ cos θ

term ek poore neighborhood ko numerically unstable bana deta hai, sirf single point ko nahi.

Connections

Euler angles — woh representation jo fail hoti hai
Quaternions — singularity-free fix
Rotation matrices SO(3) — woh group jo parametrize ho rahi hai
Angular velocity kinematics — $1/ cos θ$ rate blow-up ka source
Attitude determination and control — GNC mein yeh kahan daata hai
Apollo Guidance Computer — historic gimbal-lock avoidance ("gimbal lock zone mein mat udao")
Singularities of coordinate charts — deeper differential-geometry reason