Gauss's method for Lambert's problem — Hinglish notes

3.2.29 · Physics › Orbital Mechanics & Astrodynamics

WHAT is Lambert's problem?

WHY it matters: har interplanetary transfer, rendezvous, aur orbit-determination-from-two-observations isi pe reduce hota hai. Yeh mission design ka workhorse hai (porkchop plots Lambert ko lakho baar solve karke bante hain).

The geometric quantities we need

Maano $θ = Δ ν$ transfer angle hai (do points ke beech true anomaly ka change).

Gauss ka key ratio $y$ derive karna

Kepler ka 2nd law kehta hai radius vector equal areas in equal times sweep karta hai, toh $Δ t = \frac{2 ( area of elliptic sector swept )}{h}, h = μ p .$

Do radii aur chord ek triangle bhi banate hain jiska area hai $A_{△} = \frac{1}{2} r_{1} r_{2} sin θ .$

Gauss ne sector-to-triangle ratio ko semi-latus rectum $p$ ke through express kiya. Ellipse ke liye sector area work karne par do fundamental relations milti hain:

Do equations kyun? Ek ( $y^{2} = \frac{m}{ℓ + x}$ ) geometry ↔ auxiliary variable hai; doosri ( $y^{2} (y - 1) = \dots$ ) Kepler's equation ke through time ↔ auxiliary variable hai. True orbit dono simultaneously satisfy karta hai → coupled pair solve karo.

HOW to solve — the iteration

Jab $y$ known ho jaata hai, orbit recover karo:

Figure — Gauss's method for Lambert's problem

Worked example

Worked example Ek short-way transfer

Given (canonical units, $μ = 1$ ): $r_{1} = 2$ , $r_{2} = 3$ , $θ = 4 0^{\circ}$ , $Δ t = 1.8$ .

Step 1 — geometry. $cos 4 0^{\circ} = 0.766$ , $cos (2 0^{\circ}) = 0.9397$ . $ℓ = \frac{2 + 3}{4 6 ( 0.9397 )} - \frac{1}{2} = \frac{5}{9.206} - 0.5 = 0.0433.$ Kyun? Geometry constants mein seedha formula plug-in.

Step 2 — time constant. $m = \frac{1 \cdot 1. 8 ^{2}}{( 2 6 \cdot 0.9397 ) ^{3}} = \frac{3.24}{( 4.603 ) ^{3}} = \frac{3.24}{97.5} = 0.03323.$

Step 3 — iterate. $y = 1$ se start karo:

$x = m / y^{2} - ℓ = 0.03323 - 0.0433 = - 0.0101$
$X (- 0.0101) \approx \frac{4}{3} (1 + 1.2 (- 0.0101)) = 1.317$
$y = 1 + (ℓ + x) (x - \frac{1}{2} + X) = 1 + (0.0332) (- 0.0101 - 0.5 + 1.317) = 1 + (0.0332) (0.807) = 1.0268$

Agle step mein: $x = 0.03323/1.026 8^{2} - 0.0433 = 0.0315 - 0.0433 = - 0.0118$ ; $X$ recompute karo aur update karo → $y$ ~ $1.027$ ke paas settle ho jaata hai. ~3 steps mein converge.

Step 4 — recover. Converged $y$ use karke $p$ nikalo, phir $f, g, \overset{g}{˙}$ , phir $v_{1}, v_{2}$ . Isko trust kyun karein? Kyunki dono Gauss equations ab satisfy ho rahi hain, geometry AUR time-of-flight simultaneously consistent hain — yahi connecting orbit hai.

Steel-manned mistakes

Active recall

Recall Khud test karo (answers chhupa lo)

Lambert's problem ke inputs aur outputs kya hain?
Ratio $y$ physically kya compare karta hai?
Gauss ke method ko do equations ki zaroorat kyun hai?
Method kahan fail hota hai, aur kyun?
Jab $y$ mil jaaye toh $v_{1}$ kaise milta hai?

Recall Feynman: ek 12-saal ke bachche ko samjhao

Socho ek patthar phenka jaata hai taaki woh ek door ki pahaadi par bilkul 3 seconds mein gire. Agar main tumhe bataaun kahaan se shuru hua, kahaan gira, aur kitna time liya, toh exactly ek curved path hai (chosen curve direction ke liye) jo fit karta hai. Gauss ki trick: usne us "pie slice" ki compare ki jise planet sweep karta hai us flat "triangle" se jo do spots ke beech hai. Woh guess karta hai ki unka ratio lagbhag 1 hai, check karta hai ki kya woh guess sahi amount of time lega, aur guess nudge karta hai jab tak clock match na kare. Jab match ho jaaye, usne path dhundh liya — aur path se ushe pata chal jaata hai ki kitni tezi se ja raha tha.

Flashcards

Lambert's problem input mein kya leta hai aur output mein kya produce karta hai?

Input:

r_{1}, r_{2}

aur time-of-flight

Δ t

. Output: connecting Kepler orbit, yaani

v_{1}, v_{2}

Gauss ke method mein ratio

y

kya represent karta hai?

Swept elliptic sector area ka triangle area se ratio, jahan triangle do radii aur chord ke beech banta hai.

Gauss ka method do coupled equations kyun use karta hai?

Ek geometry ko auxiliary variable se link karta hai (

y^{2} = m / (ℓ + x)

); doosra time-of-flight ko link karta hai (

y^{2} (y - 1) = m (x - \frac{1}{2} + X)

). True orbit dono satisfy karta hai.

y

ke liye sahi iteration update kya hai?

y = 1 + (ℓ + x) (x - \frac{1}{2} + X (x))

, jo 2nd Gauss equation ko

y^{2}

se divide karke aur 1st use karke milta hai.

Transfer ke liye chord length formula kya hai?

c^{2} = r_{1}^{2} + r_{2}^{2} - 2 r_{1} r_{2} cos θ

(law of cosines).

Δ t

Gauss ke constant

m

mein kaise enter karta hai?

Δ t^{2}

ke roop mein:

m = μ Δ t^{2} / (2 r_{1} r_{2} cos (θ /2))^{3}

(Kepler-3 scaling).

Gauss ka method kahan break down karta hai?

θ = 180°

ke paas (antipodal), kyunki

cos (θ /2) \to 0

ℓ, m

blow up ho jaate hain aur iteration fail ho jaata hai.

Jab

y

(hence

p

) known ho, velocities kaise recover hoti hain?

Lagrange coefficients ke zariye:

v_{1} = (r_{2} - f r_{1}) / g

v_{2} = (\overset{g}{˙} r_{2} - r_{1}) / g

Iteration ke liye achha starting guess kya hai aur kyun?

y = 1

, kyunki modest transfer angles ke liye sector ≈ triangle hota hai, aur

y

1 ke paas rehta hai.

Connections

Lambert's problem — parent problem
Kepler's equation — time↔eccentric-anomaly link provide karta hai
Lagrange f and g functions — velocities extract karne ke liye use hota hai
Universal variable formulation — $180°$ ke paas robust alternative
Battin's method aur Izzo Lambert solver — modern replacements
Kepler's second law — sector/triangle ratio ka equal-area basis
Porkchop plot — application: launch windows ke liye mass Lambert solves