You are on page 1of 30

ACT Global Optimization Competition Workshop

Mikhail
Mikhail S.KONSTANTINOV
S.KONSTANTINOV
Gennadiy
Gennadiy G.
G. FEDOTOV
FEDOTOV
Moscow
Moscow Aviation
Aviation Institute
Institute

Viacheslav
Viacheslav G.
G. PETUKHOV
PETUKHOV
Khrunichev
Khrunichev State Research
State Research and
and Production
Production Space
Space Center
Center
OBJECTIVE FUNCTION: PRELIMINARY CONSIDERATION

J = mf ⋅|(v-vast)Tvast|

Launch date and transfer


Impact point in the vicinity
duration should be
of asteroid’s perihelion
optimized
1) Retrograde orbit
2) Impact in the vicinity Giant planets swingbys or
of S/C perihelion low-thrust should be
3) S/C aphelion distance considered
should be maximized
1) Swingbys are prefered in
Minimum propellant comparison with thrusting
2) Transfer duration tends to
maximal
THEREFORE…

Thanks ACT & Dario Izzo


for an interesting problem
FLYBY CONSTRAINTS: ASYMPTOTIC VELOCITY ROTATION

Maximal angles ⇒ flyby sequences E…JSA or E…JSJA are prefered

180

160

140 Mercury
Venus
120
Earth
Max betta, deg.

100 Mars
Jupiter
80
- heliocentric parabola
Saturn

60

40

20

0
0 5 10 15 20 25 30 35 40 45 50
Vinf, km/s
LOW-THRUST TRAJECTORY: BOUNDARY CONDITIONS

Initial point: Swingby:

x(0 ) = x 0 , + V∞+ −
p = − pv ,
v
p v (0 )
V∞
v (0 ) = v 0 + V∞ , v + = v pl + V∞+ ,
pv (0 )
m(0 ) = m0 . x = x pl ,

where
⎛ ⎞
⎛ cos β ⎞ ⎜ ⎟
⎜ ⎟ ⎜ 1 ⎟
Final point: V∞+ = M ⋅ ⎜ sin β cos γ ⎟, β max = π − 2 arccos⎜ 2 ⎟
⎜ sin β sin γ ⎟ ⎜1+ r V ∞ ⎟
x(T ) = x ast ,
p min
⎝ ⎠ ⎜ ⎟
⎝ µ ⎠
( T
)
p v (T ) = m(T )sgn (v (T ) − v ast ) v ast v ast , ⎛ − V∞−yV∞− −
V∞xV∞z− ⎞
p m (T ) = 1.
⎜V∞x − − ⎟


(V ) + (V )
− 2
∞x
− 2
∞y (V∞−x ) + (V∞−y ) ⎟⎟
2 2

⎜ V∞−xV∞− V∞−yV∞−z ⎟
M = ⎜V∞−y − ⎟
⎜ (V ) + (V )
− 2
∞x
− 2
∞y (V∞x ) + (V∞y ) ⎟
− 2 − 2

⎜ − ⎟
⎜ V∞z 0 (V∞x ) + (V∞y ) ⎟
− 2 − 2

⎜ ⎟
⎝ ⎠
VERSIONS OF TRAJECTORY PROFILE AND COMPUTATIONAL TECHNIQUES

1. Flyby sequence in the inner Solar system


(aphelion increasing + apsidal line positioning)
1.1. EV…E or EV…V coast flyby sequence
1.2. EE, EV, VE or VV trajectory using low-thrust arc
2. Retrograde trajectory shaping in the outer Solar system
using Giants flybys and/or low thrust
EJSJA, EJSA, ESJA, ESA or EJA flyby sequence using low-thrust arcs if necessary
- or -
•• Lambert
Lambert solver
solver
EA trajectory using low-thrust arc(s). •• Conic-patched
Conic-patched interplanetary
interplanetary trajectory
trajectory calculation
calculation based
based on
on
the
the Kepler
Kepler equation
equation
•• TPBVP
TPBVP solver (numerical
solver (numerical integration,
integration, targeting
targeting to
to the
the next
next
planet varying the trajectory arc duration and departure
planet varying the trajectory arc duration and departure
asymptotic
asymptotic velocity
velocity direction).
direction). Trajectory
Trajectory arc
arc can
can include
include one
one
thrusting
thrusting arc
arc using
using parametric
parametric thrust
thrust steering
steering
•• Power-limited
Power-limited (LP)
(LP) problem
problem solver
solver
•• Constant
Constant ejection velocity (LP-to-CEV) problem
ejection velocity (LP-to-CEV) problem solver
solver
PROBLEM COMPLEXITY

1. Large number of routes under consideration:


Swingby

1st planet 2nd planet … Mth planet

1 orbit 2 orbits … N orbits

1st half of orbit 2nd half of orbit

Descending node Ascending node

2. Local optimization of low-thrust arcs: bifurcations, numerical stability and


convergence problems

Global optimization assumes using of reliable local optimization techniques.


Local optimization technique of multiply swingbys low-thrust trajectory is insufficiently elaborated.
So, some simplifications are inevitable.

Due
Due to
to problem
problem complexity
complexity
researcher’s
researcher’s intuition
intuition and
and experience
experience should
should be
be used
used
along
along with
with optimization
optimization techniques
techniques using
using
SOFTWARE

EPOCH – LP-problem solver E-ProTO – multi-orbits trajectory optimization,


CEV-problem

GALTT – Gravity-Assisted Low-Thrust


Trajectory Evaluation
PL – CEV-problem solver

Other software: Legend:


• LP → CEV continuation (frozen trajectory structure) - previously developed and exploited software
• Lambert solvers - modified software
• Conic-patched interplanetary trajectory calculation - new-developed software
based on the Kepler equation
TRAJECTORY DESCRIPTION: INITIAL FLYBY SEQUENCE

1.
1. EV,
EV, 440.667
440.667 days
days 2.
2. EVE,
EVE, +1322.383
+1322.383 days,
days,
pericenter
pericenter radius 40795.8 km
radius 40795.8 km

4.
4. EVEVE,
EVEVE, +368.930
+368.930 days
days ,,
3. pericenter
pericenter radius
radius 8756.6
8756.6 km
km
3. EVEV,
EVEV, +452.055
+452.055 days
days ,,
pericenter
pericenter radius
radius 16696.8
16696.8 km
km
TRAJECTORY DESCRIPTION: EARTH-to-EARTH TRAJECTORY ARC

1. Tangential thrust 2. LP-trajectory 3. CEV-trajectory


(braking)

1 2

Boundary conditions, Continuation 5.


5. EVEVEE,
EVEVEE, +1200.0
+1200.0 days
days ,,
continuation wrt. gravity LP → CEV, pericenter
pericenter radius
radius 7936.0
7936.0 km,
km,
transfer thrusting
thrusting 29.861 + 221.237 ==
29.861 + 221.237
parameter
251.098
251.098 days
days
duration
30
0
-30

Pitch, deg.
-60
-90
-120
-150
-180
8500 8700 8900 9100 9300 9500 9700
MJD2000
TYPICAL POWER-LIMITED PROBLEM

T
1 T d 2x
2 ∫0
Purpose: To minimize J = a adt for a dynamical system obeying the differential equations = Ω x + a,
dt 2
and having following boundary conditions: t = 0: x(0) =x0, dx(0)/dt = v0, m(0) = m0, t = T: x(T) = xf, dx(T)/dt = vf..
Here T is final time, a is thrust acceleration, x and v are vectors of spacecraft position and velocity respectively, Ω is gravity potential.

Let apply Pontryagin’s maximum principle to reduce this OCP to the TPBVP.
Hamiltonian of this dynamical system is H = -aTa/2 + pxTv + pvTΩx + pvTa. So, optimal control is a = pv and equations of optimal
motion become following:
d 2x ⎫
= Ωx + pv , ⎪
dt 2

2 ⎬
= −Ω xxp v ,⎪⎪
d pv
dt 2

The boundary conditions for rendezvous mission has form: t = 0: x(0) =x0, dx(0)/dt = v0, t = T: x(T) = xf, dx(T)/dt = vf .
⎛ x(T ) − x f ⎞ ⎛ p x ( 0) ⎞
In fact, it is necessary to solve equation f(z) = 0, where f (z ) = ⎜ ⎟ is vector of residuals, z = ⎜⎜ ⎟⎟ is vector of
⎜ v(T ) − v ⎟ ⎝ p v ( 0) ⎠
⎝ f ⎠
unknown TPBVP parameters, px = -dpv/dt.
ALGORITHMS FOR POWER-LIMITED PROBLEM

Continuation
Continuation (homotopy)
(homotopy) technique
technique is
is used
used for
for LP-problem
LP-problem solving.
solving.

Purpose
Purpose ofof continuation
continuation method
method
“Regularization”
“Regularization” of numerical trajectory
of numerical trajectory optimization,
optimization, i.e.
i.e. elimination
elimination (if
(if possible)
possible) the
the methodical
methodical
deffects
deffects of
of numerical
numerical optimization.
optimization. Particularly,
Particularly, there
there was
was stated
stated and
and solved
solved problem
problem of of trajectory
trajectory
optimization
optimization using
using trivial
trivial initial
initial approximation
approximation (the
(the coasting
coasting along
along the
the initial
initial orbit
orbit for
for example).
example).
CONTINUATION TECHNUQUE

Continuation (Homotopy) Technique Application to Optimal Control Problem (OCP)

dx ⎫
= Hp , ⎪
Problem: to solve non-linear system Optimal motion equations dt ⎪

f (z ) = 0 (1) (after applying the maximum principle): dp
= − H x ⎪⎪
with respect to vector z dt ⎭

Let z0 - initial approximation of solution. Then Boundary conditions (an example): x(0) = x 0 , x(T ) =x k
f ( z0 ) = b (2)
where b - residuals when z = z0.
Boundary value problem parameters and
Let consider immersion z(τ), where τ is a scalar parameter, residuals: z = p(0), f = x(T ) − x k
and equation f ( z) = (1 − τ )b
(3) ∂x(T )
with respect to z(τ). Obviously, z(1) is solution of eq. (1). Sensitivity matrix: fz =
∂z
Let differentiate eq. (2) on τ and resolve it with respect to dz/dτ:
dz
= − f z−1 ( z)b , z( 0) = z0 Associated system of optimal motion o.d.e.
dτ (4) and perturbation equations for residuals and
sensitivity matrix calculation: dx ⎫
Just after integrating eq. (4) from 0 to 1 we have solution of = Hp , ⎪
dt
eq. (1). ⎪
dp
= −H x , ⎪
Equation (4) is the differential equation of continuation dt ⎪⎪
algorithm (the formal reduction of non-linear system (1) into d ⎛ ∂x ⎞ ∂x ∂p ⎬
initial value problem (4)). ⎜ ⎟ = H px + H pp ,⎪
dt ⎝ ∂z ⎠ ∂z ∂z ⎪
d ⎛ ∂p ⎞ ∂x ∂p ⎪
Differential equations (4) are integrated using a high-order ⎜ ⎟ = − H xx − H xp ⎪
Runge-Kutta method with adaptive step size control. dt ⎝ ∂z ⎠ ∂z ∂z ⎪⎭

∂x ∂p
Extended initial conditions: x(0) = x 0 , x(T ) =x k , = 0, =I
∂z ∂z
OCP-SOLVER BASED ON CONTINUATION ALGORITHM

Reduction of optimal control problem


to the boundary value problem
by maximum principle

CONTINUATION METHOD
Initial residuals b calculation
Initial by optimal motion o.d.e. integrating
approximation z0 for given guess value z0
of boundary value problem parameters

1st version Continuation method’s o.d.e. integrating 2nd version


of o.d.e. right- dz of o.d.e. right-hand
= − f z−1 ( z)b , z( 0) = z0
hand side dτ side calculation
calculation with respect to τ from 0 to 1

Associated integrating of optimal motion Integrating of optimal motion equations for


equations and perturbation equations for current z(τ) to calculate current residuals f(z,τ)
current z(τ) to calculate current residuals f(z,τ) and for pertubed z(τ) to calculate fz(z,τ) by finite-
and sensitivity matrix fz(z,τ) difference technique

Solution
z(1)
CONTINUATION WITH RESPECT TO GRAVITY PARAMETER (1/3)

Occasional reasons of continuation algorithm failure: sensitivity matrix degeneration (bifurcation of optimal solutions)

Mostly bifurcations of optimal planetary If angular distance will remain constant during continuation, the continuation way in
trajectories are connected with different the parametric space will not cross boundaries of different kinds of optimal
number of complete orbits. trajectories. So, the sensitivity matrix will not degenerate.

The purpose of the technique modification - to fix angular distance of transfer during continuation

Earth-to-Mars, rendezvous,
launch date October 1, 2001,
V∞= 0 m/s, T=200 days

1 - Earth at launch
2 - Earth at arrival
3 - Mars at arrival
4 - intermediate trajectories (τ < 1)
5 - optimal trajectory (τ = 1)

Sequence of trajectory calculations using Sequence of trajectory calculations using


basic continuation method continuation with respect to gravity parameter
CONTINUATION WITH RESPECT TO GRAVITY PARAMETER (2/3)

Let x0(0), x0(T) - departure planet position when t=0 and t=T;
xk - target planet position when t=T. Let suppose primary gravity parameter to
be linear function of τ, and let choose initial value of this gravity parameter µ0
using following condition:

1) angular distances of transfer are equal when τ=0 and τ=1;


2) When τ=1 primary gravity parameter equals to its real value (1 for
dimensionless equations)

The initial approximation is SC coast motion along departure planet orbit. Let
the initial true anomaly equals to ν0 at the start point S, and the final one equals
to νk=ν0+ϕ at the final point K (ϕ is angle between x0 and projection of xk into
the initial orbit plane).

The solution of Kepler equation gives corresponding values of mean anomalies M0 and Mk (M=E-e⋅sinE, where
E=2⋅arctg{[(1-e)/(1+e)]0.5tg(ν/2)} is eccentric anomaly). Mean anomaly is linear function of time at the keplerian orbit: M=M0+n⋅(t-t0),
where n=(µ0/a3)0.5 is mean motion. Therefore, the condition of angular distance invariance is Mk+2π Nrev=nT+M0, where Nrev is number
of complete orbits. So initial value of the primary gravity parameter is
µ0=[( Mk+2π Nrev - M0)/T]2a3,
and current one is
µ(τ)=µ0+(1-µ0) τ.
The shape and size of orbits should be invariance with respect to τ, therefore
v(t, τ)=µ(τ)0.5 v(t, 1).
CONTINUATION WITH RESPECT TO GRAVITY PARAMETER (3/3)

Equations of motion: x&& = µ ( τ)Ω x + p v , p


&& v = µ ( τ)Ω xx p v
x (0) = x 0 , x& (0) = µ 1/ 2 ( τ) v 0 ,
Boundary conditions:
x (T ) = x k , x& (T ) = µ 1/ 2 ( τ) v k .
⎛ x (T ) − x k ⎞
Residuals: f =⎜ ⎟
⎝ x (T ) − µ ( τ) v k ⎠
1/ 2

z = (pv(0), dpv(0)/dt)T = ( p v o , p& v o )


T
Boundary value problem parameters:

dz ⎛ ∂f ⎞
Equation of continuation method: = −f z−1 (z )⎜ b + ⎟, z (0) = z 0
dτ ⎝ ∂τ ⎠
where
d 2 ⎛ ∂x ⎞ ∂x ∂p v
2 ⎜ ⎟ = µ ( τ)Ω xx + ,
b = f(z0) dt ⎝ ∂z ⎠ ∂z ∂z
d 2 ⎛ ∂p v ⎞ ⎡∂ ∂x ∂p v ⎤
2 ⎜ ⎟ = µ ( τ) ⎢ (Ω xx p v ) + Ω xx ,
⎛ ∂x (T ) ∂p v o ∂x (T ) ∂p& v o ⎞ dt ⎝ ∂z ⎠ ⎣ ∂x ∂z ∂z ⎥⎦
f z = ⎜⎜ d ⎟
⎝dt
( ∂x (T ) ∂p v o ) d
( ∂x ( T ) ∂p& v o ) ⎟

d 2 ⎛ ∂x ⎞ ∂µ
⎟= Ω x + µ ( τ)Ω xx
∂x ∂p v
+
2 ⎜
dt ,
dt ⎝ ∂τ ⎠ ∂τ ∂z ∂z
d 2 ⎛ ∂p v ⎞ ∂µ ⎡∂ ∂x ∂p v ⎤
2 ⎜ ⎟= Ω xx p v + µ ( τ) ⎢ (Ω xx p v ) + Ω xx ,
⎛ ∂x ⎞ dt ⎝ ∂τ ⎠ ∂τ ⎣ ∂x ∂z ∂z ⎥⎦
∂f ⎜ ⎟
= ⎜ ∂x& ∂τ ⎟
∂τ ⎜ − 1 ∂µ ⎟ ∂x (0) ∂x& (0) ∂x (0) ∂x& (0) 1 ∂µ
⎜ v k⎟ = = = 0, = v0 ,
⎝ ∂τ 2µ 1/ 2 ( τ) ∂τ ⎠ ∂z ∂z ∂τ ∂τ 2µ ( τ) ∂τ
1/ 2

∂p v (0) ∂p& v (0) ∂p v (0) ∂p& v (0) ∂p v (0) ∂p& v (0)


= = E, = = = = 0.
∂p v0 ∂p& v0 ∂p& v0 ∂p v0 ∂τ ∂τ
TYPICAL CEV-PROBLEM WITH THRUST SWITCHINGS

d 2x P ⎫
T
P = Ω x + δ e,⎪⎪
Purpose: To minimize J = ∫ δ dt for a dynamical system obeying the differential equations dt 2
m ⎬
w dm P
0
= −δ , ⎪
dt w ⎪⎭
and having following boundary conditions: t = 0: x(0) =x0, dx(0)/dt = v0, m(0) = m0, t = T: x(T) = xf, dx(T)/dt = vf.
Here δ is step-like thrusting function, P – thrust, w – exhaust velocity, m – spacecraft mass.
d 2x P pv ⎫ d 2p v ⎫
Pontryagin’s maximum principle reduces this OCP to the following TPBVP: = Ωx + δ ,⎪ = −Ω xxp v , ⎪
dt 2
m pv ⎪ dt 2

⎬ ⎬
dm P
= −δ , ⎪ dpm
= −δ 2 pv ,⎪⎪
P
dt w ⎪⎭ dt m ⎭
⎧1, if ψ s > 0,
t = 0: x(0) =x0, dx(0)/dt = v0, m(0) = m0, t = T: x(T) = xf, dx(T)/dt = vf, pm(T) = 0, where step-like thrusting function δ = ⎨
⎩0, if ψ s ≤ 0,
pv 1 + pm
and switching function ψ s = −
m w ⎛ x(T ) − x f ⎞ ⎛ p x (0) ⎞
⎜ ⎟ ⎜ ⎟
In fact, it is necessary to solve equation f(z) = 0, where f (z ) = ⎜ v(T ) − v f ⎟ is vector of residuals, and z = ⎜ p v (0) ⎟
⎜ ⎟ ⎜ p (0) ⎟
⎝ pm (T ) ⎠ ⎝ m ⎠
LP-to-CEV CONTINUATION USING FROZEN TRAJECTORY STRUCTURE

1. Sequence of thrusting and coasting arcs is assumed fixed


d 2x ⎡ P ⎤
2. Equations of optimal motion are following: = Ωx + δ ⋅ ⎢(1 − τ ) + τ ⎥p v ,
dt 2 ⎣ mp v ⎦
(the equations correspond to LP-problem
d 2p v
if τ=0 and to CEV-problem if τ=1) = Ω xx p v ,
dt 2
dm P
= −δ ,
dt w
dp m P
= δ 2 pv .
dt m
pv 1 + p m
2. Switching function: ψ = −
m w
ψ(t) Thrusting
x ( 0) = x 0 ,
3. Initial conditions:
pv
v ( 0) = v 0 + V ∞ , ψ(t1)
pv
m ( 0) = m 0 . ψ(t2)
t1 t2 T t
4. Final conditions: ⎛ x − x ast ⎞
⎜ ⎟
( )
⎜ p v − τ ⋅ m sgn v v ast v t ⎟
T
∆t1 ∆t2 ∆t3
⎜ pm ⎟
⎜ ⎟
f= ⎜ ∑ ∆t i − T ⎟=0
⎜ i ⎟ k(∆ti)
⎜ k (∆t1 )ψ (t1 ) ⎟ 1
⎜ ⎟
⎜ M ⎟
⎝ ( arc
)(
⎜ k ∆t N −1 ψ t N −1
arc
)⎟

ε 2ε ∆ti
-1
Multiplier providing minimal arc length ε
EARTH-to-ASTEROID POWER-LIMITED TRAJECTORY

Continuation wrt. Continuation wrt.


gravity parameter boundary conditions
TRAJECTORY DESCRIPTION: FINAL (CEV) EARTH-to-ASTEROID TRAJECTORY

6.
6. EVEVEEA,
EVEVEEA, +6130.0
+6130.0 days
days ,,
pericenter
pericenter radius
radius 7600.0
7600.0 km,
km,
thrusting 2798.027 days
thrusting 2798.027 days

Pitch Yaw
0 0

-30

-60 -30
Pitch, deg.

Yaw, deg.

-90

-120 -60

-150

-180 -90
9800 10800 11800 12800 13800 14800 15800 9800 10800 11800 12800 13800 14800 15800
MJD2000 MJD 2000
FINAL (CEV) EARTH-to-ASTEROID TRAJECTORY

180
169.233°
150

Inclination, deg.
120

90

60

30

0
9800 10800 11800 12800 13800 14800 15800
MJD2000

0.95
Eccentricity

0.9

0.85 eemax = 0.998161


max = 0.998161

0.8

0.75
9800 10800 11800 12800 13800 14800 15800
MJD2000
FINAL (CEV) EARTH-to-ASTEROID TRAJECTORY

3.00E+08

2.50E+08 rrppmin = 1.755×1066 km


min = 1.755×10 km

Pericenter radius, km
2.00E+08

1.50E+08

1.00E+08

5.00E+07

0.00E+00
9800 10800 11800 12800 13800 14800 15800
MJD2000

2.04E+09
2.02E+09
Apocenter radius, km

2.00E+09
1.98E+09
1.96E+09
1.94E+09
1.92E+09
1.90E+09
1.88E+09
9800 10800 11800 12800 13800 14800 15800
MJD2000
FINAL (CEV) EARTH-to-ASTEROID TRAJECTORY

1.14E+09
1.12E+09
1.10E+09
1.08E+09

SMA, km
1.06E+09
1.04E+09
1.02E+09
1.00E+09
9.80E+08
9.60E+08
9.40E+08
9800 10800 11800 12800 13800 14800 15800
MJD2000

180 180
150 150

Argument of pericenter, deg.


120 120
RAAN, deg.

90 90
60 60
30 30
0 0
-30 -30
-60 -60
9800 10800 11800 12800 13800 14800 15800 9800 10800 11800 12800 13800 14800 15800
MJD2000 MJD2000
FINAL (CEV) EARTH-to-ASTEROID TRAJECTORY

2.00E+09
1.80E+09

Heliocentric distance, km
1.60E+09
1.40E+09
1.20E+09
1.00E+09
8.00E+08
6.00E+08
4.00E+08
2.00E+08
0.00E+00
9800 10800 11800 12800 13800 14800 15800
MJD2000

4.50E+01
4.00E+01
3.50E+01
Velocity, km/s

3.00E+01
VVfin = 28712.01 m/s
fin = 28712.01 m/s
2.50E+01
2.00E+01
1.50E+01 VVmin = 686.42 m/s
min = 686.42 m/s
1.00E+01
5.00E+00
0.00E+00
9800 10800 11800 12800 13800 14800 15800
MJD2000
COMPLETE TRAJECTORY

Objective
Objective function
function value:
value: 1364042.86 kg⋅km22/sec
1364042.86 kg⋅km /sec22;;
Route:
Route: EVEVEEA
EVEVEEA
Launch
Launch date:
date: JD
JD 2457556.7
2457556.7 (June
(June 17.2,
17.2, 2016)
2016)
Escape
Escape velocity:
velocity: 2.474133
2.474133 km/sec
km/sec
Total duration:
Total duration: 9914.034 days
9914.034 days
Final
Final S/C
S/C mass:
mass: 1070.187
1070.187 kg
kg

10. 2001 TW229 arrival

8. 5th swingby (Earth)

4. 3rd swingby (Venus)


7. 2nd thrusting arc
rd
9. 3 thrusting arc 2. 1st swingby (Venus)

3. 2nd swingby (Earth)


6. 1st thrusting arc
1. Earth departure

5. 4th swingby (Earth)


EXAMPLE OF LOCAL OPTIMIZATION COMPLEXITY:
DIRECT EARTH-to-ASTEROID TRAJECTORY
1. Initial S/C orbit: line of apsides along to asteroid’s line of apsides;
pericenter radius equals to earth orbit radius at departure date;
J=89000 J=330000 J=351000 apocenter radius corresponds to asymptotic velocity 2.5 km/s;
inclination equals to 0.
2. Final S/C orbit: line of apsides along to asteroid’s line of apsides;
pericenter radius equals to asteroid’s pericenter radius;
inclination and apocenter radius are varied.
3. Problem: minimum-time transfer to the final orbit
J=529000 with constrained minimal heliocentric distance (0.2 AU).
The constraint is regulated by number of orbits (continuation
wrt. gravity parameter), final inclination, and final apocenter
radius.
4. Solvers: a) Averaged optimal control problem (maximum principle,
continuation technique, E-ProTO software).
b) Unaveraged optimal control problem (maximum principle,
continuation technique, averaged solution as an intial guess,
E-ProTO software)
DIRECT EARTH-to-ASTEROID TRAJECTORY: RESULTS

INITIAL ORBIT: 1.0 × 1.4239976 AU, i = 0°


FINAL ORBIT: 1.8815331 × 8.1846558 AU, i = 136.5°
FINAL MASS: 436.365 kg
MINIMAL HELIOCENTRIC DISTANCE: 0.2009528 AU
TRANSFER DURATION: 20.658 (thrusting) + 4.784 (coasting) = 25.442 years
J = 529207.74 kg⋅km2/s2
TECHNIQUES OF MULTIREVOLUTIONAL OPTIMIZATION

Equinoctial orbital elements are used:


i
, e x = e cos (Ω + ω ), e y = e sin (Ω + ω ), i x = tan cos Ω, i y = tan sin Ω, F = ν + ω + Ω
p i
h=
µ 2 2

Here p, e, ω, ν, i, Ω are Keplerian elements, µ - primary gravity parameter. It is considered conventional CEV-problem without any
constraints on thrust direction.
Maximum principle reduces the problem into TPBVP. The numerical averaging over the orbital period is used for computational
consumption reducing and numerical stability increasing. A numerous versions of boundary conditions were considered.
The continuation (homotopy) procedure was used to solve minimum-time problem (see 4.1.1). The simple typical guess values:
ph0 = ±1, pex0 = pey0 = pix0 = piy0 = 0 (initial values of co-state variables), T = 1 (dimensionless orbital period referred to the initial
orbit) as a rule provides stable convergence of optimization. Of course, initial values of co-states from the OCP solution having
close boundary conditions provides improved convergence.
Solver of minimum-propellant problem uses minimum-time solution as initial approximation. The factored secant update algorithm
is used for minimum-propellant problem.
Both techniques demonstrates their robustness and efficiency and there were used for a numerous applied problems.
CONCLUSION

1. Tolerable objective function value was obtained without using


Jupiter/Saturn flybys
2. Global optimization should be supported by reliable methods of local
optimization
3. Continuation technique allows to find “global” minimum among local
minimums depending on restricted number of parameters (boundary
conditions, transfer duration, number of orbits)
4. THANK YOU FOR ATTENTION