力学変容に基づく二脚ロボットの左右方向移動制御

1B4
SY0004/14/0000-0049 © 2014 SICE
ྗֶม༰ʹ‫ͮ͘ج‬ೋ‫٭‬ϩϘοτͷࠨӈํ޲Ҡಈ੍‫ޚ‬
೤ా ༸࢙ ∗1 , ਿ‫ ݪ‬஌ಓ ∗2
Sideway Locomotion Control of Biped Robots Based on Dynamics Morphing
Hiroshi Atsuta∗1 and Tomomichi Sugihara∗2
∗1∗2
Department of Adaptive Machine Systems, Graduate school of engineering, Osaka University
2-1 Yamadaoka, Suita-shi, Osaka 565-0871, Japan
This paper presents a biped locomotion control to step sideways based on the framework
of the dynamics morphing. Since the proposed controller doesn’t require detailed referential
motion trajectories, it enables a robot to walk sideway at arbitrary velocity given at random
timing. Sideway locomotion is realized by alternating the velocity following control and
the self-excited oscillating control accompanying with the exchange of the supporting foot.
Techniques to automatically update the referencial position of COM and phase for the
consistent footstep are also proposed.
Key Words : Biped robot, Sideway locomotion, COM-ZMP regulator, Dynamics morphing
1. ͸
͡
Ί
(8)
ʹ
ͨɽ‫ཱʹط‬Ґ੍‫ ޚ‬ɼఆৗతͳ౿Έସ͑
ೋ‫٭‬ϩϘοτ͸ɼਓͷੜ‫ۭؒ׆‬΍ࡂ֐‫ݱ‬৔ͷΑ͏ͳ
Έग़͠
(11)
ɼલ‫ޙ‬าߦ
(12)
(10)
ɼᄉᅀͷ౿
Λ౷߹͢Δ੍‫ثޚ‬ΛఏҊͯ͠
ෳࡶͳ‫͍͓ͯʹڥ؀‬΋ߴ͍ҠಈೳྗΛൃ‫͢ش‬Δ͜ͱ͕
͍ΔɽຊߘͰ͸͜ΕΛ֦ு͠ɼཱҐ੍‫ޚ‬΍લ‫ޙ‬าߦͱ
‫ظ‬଴͞Ε͍ͯΔɽϩϘοτͷஔ͔Εͨঢ়‫گ‬΍஍‫ͭʹܗ‬
౷߹Մೳͳࠨӈํ޲Ҡಈ੍‫ޚ‬ΛఏҊ͢Δɽ
͍ͯࣄલʹे෼ͳ৘ใ͕ಘΒΕ͍ͯΕ͹ɼͦΕΒʹ‫ج‬
ࠨӈํ޲Ҡಈ͸લ‫ޙ‬าߦͷͨΊͷ੍‫ثޚ‬Λͦͷ··
͍ͮͯৄࡉͳࢀর‫ي‬ಓΛ࣌ࠁͷؔ਺ͱͯ͠‫ٻ‬Ίɼൺֱ
ద༻͢Δ͜ͱͰ࣮‫͢ݱ‬Δ͜ͱ͸Ͱ͖ͳ͍ɽͳͥͳΒલ
త༰қʹϩϘοτʹߴ౓ͳ࡞‫ۀ‬ΛߦΘͤΔ͜ͱ͕Ͱ͖
‫͋Ͱ޲ํޙ‬Ε͹ৗʹ໨ඪ଎౓ͷํ޲ʹ଍Λ౿Έग़͢͜
ɽ͔͠͠ෳࡶͳ‫͍͓ͯʹڥ؀‬࿏໘‫ܗ‬ঢ়΍‫͜ى‬Γ
ͱ͕Ͱ͖Δ͕ɼࠨӈํ޲Ͱ͸଍ͷަࠩΛ‫ݶ͍ͳ͞ڐ‬Γ
͏Δ֎ཚʹ͍ͭͯࣄલʹ஌Δ͜ͱ͸೉͘͠ɼෆ࣮֬͞
ৗʹ໨ඪ଎౓ͷํ޲ʹ଍Λ౿Έग़͢͜ͱ͸ෆՄೳ͔ͩ
ʹຬͪͨ‫ڥ؀‬தΛҠಈ͢Δํ๏ͱͯ͠͸‫ݶ‬ք͕͋Δɽ
ΒͰ͋Δɽ͕ͨͬͯ͠ɼࢦྩ͞Εͨ଎౓ํ޲ʹରͯ͠
(1) (2)
Δ
ࢀর‫ي‬ಓͷ‫ܭ‬ըΛϑΟʔυόοΫϧʔϓʹ૊ΈࠐΉ
Ұ౓‫ݮ‬଎ͨ͠‫ޙ‬ɼ࠶ͼՃ଎ͯ͠໨ඪ଎౓ʹ௥ैͤ͞ͳ
͜ͱͰ֎ཚ΍ࢦྩ஋ͷมߋʹରԠ͢Δํ๏͸ɼ͜Ε·
͚Ε͹ͳΒͳ͍ɽ͜ͷෳࡶͳ੍‫ޚ‬Λࢀর‫ي‬ಓͷࣄલ‫ܭ‬
(3)∼(5)
Ͱͷͱ͜Ζ੒ՌΛ͍͋͛ͯΔ
ɽ͔͜͠͠ΕΒͷํ
ըʹཔΒͣঢ়ଶϑΟʔυόοΫͷΈʹΑ࣮ͬͯ‫͢ݱ‬Δ
๏͸ɼ‫ݩ‬ͷࢀর‫ي‬ಓͱͷ͕ࠩͳΔ΂͘খ͘͞ͳΔΑ͏
͜ͱ͕ຊߘͷ໨తͰ͋Δɽ͜Ε͸ (i) ࢧ࣋ঢ়ଶʹ‫ͮج‬
ʹ‫ي‬ಓΛमਖ਼͢ΔͷͰɼࢀর‫ي‬ಓΛେ͖͘֎Εͳ͚Ε
ࣗ͘ྭৼಈ੍‫ͱثޚ‬଎౓௥ै੍‫ثޚ‬ͷ੾Γସ͑ɼ(ii)
͹ͳΒͳ͍Α͏ͳ૝ఆ֎ͷঢ়‫ʹگ‬ॊೈʹৼΔ෣͏͜ͱ
଎౓௥ै੍‫΁ޚ‬ͷม‫ʹܗ‬ಉ‫ͨ͠ظ‬ӡ଍੍‫ޚ‬ɼ(iii) ଎౓
͕Ͱ͖ͳ͍ɽ
௥ै੍‫͍͓ͯʹޚ‬΋ഁ୼͠ͳ͍଍ঢ੍߱‫ޚ‬ɼͷࡾऀʹ
ࢀর‫ي‬ಓʹཔΒͳ੍͍‫ํޚ‬๏΋ఏҊ͞Ε͍ͯΔ
(6) (7)
ΑͬͯՄೳͱͳΔɽ
͕଍Λ఺઀৮ͱ͢Δͱ੍͍ͬͨ໿͕͋Δɽ͞Βʹਂࠁ
2. ྗֶม༰ʹ‫ثޚ੍ͮ͘ج‬ઃ‫ܭ‬
ͳͷ͸ɼ͜ΕΒͷํ๏͸ࠓͷͱ͜ΖཱҐ੍‫ͱޚ‬ซ༻Ͱ
͖ͳ͍ɽSugihara
(8)∼(12)
͸ɼϩϘοτͷӡಈΛϩϘοτ
ࣗ਎ͷμΠφϛΫεͱ੍‫ثޚ‬ͷͳ͢ྗֶ‫ͯ͠ͱܥ‬ઃ
‫͠ܭ‬ɼྡ઀͢ΔӡಈؒͷભҠΛ੍‫ثޚ‬ͷ࿈ଓతม‫ʹܗ‬
Α࣮ͬͯ‫͢ݱ‬Δɼྗֶม༰ͱ͍͏࿮૊ΈΛఏҊ͖ͯ͠
∗1
∗2
େࡕେֶେֶӃ޻ֶ‫ڀݚ‬Պ஌ೳɾ‫ػ‬ೳ૑੒޻ֶઐ߈ʢ˟ 5650871 େࡕ෎ਧాࢢࢁాٰ 2-1ʣ[email protected]
ಉ্ [email protected]
ຊߘͰ͸ϩϘοτͷશ࣭ྔ͕ॏ৺ʹूதͨ͠Ϟσϧ
Λ༻͍ΔɽFig. 1 ͷΑ͏ʹɼϩϘοτͷਐߦํ޲ʹ x
࣠ɼࠨํ޲ʹ y ࣠ɼԖ௚্ํ޲ʹ z ࣠ΛͱΔɽ؆୯ͷ
ͨΊɼॏ৺·ΘΓʹൃੜ͢ΔτϧΫ͸ແࢹ͠ɼॏ৺ߴ
͞͸Ұఆͱ͢Δɽ͜ͷ؆୯Խ͸ఆੑతʹ͸಑ମ෦෼ʹ
ൺ΂ͯ‫٭‬෦෼͕े෼͚ܰΕ͹ଥ౰Ͱ͋Δɽॏ৺ҐஔΛ
p = [x y z]T ɼZMP ҐஔΛ p Z = [xZ yZ zZ ]T ͱ͢Δͱɼॏ
第 19 回ロボティクスシンポジア(2014 年 3 月 13 日 -14 日・兵庫)
- 49 -
(a)
(b)
(c)
(d)
Fig. 3 Position of equilbrium point
Fig. 1 COM-ZMP model
Ͱ͋Γɼd x, d y ͸ࢀরॏ৺Ґஔɼd vx ͸લ‫޲ํޙ‬ͷࢀ
৺ͷӡಈํఔࣜ͸࣍ࣜͷΑ͏ʹද͞ΕΔɽ
x¨ = ω 2 (x − xZ )
√
র଎౓ɼఆ਺ qx (≥ 0), qy (≥ 0), k(> 0), r(> 0) ͓Αͼ
(1)
y¨ = ω 2 (y − yZ )
(2)
g/zɼॏྗՃ଎౓ g = 9.8[m/s2 ] Ͱ͋Δɽ
͜Ε͸ɼZMP Ґஔ p Z Λ੍‫ޚ‬ೖྗͱͯ͠ɼॏ৺Ґஔ p
ͨͩ͠ɼω ≡
(13)
Λ੍‫ޚ‬ՄೳͰ͋Δ͜ͱΛ͍ࣔͯ͠Δ
ɽͨͩ͠ɼZMP
Ґஔ p Z ͸ࢧ࣋ྖҬ S ʹؔͯ࣍͠ͷ߆ଋ৚݅Λຬͨ͢
ඞཁ͕͋Δɽ
pZ ∈ S
(3)
͜ͷ߆ଋ৚݅Λຬͨͭͭ͠‫٭‬ҠಈΛ࣮‫͢ݱ‬ΔͨΊʹɼ
ϩϘοτ͸ࢧ࣋ྖҬͷൣғ಺Ͱ ZMP Λૢ࡞ͯ͠ॏ৺
Λ੍‫͢ޚ‬Δ͜ͱͱࢧ࣋ྖҬΛෆ࿈ଓʹม‫͢ܗ‬Δ͜ͱΛ
ཱ྆ͤ͞ͳ͚Ε͹ͳΒͳ͍ɽಛʹೋ‫٭‬ϩϘοτ͸ɼࢧ
࣋ྖҬͷม‫ܗ‬ΛࠨӈҰରͷ଍ͷ౿Έସ͑ʹΑͬͯߦ͏ɽ
Sugihara(12)͸ཱҐ੍‫ޚ‬ɼఆৗతͳ଍౿Έɼલ‫ޙ‬ਐา
ρ (≥ 0) ͸ઃ‫ܭ‬ύϥϝʔλͰ͋Δɽ·ͨࢧ࣋ྖҬ S ͸
x ํ޲ͷดྖҬ [xZmin , xZmax ] ͓Αͼ y ํ޲ͷดྖҬ
[yZmin , yZmax ] Ͱද͞ΕΔͱ؆୯Խ͍ͯ͠Δɽ͜ͷ੍‫ޚ‬
‫ࡍ࣮͕ͯͬͨ͠ʹث‬ͷ ZMP ͕໨ඪ ZMP ʹΑ͘௥ै
͢ΔͳΒ͹ɼx, y ํ޲ͷॏ৺ӡಈ͸ͦΕͧΕ࣍ͷ۠෼
ࣗ཯‫͏ैʹܥ‬ɽ



ω 2 x − ω 2 xZmax
(T1)


d
2
d
x¨ = −ω (qx + 1)(x˙ − vx ) − ω qx (x − x) (T2)(10)



ω 2 x − ω 2 x
(T3)
Zmin


ω 2 y − ω 2 yZmax
(S1)


2
d
y¨ = −ω (qy + 1) f (ζ )y˙ − ω qy (y − y) (S2) (11)



ω 2 y − ω 2 y
(S3)
Zmin
ࣜ (10) ͓Αͼ (11) ʹै͏ॏ৺ͷৼΔ෣͍Λ૬ਤΛ
༻͍ͯઆ໌͢Δɽ·ͣ໼ঢ়໘಺ͷӡಈɼ͢ͳΘͪ x ํ
ߦΛ౷߹͢Δ੍‫( ࣜͯ͠ͱثޚ‬1)(2) ͷ໨ඪ ZMP Ґஔ
޲ͷӡಈʹ͍ͭͯड़΂Δɽલ‫ʹ޲ํޙ‬Ҡಈͤ͞ΔͨΊ
xZ ͓Αͼ yZ Λ࣍ࣜͰఆΊΔ͜ͱΛఏҊͨ͠ɽ
)
(
x˙ − d vx
x˜Z = d x + (qx + 1) x − d x +
ω



x
(T1 : x˜Z > xZmax )

 Zmax
xZ = x˜Z
(T2 : xZmin ≤ x˜Z ≤ xZmax )



x
(T3 : x˜Z < xZmin )
Zmin
ʹ͸ॏ৺଎౓Λࢀর଎౓ d vx ʹऩଋͤ͞ΔΑ͏ͳ੍‫ޚ‬
)
y˙
y˜Z = y + (qy + 1) y − y + f (ζ )
ω


yZmax (S1 : y˜Z > yZmax )


yZ = y˜Z
(S2 : yZmin ≤ y˜Z ≤ yZmax )



y
(S3 : y˜ < y
)
d
Zmin
ͨͩ͠
(
Z
{
d
(4)
(5)
+ 1)2 ζ
(qy
f (ζ ) ≡ 1 − ρ exp k 1 −
r2
√
y˙2
ζ ≡ (y − d y)2 + 2
ω qy
(T2) ʹ͓͚Δॏ৺ӡಈ͸࣍ͷඍ෼ํఔࣜʹै͏ɽ
x¨ = −ω (x˙ − d vx )
(12)
͜ͷͱ͖ฏߧ఺͸ଘࡏͤͣɼx˙ ͸ d vx ʹऩଋ͠Α͏ͱ
͢Δɽd vx > 0 ͳΒ͹ঢ়ଶ͸͍ͣΕ T1 ʹɼd vx < 0 ͳ
(6)
Β͹ঢ়ଶ͸͍ͣΕ T3 ʹͦΕͧΕҠߦ͠ɼx˙ ͸ൃࢄ͢
√
ΔɽFig. 2(a) ͸ ω = g/0.27, xZmin = −0.4, xZmax =
(7)
ͷΑ͏ʹย଍ཱͪͷঢ়ଶͰ΋྆଍தԝΛฏߧ఺ͱ͢Δ
0.55, qx = 0 ʹର͢Δ૬ਤͰ͋ΔɽFig. 3(a) ʹࣔ͢Α͏
ʹ໼ঢ়໘಺ͷӡಈͰ͸ɼ྆଍ཱͪͷঢ়ଶͰ΋ Fig. 3(b)
͜ͱ͕ՄೳͰ͋Δɽͦ͜Ͱ x˙ ͷൃࢄΛ๷͗‫ܧ‬ଓతʹӡ
Zmin
2}
‫͕ث‬ඞཁͰ͋Δɽࣜ (10) ʹ͓͍ͯ qx = 0 ͷͱ͖ঢ়ଶ
(8)
ಈΛߦ͏ͨΊʹɼཱҐՄ҆ఆ৚݅ Λৗʹຬ଍͢ΔΑ
(12)
͏ͳӡ଍੍‫͕ޚ‬ఏҊ͞Ε͍ͯΔ
(8)
(9)
dv
x
ɽ Fig. 4 ʹɼࢦྩ஋
ʹΑཱͬͯҐ҆ఆԽ੍‫͔ثޚ‬Β଎౓௥ै੍‫ʹثޚ‬
࿈ଓతʹม༰͢Δ༷ࢠΛࣔ͢ɽ
࣍ʹલֹ໘಺ͷӡಈɼ͢ͳΘͪ y ํ޲ͷӡಈʹ
√
͍ͭͯड़΂ΔɽFig. 2(b) ͸ ω = g/0.27, yZmin =
- 50 -
S1
T1
T2
S1
S2
T3
(a) Velocity follower
S2
S3
S3
(b) COM-ZMP regulator
(c) Self-excited oscillator
Fig. 2 Solution curves of the piecewoise linear autonomous system
Fig. 4 Dynamics morphing along x axis
Fig. 5 Dynamics morphing along y axis
−0.7, yZmax = 0.7, qy = 0.5, d y = 0, k = 1, r = 0.5, ρ =
0 ʹର͢Δ૬ਤͰ͋Δɽͨͩ͠ɼԣ͕࣠ࠨӈํ޲ͷॏ
৺Ґஔ yɼॎ͕࣠ॏ৺଎౓ y˙ Ͱ͋Δɽρ = 0 ͷͱ͖ࣜ (8)
ΑΓ f (ζ ) = 1 (const.) ͱͳΓɼ͜Ε͸ Fig. 3(c) ͷΑ͏
ʹϩϘοτͷॏ৺ΛࢀরҐஔ d y ʹ઴ۙऩଋͤ͞Δ࠷
(8)
ྑॏ৺-ZMP ϨΪϡϨʔλ ʹҰக͢Δɽ
੍‫ثޚ‬Λॏ৺-ZMP ϨΪϡϨʔλʹͨ͠··Ͱ͸଍
౿ΈΛ͢Δ͜ͱ͸Ͱ͖ͳ͍ɽͳͥͳΒ Fig. 3(c) ͷΑ͏
ʹॏ৺Λ྆଍தԝʹ઴ۙ҆ఆͤͨ͞ঢ়ଶͰɼย଍Λ্
ZMP
Fig. 6
͛Δ͜ͱ͸Ͱ͖ͳ͍͔ΒͰ͋Δɽ͕ͨͬͯ͠ɼย଍Λ
ZMP
Difference of COM motion depending on
supporting foot
্͛Δʹ͸ Fig. 3(d) ͷΑ͏ʹࢧ࣋‫٭‬ଆʹॏ৺ΛҠಈ͞
͕Β΋ఆৗతʹ଍౿ΈΛߦ͏੍‫ʹط͕ޚ‬ఏҊ͞Ε͍ͯ
ͤΔඞཁ͕͋Γɼఆৗతʹ྆଍౿ΈΛ͓͜ͳ͏ͨΊʹ
Δ
͸ɼॏ৺ʹࣗྭৼಈΛੜͤ͡͞Δඞཁ͕͋Δɽࣜ (8)
૬ਤͷมԽͷ༷ࢠΛ Fig. 5 ʹࣔ͢ɽ
Ͱఆٛ͞ΕΔඇઢ‫ʹ߲ܗ‬Αͬͯɼρ >
e−1
(10)
ͷͱ͖ঢ়ଶ
(S2) ʹ͓͍ͯ࣍ͷϦϛοταΠΫϧΛൃ‫͢ݱ‬Δɽ
√
(1 + log k ρ r2 )
y˙2
(y − d y)2 + 2 =
(13)
ω qy
(qy + 1)2
r
2π
ɼप‫ √ ظ‬ͷࣗ
ಛʹɼρ = 1 ͷͱ͖ৼ෯
qy + 1 √
ω qy
ྭৼಈͱͳΔɽFig. 2(c) ͸ ω = g/0.27, yZmin =
−0.7, yZmax = 0.7, qy = 0.5, d y = 0, k = 1, r = 0.5, ρ =
1 ʹର͢Δ૬ਤͰ͋Δɽঢ়ଶ (S2) ʹ͓͍ͯϦϛοτα
ΠΫϧ͕‫ܗ‬੒͞Ε͍ͯΔɽఆৗঢ়ଶͰ͸ɼZMP ͸ৼ
෯ r Ͱॏ৺ͷࣗྭৼಈʹ஗Εͳ͠Ͱಉ‫͢ظ‬Δɽ͜ͷࣄ
࣮ʹ‫͍ͯͮج‬ɼZMP ͷ߆ଋ৚݅Λࣗಈతʹຬ଍͠ͳ
ɽͦͷͱ͖ʹࢧ࣋ྖҬ͕ෆ࿈ଓʹมԽ͢Δ༷ࢠͱ
3. ࠨӈํ޲ͷҠಈ੍‫ޚ‬ͷͨΊͷྗֶม༰
3·1 ଍ͷࢧ࣋ঢ়ଶʹΑΔ଎౓௥ै‫ৼྭࣗͱܥ‬ಈ‫ܥ‬
ͷ੾Γସ͑
ॏ৺ͷӡಈํఔࣜ (1)(2) ͸ x, y ʹͭ
͍ͯରশͰ͋Δ͕ɼࠨӈํ޲ҠಈͷͨΊʹɼલઅͰड़
΂ͨ x ํ޲ͷ଎౓௥ै੍‫ثޚ‬Λͦͷ·· y ํ޲ʹద༻
͢Δ͜ͱ͸Ͱ͖ͳ͍ɽͳͥͳΒલ‫͋Ͱ޲ํޙ‬Ε͹ৗʹ
໨ඪ଎౓ͷํ޲ʹ଍Λ౿Έग़͢͜ͱ͕Ͱ͖Δ͕ɼࠨӈ
ํ޲Ͱ͸ɼ଍ͷࢧ࣋ঢ়ଶʹΑͬͯ͸଍ͷަࠩΛ‫ͳ͞ڐ‬
͍‫ݶ‬Γৗʹ໨ඪ଎౓ͷํ޲ʹ଍Λ౿Έग़͢͜ͱ͕Ͱ͖
ͳ͍͔ΒͰ͋ΔɽFig. 6 ͷΑ͏ʹࢦྩ஋ͱͯ͠ d vy > 0
Λ༩͑ͨ৔߹ɼ͢ͳΘͪࠨํ޲ʹҠಈ͢Δ৔߹Λߟ͑
- 51 -
(8)
લ‫޲ํޙ‬ͷӡ଍͸ɼཱҐՄ҆ఆ৚݅ ʹ‫͍ͯͮج‬໨
ඪ౿Έग़͠Ґஔ d xS Λܾఆ͢Ε͹Α͔ͬͨɽ
d
xS = x +
x˙
ω
(16)
͔͠͠ࠨӈํ޲ʹ͍ͭͯ͸ɼॏ৺͸ৼಈ͍ͯ͠ΔͨΊɼ
ཱҐՄ҆ఆ৚݅ʹ‫ͮ͘ج‬໨ඪ౿Έग़͠Ґஔ΋·ͨৼಈ
ͯ͠͠·͏ɽͦ͜Ͱࠨӈํ޲ʹ͍ͭͯ͸ॏ৺ͷৼಈத
৺ɼͭ·ΓࢀরҐஔ d y ͕྆଍தԝʹҐஔ͢ΔΑ͏ʹ
Fig. 7
Dynamics morphing along y axis with a desired
velocity
ܾΊΔɽ͢ͳΘͪɼࠨӈํ޲ͷ໨ඪ౿Έग़͠Ґஔ d yK
Λ࣍ͷΑ͏ʹఆΊΔɽ
Α͏ɽ͜ͷͱ͖ɼࠨ଍͕཭চ͍ͯ͠ΔͳΒ͹଎౓௥ै
d
y K = 2 d y − yP
੍‫ʹثޚ‬ม‫͢ܗ‬Δ͜ͱͰࠨํ޲ʹҠಈ͢Δ͜ͱ͕Ͱ͖
ͨͩ͠ɼK ɼP ͸ͦΕͧΕ༡‫ࠨ͕٭‬଍ͷͱ͖ LɼRɼ༡
Δɽ͔͠͠ɼࠨ଍͕ணচ͍ͯ͠Δঢ়ଶͰґવ଎౓௥ै
‫͕٭‬ӈ଍ͷͱ͖ RɼL Ͱ͋Δɽ
(17)
੍‫͞ࢪ͕ޚ‬Εͨ··Ͱ͋Δͱస౗ͯ͠͠·͏ɽͦ͜Ͱɼ
͜ͷํ๏ʹΑͬͯ྆଍෯ΛมԽͤ͞Δ͜ͱʹͳΔ͕ɼ
ࠨ଍͕ணচ͍ͯ͠Δͱ͖͸ɼࣗྭৼಈ੍‫ʹثޚ‬Ҡߦ͠ɼ
ͦΕʹ൐ͬͯࣗྭৼಈͷৼ෯ r Λௐઅ͢Δඞཁ͕͋Δɽ
ॏ৺଎౓Λ‫ݮ‬଎ͤ͞Δඞཁ͕͋Δɽ͜ͷΑ͏ͳ੍‫ثޚ‬
͢ͳΘͪɼr ͸࣍ࣜʹΑͬͯௐઅ͢Δɽ
Λ࣮‫͢ݱ‬ΔͨΊʹ y ํ޲ͷ੍‫ثޚ‬Λࣜ (6) ͷ୅ΘΓʹ
r=
࣍ࣜͰஔ͖‫͑׵‬Δɽ
)
(
y˙ − d vy
y˜Z = d y + (qy + 1) y − d y + f (ζ )
ω
ͨͩ͠ɼ f (ζ ) ͸ࣜ (8)
Ͱఆٛ͞Εɼd vy
(14)
͸ y ํ޲ͷࢀ
র଎౓Λද͢ɽ͜Ε͸ ρ = 0 ͔ͭ qy = 0 ͱͨ͠ͱ͖
3·3 ଍ঢ੍߱‫ޚ‬
y
pZ ≡ yZ − d y −
z∗ ≡
ৼಈ੍‫ͳʹثޚ‬Δɽͭ·Γɼ଎౓௥ै੍‫ৼྭࣗͱثޚ‬
Fig. 7 ʹࢧ࣋‫٭‬ͷঢ়ଶʹΑͬͯ଎౓௥ै੍‫ࣗͱثޚ‬
ྭৼಈ੍‫͕ࣗثޚ‬ಈతʹ੾ΓସΘΔ༷ࢠΛࣔ͢ɽ
3·2 ࢀরॏ৺Ґஔͷߋ৽ͱӡ଍੍‫ޚ‬
଎౓௥ै
ͷͨΊཱҐ੍‫ʹޚ‬෮‫͢ؼ‬ΔࡍɼҎલͷࢀরҐஔ͕࢒ͬ
͍ͯͯ͸‫͋Ͱݥة‬Δɽ͜ͷΑ͏ͳໃ६͕ੜ͡ͳ͍Α͏
ʹલ‫޲ํޙ‬ͷҠಈ੍‫Ͱޚ‬͸ɼࢀরҐஔΛ‫ࡏݱ‬ͷॏ৺Ґ
ஔɼ͢ͳΘͪ d x = x ͱ͢Ε͹Α͔ͬͨɽ͔͠͠ࠨӈํ
޲ʹ͍ͭͯ͸ɼࢀরҐஔ d y ͸ॏ৺ͷࣗྭৼಈͷৼಈ
த৺ͳͷͰɼࢀরҐஔΛ‫ࡏݱ‬ͷॏ৺ҐஔͰߋ৽͢Δ͜
= 0 ͱͨ͠ͱ͖ʹ଎΍͔
ʹఀࢭɼ·ͨ͸ͦͷ৔Ͱ଍౿Έ͕ߦ͑ΔΑ͏ɼॏ৺଎
౓ y˙ ͷਖ਼ෛ͕ d vy ͷਖ਼ෛͱҰக͢Δͱ͖ʹ
√
√
1 + log k ρ
d
r
y = y±
qy + 1
(19)
1 h|pZ |
σ (ρ¯ )(1 − cos 2πϕ∗ )
2 r
(20)
ͨͩ͠ ∗ ʹ͸ L ·ͨ͸ R ͕ೖΔɽϕL ɼϕR ͸ pZ ͕࣠
଍઀஍ྖҬ಺ʹ͋Δͱ͖ʹஞ࣍ਪఆ͞Εͨ૬ରҐஔ͔
(10)
·ΕͣɼϩϘοτ͸ࢀরҐஔͱແؔ܎ʹҠಈ͢Δɽͦ
y
(q¯y + 1)(y˙ − d vy )
i
√
ω q¯y
Βఆٛ͞ΕΔҐ૬Ͱ͋Δɽৄࡉ͸จ‫ݙ‬
੍‫͍͓ͯʹޚ‬͸ࢀরҐஔ d x ͋Δ͍͸ d y ੍͕‫ؚʹثޚ‬
ͱ͸Ͱ͖ͳ͍ɽͦ͜Ͱɼd v
଍ঢ੍߱‫ޚ‬Λઆ໌͢Δɽy ํ޲
͓Αͼ଍্͛ߴ͞ z∗ ΛͦΕͧΕ࣍ͷΑ͏ʹఆٛ͢Δɽ
= 0 ͔ͭ ρ = 1 ͱͨ͠ͱ͖͸ࣜ (6) ʹҰகͯࣗ͠ྭ
ಈ੍‫ثޚ‬Λ੾Γସ͑ΒΕΔΑ͏ͳ੍‫ͳͱثޚ‬Δɽ
(18)
ͷ ZMP Ґஔͱॏ৺଎౓ʹΑͬͯఆٛ͞ΕΔෳૉ਺ pZ
͸ࣜ (4) ͱಉ͡ߏ଄Λͱͬͯ଎౓௥ै੍‫ͳʹثޚ‬Γɼ
dv
y L − yR
2
Λࢀর͞Εͨ
͍ɽh ͸ఆ਺Ͱ͋Γɼσ (ρ¯ ) ͸࣍ͷΑ͏ʹఆٛ͞ΕΔɽ



1
(ρ¯ > 1)



−1
ρ¯ − e
(21)
σ (ρ¯ ) ≡
(e−1 ≤ ρ¯ ≤ 1)
 1 − e−1



0
(0 ≤ ρ¯ < e−1 )
·ͨɼࣜ (19) ͓Αͼ (21) ͷ q¯y ͓Αͼ ρ¯ ͸ͦΕͧΕ࣍
ͷΑ͏ʹఆٛ͞ΕΔɽ

q (q ̸= 0)
y
y
,
q¯y ≡
′
q (q = 0)
y
y

ρ
ρ¯ ≡
ρ ′
(ρ ̸= 0)
(ρ = 0)
(22)
3.1 અͰઆ໌ͨ͠Α͏ʹࢀর଎౓Λ༩͑ͯ଎౓௥ै
(15)
ͱ͢Δɽͨͩ͠ෳ߹͸ y˙ ͕ਖ਼ͷͱ͖ +ɼy˙ ͕ෛͷͱ͖ʹ
− ͱ͢Δɽࣜ (15) ͷӈลୈ 2 ߲͸ॏ৺ͷৼ෯Λද͢ɽ
੍‫੾ʹثޚ‬Γସ͑Δ৔߹ɼqy = 0, ρ = 0 ͱ͢Ε͹Α͍
′
͕ɼͦͷࡍͰ΋଍ͷ‫ڍ‬ಈ͕੔߹͢ΔΑ͏ʹ qy ͓Αͼ
′
ρ ͸ͦΕͧΕ଎౓௥ै੍‫੾ʹثޚ‬Γସ͑Δ௚લͷ qy
͓Αͼ ρ ͷϦβʔϒ஋ͱ͢Δɽ
- 52 -
Fig. 8 Snapshots of the movie of the simulation of the locomotion along y axis
4. γϛϡϨʔγϣϯ
ΔͨΊͰ͋ΔɽલઅͰड़΂੍ͨ‫ߏ͕ثޚ‬੒͢Δྗֶ‫ܥ‬
ਓ‫ܕ‬ϩϘοτ mighty Λ૝ఆ͠ɼγϛϡϨʔγϣϯ
Λߦͬͨɽ؆୯ͷͨΊɼશ࣭ྔ͕ॏ৺ʹूத͍ͯ͠Δ
ϞσϧΛ༻͍ɼॏ৺ߴ͞Λ z = 0.26[m]ɼͦͷଞͷύϥ
ϝʔλΛ h = 0.02[m]ɼk = 1 ͱͨ͠ɽ྆଍ͱ΋લํʹ
0.055[m]ɼ‫ ʹํޙ‬0.04[m]ɼࠨӈʹͦΕͧΕ 0.035[m]
ͷ෯Λ΋ͭ௕ํ‫Ͱܗ‬ද‫͞ݱ‬Ε͍ͯΔɽॳ‫ظ‬ঢ়ଶʹ͓͍
ͯ྆଍ؒͷ‫཭ڑ‬͸ 0.084[m] ͱͨ͠ɽ·֤ͨ଍͕༡‫٭‬
ͱͳΔؒɼ໨ඪ౿Έग़͠Ґஔʹ௥ै͢ΔΑ͏ͳೋ࣍஗
Ε੍‫ثޚ‬Λઃ‫ͨ͠ܭ‬ɽ
x¨∗ = K(d xS − x∗ ) −Cx˙∗
y¨∗ = K(d yK − y∗ ) −Cy˙∗
ʹ͓͍ͯɼॏ৺͸ϦϛοταΠΫϧͱࢦྩ஋ͷେ͖͞
ͷؔ܎ʹΑͬͯ૬ਤ্Ͱ Fig. 10 ͷΑ͏ͳ‫ي‬ಓΛඳ͘ɽ
Fig. 10(a) ͷΑ͏ʹࢦྩ஋ d vy ͕ϦϛοταΠΫϧΑ
Γେ͖͍৔߹ɼ੍‫ثޚ‬͸ॏ৺͕଎౓௥ै‫͋ʹܥ‬Δͱ͖
͸ࢦྩ஋ʹऩଋͤ͞Α͏ͱ͢Δ͕ɼࣗྭৼಈ‫੾ʹܥ‬Γ
ସΘΔͱϦϛοταΠΫϧʹҾ͖ࠐ΋͏ͱ͢ΔͨΊɼ
࣮ࡍͷॏ৺଎౓͸ࢦྩ஋ʹ౸ୡ͢Δલʹ଎౓Λ൓స͞
ͤͯ͠·͏ɽҰํɼFig. 10(b) ͷΑ͏ʹࢦྩ஋ d vy ͕Ϧ
ϛοταΠΫϧΑΓখ͍͞৔߹΋ಉ༷ʹɼ੍‫ثޚ‬͸ॏ
(23)
(24)
৺͕଎౓௥ै‫͋ʹܥ‬Δͱ͖ʹࢦྩ஋ʹऩଋͤ͞Α͏ͱ
͢Δ͕ɼࣗྭৼಈ‫੾ʹܥ‬ΓସΘΔͱϦϛοταΠΫϧ
ʹҾ͖ࠐ·ΕΔͨΊɼ࣮ࡍͷॏ৺଎౓͸ࢦྩ஋ΑΓ΋
ͨͩ͠ ∗ ʹ͸ L ·ͨ͸ R ͕ೖΔɽK = 3000, C = 50 ͱ
େ͖ͳ଎౓Λग़ͯ͠͠·͏ɽ
ඪ ZMP ͷࢧ࣋ྖҬ্࠷ۙ๣఺ΛͱΔ͜ͱͰ୅ସͨ͠ɽ
ͭʹͳΔ఺͕ଟ਺‫ݟ‬ΒΕΔ͕ɼ͜Ε͸ Fig. 10 ͷ‫ؙ‬ҹ
ඍ෼ํఔࣜͷ਺஋ղ๏ʹ 4 ࣍ͷ Runge-Kutta ๏Λ༻͍ɼ
Ͱࣔ͢Α͏ͳ଎౓௥ै੍‫͔ثޚ‬Βࣗྭৼಈ੍‫੾ʹثޚ‬
ੵ෼࣌ؒ෯͸ 0.01[s] ͱͨ͠ɽ
Γସ͑ΔॠؒʹɼϦϛοταΠΫϧʹҾ͖ࠐ΋͏ͱ͠
ͨ͠ɽࣜ (5) (7) ʹ͓͚Δ؆୯Խ͞Εͨ๞࿨ଇ͸ɼ໨
ຊߘͰड़΂੍ͨ‫ثޚ‬ΛγϛϡϨʔγϣϯ্Ͱ࣮૷͠
·ͨɼFig. 9(b) ʹ͓͍ͯ ZMP Ґஔ yZ ͷ‫ي‬ಓ͕͍ͼ
ͯ‫ʹܹٸ‬଎౓ΛมԽͤ͞ΔͨΊͰ͋Δɽ
ͨɽͨͩ͠ϩϘοτͷঢ়ଶʹΑͬͯࢦྩ஋ͱ͸ؔ܎ͳ
5. ͓
੍͘‫ޚ‬ύϥϝʔλΛ‫੍ڧ‬తʹมߋ͢Δ͜ͱ͕͋ΔͷͰɼ
Θ
Γ
ʹ
ࢦྩ஋͸੍‫ޚ‬ύϥϝʔλͱ͸ผʹอ͓͖࣋ͯ͠ɼ੍‫ޚ‬
࣌ࠁͷؔ਺ͱͯ͠ఆٛͨ͠ࢀর‫ي‬ಓΛ‫ܭ‬ըͤͣʹཱ
‫͕ث‬ඞཁͳλΠϛϯάͰࢦྩ஋ͷ஋Λ੍‫ޚ‬ύϥϝʔλ
Ґ੍‫͔ޚ‬Βࠨӈํ޲ͷҠಈ͕ߦ͑Δ੍‫ثޚ‬Λ։ൃ͠
ͱͯ͠൓өͤ͞ΔΑ͏ʹͨ͠ɽ
ͨɽࠨӈํ޲ͷҠಈ͸଍ͷࢧ࣋ঢ়ଶʹΑͬͯॏ৺଎౓
˙ y, y)
˙ = (0.0409, 0, 0, 0)ɼqx = 0.5ɼ
ॳ‫ظ‬ঢ়ଶΛ (x, x,
dv
y
ΛཹΊΔඞཁ͕͋Δ͕ɼ଎౓௥ै੍‫ৼྭࣗͱثޚ‬ಈ੍
qy = 0.5ɼρ = 0 ͱ͠ɼࢦྩ஋
Λ −3.0 ∼ 3.0[m/s]
ͷൣғͰແ࡞ҝʹมߋͨ͠ɽಈ࡞ͷ࠷‫͍͓ͯʹޙ‬͸
‫ثޚ‬Λ੾Γସ͑Δ͜ͱʹΑΓ͜Ε͕Մೳͱͳͬͨɽ೚
dv
= 0[m/s] ͱ͠ɼཱҐ੍‫ʹޚ‬໭ͨ͠ɽಈըεφοϓ
γϣοτΛ Fig. 8 ʹɼॏ৺ࢀর଎౓ɾॏ৺଎౓Λϓϩο
ߦ͑ΔΑ͏ʹͳͬͨɽࠨӈํ޲ࢀর଎౓͚ͩͰͳ͘લ
τͨ͠άϥϑΛ Fig. 9(a)ɼॏ৺ҐஔɾZMP Ґஔɾ଍ઌ
ટճΛ‫ؚ‬Ίͨશํ޲Ҡಈʹ֦ு͢Δͷ͕ࠓ‫ޙ‬ͷ՝୊Ͱ
Ґஔɾࢧ࣋ྖҬΛϓϩοτͨ͠άϥϑΛಉਤ (b) ʹͦ
͋Δɽ
y
ҙͷλΠϛϯάͰ໨ඪ଎౓Λ༩͑ͯ΋ഁ୼ͤͣӡಈ͕
‫ࢀ޲ํޙ‬র଎౓΋༩͑Δ͜ͱͰࣼΊาߦ΋ՄೳʹͳΔɽ
ࢀ ߟ จ ‫ݙ‬
ΕͧΕࣔ͢ɽ
Fig. 9(a) ʹࣔ͢Α͏ʹࢦྩ஋ d vy Λແ࡞ҝʹมԽ͞
ͤͨʹ΋ؔΘΒͣɼಉਤ (b) ͷΑ͏ʹస౗͢Δ͜ͱͳ͘
‫ܧ‬ଓతʹӡಈΛߦ͍͑ͯΔ͜ͱ͕֬ೝͰ͖ΔɽFig. 9(a)
ʹ͓͍ͯॏ৺଎౓ y˙ ͕ࢦྩ஋ d vy ʹൺ΂খ͘͞ͳΔ఺
ͱେ͖͘ͳΔ఺͕͋Δ͕ɼ͜Ε͸ࣗྭৼಈ੍‫ࢦ͕ثޚ‬
ྩ஋ d vy ͱແؔ܎ʹϦϛοταΠΫϧ (13) Λ‫ܗ‬੒͢
(1) A. Takanishi, Y. Egusa, M. Tochizawa, T. Takeya,
and I. Kato. Realization of Dynamic Walking
Stabilized with Trunk Motion. In ROMANSY 7, pp.
68–79, 1988.
(2) K. Hirai, M. Hirose, Y. Haikawa, and T. Takenaka.
The development of Honda humanoid robot.
In Proceedings of the 1998 IEEE International
- 53 -
reference
velocity
[m/s]
COM
limit cycle
[s]
switching point of
the controllers
(a) d vy , y˙
[m]
supporting region
COM
(a) The case which d vy is large
left foot
limit cycle
reference
velocity
ZMP
right foot
switching point of
the controllers
[s]
(b) y, yZ , yL , yR
(b) The case which d vy is small
Fig. 9 Results of the simulation
(3)
(4)
(5)
(6)
(7)
(8)
Conference on Robotics and Automation, Vol. 2, pp.
1321–1326, 1998.
S. Kajita, M. Morisawa, K. Harada, K. Kaneko,
F. Kanehiro, K. Fujiwara, and H. Hirukawa. Biped
Walking Pattern Generator allowing Auxiliary
ZMP Control. In Proceedings of the 2006 IEEE
International Conference on Intelligent Robots and
Systems, pp. 2993–2999, 2006.
K. Nishiwaki and S. Kagami. Online Walking Control System for Humanoids with Short Cycle Pattern Generation. International Journal of Robotics
Research, Vol. 28, No. 6, pp. 729–742, 2009.
A. Herdt, N. Perrin, and P. B Wieber. Walking
without thinking about it. In Proceedings of
the 2010 IEEE/RSJ International Conference on
Intelligent Robots and Systems, pp. 190–195, 2010.
S.-H. Hyon and T. Emura. Symmetric Walking
Control: Invariance and Global Stability. In
Proceedings of the 2005 IEEE International
Conference on Robotics & Automation, pp. 1455–
1462, 2005.
C. Chevallereau, J.W. Grizzle, and Ching-Long
Shih. Asymptotically Stable Walking of a FiveLink Underactuated 3-D Bipedal Robot. IEEE
Transactions on Robotics, Vol. 25, No. 1, pp. 37–
50, 2009.
T. Sugihara. Standing stabilizability and stepping
maneuver in planar bipedalism based on the best
Fig. 10 Loci of COM in sideway locomotion
(9)
(10)
(11)
(12)
(13)
- 54 -
COM-ZMP regulator. In Proceedings of the 2009
IEEE International Conference on Robotics &
Automation, pp. 1966–1971, 2009.
T. Sugihara. Dynamics morphing from regulator
to oscillator on bipedal control. In Proceedings of
the 2009 IEEE/RSJ International Conference on
Intelligent Robots and Systems, pp. 2940–2945,
2009.
T. Sugihara. Consistent biped step control with
COM-ZMP oscillation based on successive phase
estimation in dynamics morphing. In Proceedings
of the 2010 IEEE International Conference on
Robotics & Automation, pp. 4224–4229, 2010.
T. Sugihara. Reflexive Step-out Control Superposed
on Standing Stabilization of Biped Robots. In
Proceedings of the 2012 IEEE-RAS International
Conference on Humanoid Robots, pp. 741–746,
2012.
T. Sugihara. Biped control to follow arbitrary
referential longitudinal velocity based on dynamics
morphing. In Proceedings of the 2012 IEEE/RSJ
International Conference on Intelligent Robots and
Systems, pp. 1892–1897, 2012.
K. Mitobe, G. Capi, and Y. Nasu. Control of
walking robots based on manipulation of the zero
moment point. Robotica, Vol. 18, pp. 651–657,
2000.