㐼㞴⪅᥈⣴䛾䛯䜑䛾⮬ᚊᆺ㣕⾜⯪䝻䝪䝑䝖䛾䝥䝻䝖䝍䜲䝥㛤Ⓨ㻌
ࣁࢺࣈࢻࣛ
࣭ࣅ࢝ࢫ ࣒ࣛࢧࣝ
2࣭ඵᮌᗣ
࣭ᑠᓥ⠊Ꮚ
࣭ᯇᮏ┤ᩥ
Ꮫ㝔ᕤᏛ◊✲⛉ಟኈㄢ⛬ሗ࣭⏕⏘ᕤᏛᑓᨷ 㸦 ಟ࣭⌧ᅾ࣎ࢫࢺࣥ␃Ꮫ୰㸧 Ꮫ㝔ᕤᏛ◊✲⛉༤ኈᚋᮇㄢ⛬ሗ࣭⏕⏘ᕤᏛᑓᨷ Ꮫ㝔ᕤᏛ◊✲⛉ಟኈㄢ⛬ሗ࣭⏕⏘ᕤᏛᑓᨷ ሗࢩࢫࢸ࣒ࢹࢨࣥᏛ⣔ ᩍᤵ
3URWRW\SH'HYHORSPHQWRI$XWRQRPRXV$LUVKLS5RERWIRU6HDUFKLQJ9LFWLPV
A. Aihaitijiang, B. Lamsal, Y. Yagi, N. Kojima and N. Matsumoto
Abstract
In this paper, we describe the prototype development of an autonomous airship. It is intended to explore the victims in a wide area, such as deserts. At first, we derive individual mathematical models divided into upward movement and rectilinear movement of the airship. Then, we perform system identification for the airship models. Next, we describe design and simulation of a PID controller for hovering and an optimal regulator for forward motion using MATLAB. By utilizing Arduino to implement these control systems, we got the prospect of autonomous area scan, compering between simulation and experiment results. Finally, we show the feasibility of searching victims by adding an image detection function, using Web camera at the tip of the airship.
Keywords:Victim search, airship robot, system identification, simulation, prototype development 㸯㸬ࡣࡌࡵ ࣟ࣎ࢵࢺᢏ⾡ࡢ㛤Ⓨࡣே㛫ࣟ࣎ࢵࢺࡢඹᏑࡍ ࡿ♫ྥࡅ࡚άⓎ⾜ࢃࢀ㸪ᵝࠎ࡞ᛂ⏝ࡀ◊ ✲ࡉࢀ࡚࠸ࡿ>@ࠋ࠼ࡤ㸪◁₍࡞ࡢᗈ࠸㡿ᇦ࡛㐼 㞴⪅ࡸ⿕⅏⪅ࢆ᥈⣴ࡍࡿࢣ࣮ࢫ࡛ࡣ㸪ேᡭ㢗ࡗ ࡓ᥈⣴࡛ࡣḟ㐼㞴ࡸḟ⅏ᐖࡢ༴㝤ᛶࡶ࠶ࡾ>@㸪 ࡇࢀࡽࡢ༴㝤ᛶࢆ㍍ῶࡋ࡚㸪㐼㞴⪅࡞ࢆ࠸ࡕ᪩ ࡃⓎぢࡍࡿࡓࡵ㸪㣕⾜ᆺࣟ࣎ࢵࢺᢏ⾡⏝ࡢ᭷ ຠᛶࡀᣦࡉࢀ࡚࠸ࡿ>@ࠋ ➹⪅ࡢ୍ேࡣ୰ᅜ㸪᪂࢘ࢢࣝ⮬༊༡㒊㸪 ࢱ࣒ࣜ┅ᆅࡢ୰ኸ࠶ࡿࢱࢡ࣐࡛ࣛ࢝ࣥ㸪 ᖺ ◁₍ࡢ㏆ࡃࡢ⾤ࡢᏊ౪ࡀ⾜᪉᫂࡞ࡗࡓ㝿㸪 ࣃࣛࢢࣛࢲ࣮ࡽ┠どࡼࡾ᥈⣴ࡍࡿࢆ㢗 ࡲࢀࡓ⤒㦂ࡀ࠶ࡿࠋᏊ౪ࡣᨾࡀⓎ⏕ࡋࡓ ࣨ᭶ ᚋᆅୖ࡛㌷㝲ࡼࡾṚయ࡛Ⓨぢࡉࢀࡓࠋᗈ࠸㡿 ᇦࡢ◁₍ࡢ୰࡛㸪㐼㞴⪅ࢆ᥈ࡍࡣࢥࢫࢺ㛫 ࡀࡾ㸪ᩆຓ⪅ࡀḟ㐼㞴ྜ࠺ྍ⬟ᛶࡶ࠶ࡗ ࡓࠋࡇ࠺ࡋࡓࢣ࣮ࢫ࡛㸪㣕⾜ᆺࣟ࣎ࢵࢺࢆά⏝࡛ ࡁࢀࡤ㸪ᆅ⾲㠃ࡢ㸰ḟඖⓗ࡞᥈⣴άືࡶ㸪
Fig. 1: Conceptual of the Airship with search function ✵୰ࡽ ᥈⣴άືࢆຍ࠼ࡿࡇ࡛Ᏻ ࡘ▷㛫ࡢ㐼㞴⪅᥈⣴ࡢྍ⬟ᛶࡀ୍ᒙ㧗ࡲࡿ⪃ ࠼ࡽࢀࡿ>@ࠋ ⅏ᐖࡢ⿕⅏⪅᥈⣴ᙜࡓࡿࢩࢫࢸ࣒ࡢ◊✲>@ ➼ࡀ࠶ࡿࡀ㸪⏬ീሗࢆᇶ࣮࢜࣌ࣞࢱࡀไᚚ ᅾࡋ࡚࠸࡚⮬ᚊᆺ࡛ࡣ࡞࠸ࠋ ࡇࡇ࡛ࡣ㸪㣕⾜⯪ࣟ࣎ࢵࢺࢆ⮬సࡋ㸪⏬ീㄆ㆑ᶵ ⬟ࢆ⏝ࡋࡓ㸪᪂ࡋ࠸᥈⣴ࣜࢦࣜࢬ࣒ࢆ⪃࠼㸪ࡑ ࢀ㐺ᛂࡍࡿ㣕⾜⯪ࡢᩘᏛࣔࢹࣝࢆᑟฟࡋ㸪ᩘᏛࣔ ࢹࣝࢆ⥺ᙧࡋ㸪ྍไᚚᛶࢆุูࡋ࡚タィᐇࢆ の三次元的な
ヨࡳࡓࠋᮏ✏ࡣ㸪ୖ㏙ࡢࡼ࠺࡞ࢣ࣮ࢫ㸪ไᚚჾࡢタ ࡛㸪㐼㞴⪅ࡸᑐ㇟≀᥈⣴ࢆ࡛ࡁࡿࡔࡅࢫ࣒࣮ࢬ ⾜࠺ࢩࢫࢸ࣒㛤Ⓨࡢࡓࡵࡢึᮇẁ㝵ࡢ◊✲ሗ࿌࡛ ࠶ࡾ㸪⏬ീㄆ㆑ᶵ⬟ࡼࡾ᥈⣴ືస☜ㄆࢆ┠ᣦࡋ ࡓࡶࡢ࡛࠶ࡿ㸦Fig.1 ཧ↷㸧ࠋ ᮏ✏࡛ࡣ➨ ❶࡛㸪◊✲ᑐ㇟ࡋ࡚㣕⾜⯪ࡢඛ ➃ :HE ࣓࢝ࣛࢆྲྀࡾࡅ㸪⏬ീㄆ㆑ࡼࡾ≉ᐃ ࡢ࣐࣮࢝ࢆ᥈⣴ࡍࡿࡓࡵࡢ㣕⾜⯪ࡢᵓᡂࡑࢀࢆ ⏝࠸ࡓ᥈⣴ࣝࢦࣜࢬ࣒ࡘ࠸࡚㏙ࡿࠋ➨ ❶ ࡛ࡣ㸪㣕⾜⯪ࡢ≧ែ✵㛫ࣔࢹࣝᑟฟࢆᆶ┤᪉ྥࡢ 㐠ືỈᖹ᪉ྥࡢ㐠ືศࡅ࡚⾜ࡗࡓࠋ≉㸪Ỉ ᖹ᪉ྥࡢ㐠ື㛵ࡋ࡚ࡣ㸪ྍไᚚᛶࢆ☜ಖࡍࡿࡓ ࡵ㸪┠ᶆ್ࡢࡎࢀࢆ⪃࠼ࡓࣔࢹࣝࡢኚࢆ⾜ ࡗ࡚࠸ࡿࠋࡉࡽ㸪ᑐ㇟ࡢ㣕⾜⯪ࡘ࠸࡚㸪ࣉࣟ ࣌ࣛ᥎ຊࡢ ᐃࡸᦶ᧿ಀᩘࡢᑟฟࢆྵࡵࡓࢩࢫࢸ ࣒ྠᐃࡘ࠸࡚㏙㸪ࡉࡽ㧗ᗘ⥔ᣢỈᖹ┤㐍 㐠ືࡢ PID ୪ࡧ LQR ไᚚ⣔タィࢩ࣑࣮ࣗࣞࢩ ࣙࣥ⤖ᯝࡘ࠸࡚㏙ࡿࠋ➨ ❶࡛ࡣ㸪AR ࣐࣮࢝ ࡢ᥈⣴ᐇ㦂ࡘ࠸࡚㏙ࡿࠋ 㸰㸬◊✲ᑐ㇟ࡢタᐃ 㣕⾜⯪ᆺࣟ࣎ࢵࢺࢆ㑅ࢇࡔ⌮⏤ ◁₍࡛᭱ࡶ㏿ࡃᗈ࠸⠊ᅖ࡛┠ᶆࢆほ ࡋࠊ㏿࠸ ࢫࣆ࣮ࢻ࡛⛣ືࡉࡏࡿᚲせࡀ࠶ࡿࠋ㣕⾜ࣟ࣎ࢵࢺ ࡣᩚᆅࡉࢀ࡚࡞࠸ሙᡤࡸࠊ㞀ᐖ≀ࡀከ࠸ሙᡤ࡛ࡶ ⛣ືྍ⬟㸦୕ḟඖ⛣ືྍ⬟㸧ࠋ㧗࠸⨨ࡽほ ࡍ ࡿࡇ࡛ࠊ୍ᗘᵝࠎ࡞ሗࢆྲྀᚓࡍࡿࡇࡀ࡛ ࡁࡿⅬࡀ࠶ࡿࠋ㣕⾜⯪ᆺࣟ࣎ࢵࢺࡣࡢ㣕⾜≀ ẚࡿࡼࡾ㛗㛫㣕⾜࡛ࡁ㸪ሙྜࡼࡗ୍࡚ ᐃࡢሙᡤ࡛࣍ࣂࣜࣥࢢࡀྍ⬟࡛࠶ࡾ㸪✵୰ほ ౽࡞㣕⾜ᶵ࡛࠶ࡿࠋ୍ᐃࡢ㧗ᗘࡽ୍ᗘ࣓࢝ ࣛࡽࡗࡓ⏬ീࡼࡗ࡚ᵝࠎ࡞ሗࢆྲྀࢀࡿ Ⅼࡽᗈ࠸⠊ᅖ࡛ࡢᑐ㇟≀ࡢ᥈⣴ࡣࡼࡾ㏿࠸࡛࠶ ࡿࠋ㣕⾜⯪ࡢᙺࡋ࡚ࡣ✵୰ࡽ⏬ീฎ⌮ࢆ⾜ ࠸㸪㐼㞴⪅ࡸᑐ㇟≀ࢆ᥈⣴ࡋᗈ࠸㡿ᇦ࡛ࡼࡾ㎿ ㏿࡞ᩆຓάືࡀᮃࡲࢀࡿࠋ ࢭࣥࢧሗࡋ࡚⏬ീㄆ㆑ὀ┠ࡋࡓ⌮⏤ እ⏺ࢆㄆ㆑ࡍࡿࢭࣥࢧࡋ࡚ࡢ࣓࢝ࣛ⏬ീࡣ㸪 ࣟ࣎ࢵࢺࡢ┠ࡢᙺࢆᯝࡓࡋ࡚㸪⏬ീࡽ࿘㎶ࡢ ᵝࠎ࡞ሗࢆྲྀᚓࡍࡿࡇࡀ࡛ࡁࡿࠋ⏬ീࢆฎ⌮ ࡍࡿࡇࡼࡗ࡚≉ᐃࣃࢱ࣮ࣥ㸪㢦㸪ேయ࡞ࢆ ㄆ㆑ࡋ㸪ࣟ࣎ࢵࢺࢆ⮬ᚊไᚚࡍࡿࡓࡵ⏝࡛ ࡁࡿࠋ࣓࢝ࣛࡣࡢࢭࣥࢧẚ㸪ẚ㍑ⓗᏳ౯࡛ ࠶ࡾ㸪ࡲࡓ㍍㔞ࡢࡶࡢࡶ࠶ࡾ㸪㠀ᖖ㌟㏆࡞ࢭࣥ ࢧ࡞ࡗ࡚࠸ࡿࠋࣇ࣮࣒ࣞࡢ୰ከࡃࡢሗࢆྵ ࡴ࠸࠺Ⅼࡀ࠶ࡿࠋ Fig. 2: Controller and drive unit of the Airship. ࡇࡢࡼ࠺࡞⌮⏤ࡼࡾ㸪ᮏㄽᩥ࡛ࡣ㸪AR ࣐࣮࢝ ࢆᑐ㇟ࡋ࡚⏬ീㄆ㆑ࡼࡿࣟ࣎ࢵࢺ᥈⣴ࢩࢫࢸ ࣒ࡢ᳨ウࢆ⾜ࡗࡓࠋ ᚋ㸪AR ࣐࣮࢝ࢆ࣐࣮࣮࢝ࣞࢫ AR ௦࠼ࢀࡤ㸪 ᑐ㇟≀ࡢࣃࢱ࣮࣐ࣥࢵࢳࣥࢢ࣭㑇㊧ࡢ᳨ฟ➼⧅ ࡀࡿ⪃࠼ࡿࠋ⮬ᚊⓗ㐼㞴⪅ࢆ᥈⣴ࡍࡿࣟ࣎ࢵ ࢺࡣ㸪ᑐ㇟≀ࡀே㛫࠺ࢆ㆑ูࡍࡿᚲせࡀ ࠶ࡾ㸪㢦ㄆ㆑ࡲࡓࡣ Kinect ࢭࣥࢧࢆ⏝࠸ࡓேయㄆ ㆑㸪ࣁࣥࢻࢪ࢙ࢫࢳࣕㄆ㆑➼ࡢ⏝ࡀ⪃࠼ࡽࢀࡿࠋ 㣕⾜⯪ࡢせồᵝ㸪タィヨస ⏬ീㄆ㆑ᶵ⬟ࢆᣢࡓࡏࡓ㣕⾜⯪ࡢணࡋࡓ࣌ ࣮ࣟࢻࡢ㔜㔞ࡣ J ࡛࠶ࡗࡓࠋࡇࡢ࣮࣌ࣟࢻ ࢆᦚ㍕ࡍࡿࡓࡵ㸪Ẽᄞࡢయ✚㠃✚ࢆィ⟬ࡋ㸪 タィᵝࢆỴࡵࡓࠋ⯪యࡢഃ㠃ᙧ≧ࡣ 7KH13//RZ 'UDJ$LUVKLS%RG\6KDSH>@ࢆཧ⪃ࡋࡓࠋ ࣭㛗㸸 >㹫@ ࣭㛗༙ᚄ㸦๓㒊㸧㸸>㹫@ ࣭┤ᚄ㸸>㹫@ ࣭య✚㸸>ੑ@ ࣭⾲㠃㠃✚㸸>੍@ ࣭࣒࣊ࣜ࢘࢞ࢫ㸸>/@ タィᵝࡋࡓࡀ࠸㸪⯪యࡢᆺ⣬ࢆసࡾ࣑ࣝ ࣒ࣝࢆ୕ᯛษࡾฟࡋ㸪ࢩ࣮࡛ࣛ᥋╔ࡋࡓࠋ ⣽࠸㒊ศࡣ㸪༙⏣ࡈ࡚࡛᥋╔ࡋࡓࠋ㣕⾜⯪ඛ➃ ↓⥺࣓࢝ࣛࢆྲྀࡾࡅ㸪ୗ㒊ไᚚᇶᯈ⏝ࢦࣥ ࢻࣛࢆྲྀࡾࡅࡓ㸦ᡂ⯪యࢆ )LJ ♧ࡍ㸧ࠋ ไᚚ⨨ࡋ࡚ Arduinoࢆࡗ࡚ࡑࢀࡒࢀ㸱ࡘࡢ ࣮ࣔࢱ㸦๓㐍͌᪕ᅇ⏝࣮ࣔࢱ ྎ㸪ୖ᪼⏝࣮ࣔࢱ ྎ㸹Fig. 3 ࢆཧ↷㸧ࢆṇ࣭㏫ᅇ㌿ࡋ㸪PMW㸦ࣃࣝ ࢫไᚚ㸧ࢆ⏝࠸࡚࣮ࣔࢱࡢᅇ㌿ᩘࢆኚ࠼ࡿࡇ࡛ 㣕⾜⯪ࡣ㸪࣍ࣂࣜࣥࢢ≧ែ࡛᪕ᅇࡍࡿࠋࡇࡢࢦࣥ ࢻࣛୗ㒊ᆅ㠃ࡢ㊥㞳ࢆ ᐃࡍࡿ㉸㡢Ἴࢭࣥࢧ 蒸着フィ
Fig. 3: System configuration of our airship
Fig. 4: Search algorithm of the Airship robot ࢆୗྥࡁྲྀࡾࡅࡓࠋ ᦚ㍕ไᚚࢩࢫࢸ࣒ࡢᵝࡣ௨ୗࡢ㏻ࡾ࡛࠶ࡿࠋ ࣭ ไᚚࢩࢫࢸ࣒㸸Arduino Fio ⤌㎸ࡳ ࣭ Xee ↓⥺㏦ཷಙࢩࢫࢸ࣒ ࣭ 㧗ᗘ⥔ᣢ⏝ࢭࣥࢧ㸸㉸㡢Ἴ ㊥ࢭࣥࢧ ࣭ *36 ࢭࣥࢧ ࣭ ↓⥺࣓࢝ࣛ㸦 ⏬⣲㸪ゎീᗘ 㸪 LQFK&026㸪ど㔝ゅ r㸧 ⏬ീㄆ㆑ࡼࡿ AR ࣐࣮࢝ࡢ᥈⣴ ◊✲ࡢึᮇẁ㝵ࡋ࡚ $5 ࣐࣮࢝ࢆ㐼㞴⪅ࡲࡓࡣ ⿕ᐖ⪅࡛࠶ࡿ௬ᐃࡋ㸪᥈⣴ᑐ㇟≀タᐃࡍࡿࠋ $5 ࣐࣮࢝ㄆ㆑ࢩࢫࢸ࣒ࡣ Processing ᑐᛂࡍࡿ NyARToolkit>@ࢆ⏝࠸ࡓࠋ 㣕⾜⯪ඛ➃ࡢ↓⥺࣓࢝ࣛࡢ⏬ീࢆ 3& ഃ࡛ฎ⌮ࢆ ⾜࠸㸪࣓࢝ࣛࡢ┘ど࢚ࣜࣖỴࡵࡽࢀࡓࢱࣉࡢ $5 ࣐࣮࢝ࡀᫎࡗࡓࡽ $5 ࣐࣮࢝ㄆ㆑ࢆ⾜࠸㸪Arduino ഃไᚚࢥ࣐ࣥࢻࢆⓎಙࡋ㸪᥈⣴ᑐ㇟≀ࢆぢࡘ ࡗࡓࡇࢆ⾲ࡍືసࢆ⾜࠺ࠋ Fig.4 ♧ࡍࡼ࠺㸪AR ࣐࣮࢝ࢆ᥈⣴ࡍࡿ㡿ᇦ ࢫ࣭࢟ࣕࣥࣜࢦࣜࢬ࣒ࢆ⪃࠼ࡓࠋࢫࢸࢵࣉ ࡛ ࡣ┠ᶆ⨨ࢆタᐃࡋ㸪ࢫࢱ࣮ࢺⅬࡽ᥈⣴ࢆጞ ࡵࡿࠋ㛫ࡢኚࡼࡗ࡚㣕⾜⯪ࡀ┠ᶆ್ྥ ࡗ࡚┤⥺㐠ືࢆࡍࡿࠋࢫࢸࢵࣉ ࡛ࡣࢫࢸࢵࣉ㸯 ㏫᪉ྥ࡛┤⥺㐠ືࢆࡋ㸪᥈⣴ࢆ⾜࠺ࠋࢫࢸࢵࣉ 㸪 ࡛ࢫࢸࢵࣉ ࡢ⧞ࡾ㏉ࡋ࡞ࡿண࡛ࡁࡿࠋ ❶࡛㸪ୖ㏙ࡋࡓ᥈⣴ࣜࢦࣜࢬ࣒ࢆᐇࡍࡿࡓ ࡵᚲせ࡞ࡿ㣕⾜⯪ࡢ┤⥺㐠ືࡘ࠸᳨࡚ウࡋ㸪┤ ⥺㐠ື᪉⛬ᘧࡽᩘᏛࣔࢹ࣭ࣝ⥺ᙧࢆ⾜࠸㸪㣕 ⾜⯪ࡀྍไᚚ࡛࠶ࡿࢆุูࡍࡿࠋࡑࢀࡼࡗ࡚ ᐇ㝿 AR ࣐࣮࢝ࡢ᥈⣴ᐇ㦂ࢆ⾜࠺ࠋ⌧ẁ㝵࡛ $5 ࣐࣮࢝ࢆ᥈⣴ᑐ㇟≀タᐃࡋࡓࡀ㸪AR ࣐࣮࢝ࢆ࣐ ࣮࣮࢝ࣞࢫ $5 ௦࠼ࢀࡤ㸪ᑐ㇟≀ࡢࣃࢱ࣮࣐ࣥࢵ ࢳࣥࢢ࣭㑇㊧ࡢ᳨ฟ➼⧅ࡀࡿ⪃࠼ࡿࠋ⮬ᚊⓗ 㐼㞴⪅ࢆ᥈⣴ࡍࡿࣟ࣎ࢵࢺࡣ㸪ᑐ㇟≀ࡀே㛫 ࠺ࢆ㆑ูࡍࡿᚲせࡀ࠶ࡾ㸪㢦ㄆ㆑ࡲࡓࡣ Kinect ࢭࣥࢧࢆ⏝࠸ࡓேయㄆ㆑㸪ࣁࣥࢻࢪ࢙ࢫࢳ ࣕㄆ㆑➼ࡢ⏝ࡀ⪃࠼ࡽࢀࡿࠋ ᮏ❶࡛㸪㣕⾜⯪ࡢ ㍈ୖࡢୖ᪼㐠ືࡲࡓࡣỈᖹ᪉ ྥࡢ┤⥺㐠ືࢆࢃࡅ᳨࡚ウࢆ⾜࠺ࠋ ࡑࡢ㝿ࡢ௬ᐃࡣ௨ୗࡢ㏻ࡾ࡛࠶ࡿࠋ (1) 㣕⾜⯪ࡢࣆࢵࢳ࣮ࣚࡣ⪃៖ࡋ࡞࠸ࠋ (2) 㣕⾜⯪✵Ẽࡢᦶ᧿ࡣ⪃៖ࡍࡿࠋ (3) 㣕⾜⯪ࡢ࿘ࡾࡢ✵Ẽࡢὶࢀࡣ࡞࠸ࡶࡢࡍࡿࠋ (4) 㣕⾜⯪ࡢỈᖹ᪉ྥࡢ⨨ࡣィ ࡲࡓࡣ᥎ᐃ࡛ ࡁࡿࠋ 㣕⾜⯪ࢆ୍ᐃࡢ㧗ᗘࡲ࡛ୖ᪼ࡋ㸪࣍ࣂࣜࣥࢢࡉࡏ ࡿࡓࡵᆶ┤㐠ືࢆ⪃࠼ࡿࠋࢩࢫࢸ࣒ࡢධຊࡣୖ᪼ ⏝ࣉࣟ࣌ࣛࢆ᭷ࡋ࡚࠸ࡿ㸦Fig. 5ࢆཧ↷㸧ࠋ㣕⾜⯪ ࡢᆶ┤᪉ྥ㐠ືࡣ㸪(1)ᘧ࡛⾲ࡍࡇࡀ࡛ࡁࡿࠋ
mg
T
t
F
t
z
D
t
z
m
(
)
z(
)
z(
)
z (1) ࡇࡇ࡛㸪㣕⾜⯪ࡢ㉁㔞 m[kg]㸹ᆶ┤᪉ྥࡢᦶ᧿ຊ Dz[kg/s]㸹୰ᚰࣉࣟ࣌ࣛࡢ᥎ຊFz㸹ᾋຊTzࡍࡿ 㸦Tz|mg㸧ࠋ ≧ែ✵㛫ࣔࢹࣝᑟฟࡢࡓࡵ㸪z
⨨ Ѝ x1z
㏿ᗘ Ѝ x2 zF
᥎ຊ Ѝ u ࠾ࡁ㸪ࡉࡽ'
T
mg
T
z ࠾ࡃ㸪x
Ax
Bu
'
T
/
m
(2) ࡢ≧ែ᪉⛬ᘧࢆᚓࡿࠋࡓࡔࡋ㸪»
¼
º
«
¬
ª
»
¼
º
«
¬
ª
»
¼
º
«
¬
ª
m
B
m
D
A
x
x
x
z1
0
,
0
1
0
,
2 1３．飛行船の数学モデルと同定
3.1 飛行船の数学モデルの導出 3.1.1 上昇運動の数学モデルFig. 5: Side view of the Airship ヨసࡋࡓ㣕⾜⯪࡛ࡣTz|mg࡛࠶ࡿࡢ࡛㸪≧ែ᪉⛬ ᘧࡣ
Bu
Ax
x
(3) ⾲ࡍࡇࡀ࡛ࡁࡿࠋࡲࡓ㸪ฟຊ᪉⛬ᘧࡣ
Du
Cx
y
(4) ࡓࡔࡋ㸪 C =[1 0]㸪D = 0 ࡍࡿࠋ ௨ୖࡢᩘᏛࣔࢹࣝࡘ࠸࡚ྍไᚚᛶࢆࢳ࢙ࢵࢡ ࡍࡿ㸪ྍไᚚ⾜ิࡣ Co==»
¼
º
«
¬
ª
17
1
.
90
.
3
90
.
3
0
(5) ࡑࡢࣛࣥࢡࡣ rank(Co)=2 ࡞ࡾࢩࢫࢸ࣒ࡣྍไᚚ࡛࠶ࡿ㸦ࡓࡔࡋ㸪ୖᘧࡢ ᩘ್ࡣḟ⠇ࡢྠᐃࡢ⤖ᯝࡼࡿ㸧ࠋ Fig. 7 ࢆཧ↷㸪ᗙᶆ⣔x-y ࢆࡾ㸪㔜ᚰࢆᅇ㌿ ㍈ࡋࡓࡁࡢ㣕⾜⯪ࡢ๓᪉ྥx ㍈ࡢ࡞ࡍゅࢆ ȟࡍࡿࠋ㣕⾜⯪ࡢỈᖹ᪉ྥࡢ㐠ື᪉⛬ᘧࢆ௨ୗ ⾲ࡍࠋ fNr
J
F
y
y
m
F
x
x
m
T
\
T
T
P
T
P
sin
cos
(6) ࡓࡔࡋ㸪F
F
RF
L,
N
F
RF
L࠾ࡃࠋ ୖࡢ㐠ື᪉⛬ᘧࡣ㠀⥺ᙧ࡛࠶ࡾ㸪ࡇࡢࡲࡲ⥺ᙧ ࡋ࡚ࡶࢩࢫࢸ࣒ࢆྍไᚚᛶࡣᡂ❧ࡋ࡞࠸[8]ࠋࡇࡢ ࡓࡵ┠ᶆ್ᑐࡍࡿ೫ᕪ⣔ࣔࢹࣝࢆኚࡍࡿ ࡇࢆ⪃࠼ࡿࠋ ≧ែࢆx
e>
x
ey
eT
ex
ey
eT
e@
T ࠾ࡁ㸪 ┠ᶆ್x
ref 㸪೫ᕪ⣔x
e ࢆ௨ୗࡢࡼ࠺࠼ࡿࠋ e ref e ref e ref e ref e refN
N
N
F
F
F
y
y
y
x
x
x
T
T
T
(7)Fig. 6: Top view of the Airship
ࡇࡇ࡛㸪ῧᏐࡢ ref ࡣ┠ᶆ್㸪e ࡣ┠ᶆ್ࡢ೫ ᕪࢆ⾲ࡍࠋ ⥺ᙧࡢࡓࡵ㸪೫ᕪࡣᚤᑠࡍࡿࠋࡉࡽ㸪 ┠ᶆ್ࡢᚤᑠኚືࢆ⪃࠼࡚㸪௨ୗࡢࡼ࠺࡞㸯ḟ㏆ ఝࢆ⾜࠺ࠋࡍ࡞ࢃࡕ㸪 ref ref ref ref ref ref
y
t
t
y
y
x
t
t
x
x
)
(
)
(
0 0 (8) ࡑࡢ⤖ᯝ㸪 (9) (10) ࡇࡇ࡛㸪t
0ࡣึᮇ้ࡍࡿࠋx ,
refy
refࡣ୍ᐃ ࡋ㸪T
refࡣ 0 ࡍࡿࠋ ࡲࡓ㸪ධຊࢆu
>
F
N
@
T࠾ࡁ㸪┠ᶆ್x
ref ࢆ ㏣㊧ࡋ࡚࠸ࡿࡁࡢධຊࢆu
ref 㸪೫ᕪ⣔ࡢධຊu
e ࢆ௨ୗࡢࡼ࠺࠼ࡿࠋ>
@
T ref ref refF
N
u
(11)>
@
ref T e e e F N u u u (12) ௨ୖࡢኚࡼࡾ㸪ḟࡢ೫ᕪ⣔ࡢ⥺ᙧࢩࢫࢸ࣒ ࡋ࡚⾲⌧࡛ࡁࡿࠋ e e e e eA
x
B
u
x
(13) ࡓࡔࡋ㸪 Ae㸻 » » » » » » » » ¼ º « « « « « « « « ¬ ª J m m F m m F ref ref ref ref
\
P
T
P
T
0 0 0 0 0 0 0 cos 0 0 0 0 sin 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0>
@
T ref ref ref ref ref ref refx
y
x
y
x
T
T
>
@
ref T e e e e e e ex
x
y
x
y
x
x
T
T
3.1.2 水平方向の直進運動の数学モデル
ฟຊ᪉⛬ᘧࡣ㸪 e e e e e
C
x
D
u
y
ሺͳͶሻ ⥺ᙧࢩࢫࢸ࣒ሺͳ͵ሻᘧࡢྍไᚚᛶࡘ࠸࡚ࡣ㸪ྍ ไᚚᛶ⾜ิࡣ>
B AB A B@
Co 5 ሺͳͷሻ ࡛࠶ࡾ㸪 rank(Co) = 6 ࡞ࡗ࡚ྍไᚚ࡛࠶ࡿࠋỈᖹ᪉ྥࡢ㐠ືࡘ࠸࡚ ࡣ㸪┠ᶆ್ᑐࡍࡿ೫ᕪ⣔࡛⾲⌧ࡍࡿࡇࡼࡾ㸪 ⥺ᙧࢩࢫࢸ࣒ࡣྍไᚚ࡞ࡿࡇࡀࢃࡿࠋ 㐠ືࢩ࣑࣮ࣗࣞࢩࣙࣥࢆ⾜࠺ࡓࡵヨసࣔࢹࣝ ࡢ㉁㔞㸪ࣉࣟ࣌ࣛࡢ᥎ຊ㸪㐠ື᪉ྥࡢ័ᛶ࣮࣓ࣔࣥ ࢺ㸪㐠ື᪉ྥࡢᦶ᧿ಀᩘ࡞ࡢࣃ࣓࣮ࣛࢱࢆྠᐃࡍ ࡿᚲせࡀ࠶ࡿࠋࣃ࣓࣮ࣛࢱࢆỴࡵࡿࡓࡵ㸪㣕⾜⯪ ୖ࣮᪼ࣔࢱ᥎ຊࡢィ ᐇ㦂㣕⾜⯪ୖ᪼᪉ྥ㏿ᗘ ࡢィ ᐇ㦂ࢆ⾜ࡗࡓࠋタᐃࡋࡓࣃ࣓࣮ࣛࢱࡢ⤖ᯝࢆ Table 1♧ࡍ 㐠ື᪉⛬ᘧ࠾ࡅࡿධຊࡣ᥎ຊ࡛࠶ࡿࠋ࣮ࣔࢱࡢ ᅇ㌿ᩘࡽ ࡢᩘ್㸦ࢹ࣮ࣗࢸẚ㸧ࢆ㣕⾜⯪ ㏦ಙࡍࡿࡇࡼࡗ࡚ࣉࣟ࣌ࣛࡢ᥎ຊࢆィ ࡋࡓࠋ Fig.7 ࡣࢹ࣮ࣗࢸẚᑐࡍࡿྛࣉࣟ࣌ࣛ᥎ຊࡢ ィ ࢹ࣮ࢱࢆ⾲ࡍࠋྑࣉࣟ࣌ࣛᕥࣉࣟ࣌ࣛࡢ᥎ຊ ࡣⱝᖸࣥࣂࣛࣥࢫࡀ࠶ࡿࡇࡀࢃࡿࠋ Fig.8 㸪ᚋ㏙ࡍࡿᅇ㌿㐠ື࡛ᚲせ࡞ࡿ㏫ᅇ㌿ ࡢሙྜࡢ᥎ຊᑐࢹ࣮ࣗࢸẚࡢ ᐃࢹ࣮ࢱࢆ♧ࡍࠋTable 1: System identification results; Value and unit of the parameters
Fig. 7: System identification results; Thrusts of each propeller for normal rotation
Fig. 8: System identification results; Thrusts of each propeller for revers rotation
Fig. 9: System identification results;
Relation between time and distance in upward direction
ୖ᪼᪉ྥ᥎㐍᪉ྥᦶ᧿ಀᩘࢆᑟฟࡍࡿࡓࡵ㸪 㣕⾜⯪ࡢᆶ┤᪉ྥࡢ㏿ᗘ┤㐍㏿ᗘࢆィ ࡍࡿᚲ せࡀ࠶ࡿࠋSimulinkࢆ⏝࠸࡚㸪ࢹ࣮ࣗࢸẚࢆኚ࠼ ࡚ୖ᪼㏿ᗘࡢ ᐃࡍࡿᐇ㦂ࢆ⾜ࡗࡓࠋ ྠᵝ㸪ྛࢹ࣮ࣗࢹẚ࠾ࡅࡿ┤㐍᪉ྥࡢ㏿ ᗘࢆィ ࡍࡿᐇ㦂ࡣ㸪๓ࡶࡗ࡚ຍ㏿ࡍࡿ㊥㞳➼㏿ ࡍࡿ㊥㞳ࢆணࡋ࡚⾜ࡗࡓࠋ㣕⾜⯪ࡀ๓㐍ࡋጞࡵ㸪 ㏿ᗘࡀ୍ᐃ࡞ࡗࡓⅬࡽ୍ᐃࡢ㊥㞳ࡢ㛫ࡢ 㛫ࢆ ᐃࡋࡓࠋFig.10 ࡣ㸪ࢹ࣮ࣗࢸẚᑐࡍࡿ 3.2 飛行船ロボットのシステム同定 3.2.1 各プロペラ（上昇・右・左）推力の測定 3.2.2 摩擦係数の導出 Be= ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣ ⎡ J r m m f ref ref 0 0 sin 0 cos 0 0 0 0 0 0
θ
θ
Fig. 10: System identification results; Upward and forward speed of the airship
㸲㸬ไᚚ⣔タィࢩ࣑࣮ࣗࣞࢩࣙࣥ
ᅇつᐃືసࡣ㸪1mୖ᪼ࡋࡓᚋ㣕⾜⯪ࢆ୍ᐃ ࡢ㧗ᗘ࡛࣍ࣂࣜࣥࢢࡍࡿࡇࢆᐃࡋ࡚ࢩ࣑ࣗࣞ ࣮ ࢩ ࣙ ࣥ ࢆ ⾜ ࡗ ࡓ ࠋ MATLAB R2014a ࡢ Control System Toolbox9.7㸸PIDㄪᩚჾࣉࣜࢣ࣮ࢩࣙࣥࢆ ࡗ ࡚ 㸪 PID ไ ᚚ ჾ ࡢ ㄪ ᩚ ࢆ ⾜ ࡗ ࡓ ࠋ Fig.11 ࡣ simulinkࡢࣈࣟࢵࢡ⥺ᅗ࡛࠶ࡿࠋPIDไᚚჾࡢ᭱㐺࡞ ࣃ࣓࣮ࣛࢱࡣ㸪Kp= 0.124㸪Ki= 0.0031㸪Kd= 0.415 ࡛ ࠶ࡗࡓࠋ ึᮇ≧ែ z0= 0㸪┠ᶆ್ࡣrz= 1 タᐃࡋࡓࢩ࣑ࣗ ࣮ࣞࢩࣙࣥ⤖ᯝࢆ Fig.12 ♧ࡍࠋ2YHU6KRRWࡀ࠶ࡿ ࡢࡣ㸪Fig.13 ♧ࡍࡼ࠺㸪ࣉࣟ࣌ࣛࡢ᥎ຊࢆࠕ㸩ࠖ ࡋ࠼࡚࡞࠸ࡼ࠺タᐃࡋ࡚࠸ࡿࡇࡼࡿࠋ ࢩ࣑࣮ࣗࣞࢩࣙࣥ⤖ᯝࡼࡾᩚᐃ㛫⣙⛊࡛㸪 ┠ᶆ್㸯Pࢆ㐩ᡂࡋ࡚࠸ࡿࡇࡀศࡿࠋࡑࡢᚋࡣ㸪 ᆶ┤ୖྥࡁࡢ᥎ຊࡔࡅ࡛┠ᶆࡢ㸯㹫ࡢ㧗ᗘࢆ⥔ᣢ ࡛ࡁࡿࡇࢆ☜ㄆࡋࡓࠋ ẚ㍑ࡢࡓࡵ㸪Fig.12ࡣ㉥⥺࡛ࡇࡢ3,'ไᚚჾࢆ
Fig. 11: Simulationblock diagram of PID control
Fig. 12: Compering simulation and experiment result
Fig. 13: PID control input for simulation
ᐇࡋࡓ㣕⾜⯪ࡢᐇ ࡢ㧗ᗘኚࢆ♧ࡍࠋᐇ ࡢ㧗 ᗘኚࡣࢩ࣑࣮ࣗࣞࢩࣙࣥࡢ⤖ᯝẚ࡚㸪❧ࡕୖ ࡀࡾࡢ㐜ࢀᩚᐃ⮳ࡿࡲ࡛ࡢᗄศࡢືࡀㄆࡵ ࡽࢀࡿࠋ୧⪅ࡢ㐪࠸ࡢཎᅉゎ᫂ࡣ㸪༑ศ࡛࠶ࡿࡀ㸪 ᐇ⏝ୖၥ㢟ࡀ࡞࠸ุ᩿ࡋࡓࠋ Fig.4 ♧ࡋࡓࡼ࠺㸪᥈⣴ࡢᇶᮏࡣ࠶ࡿ㡿ᇦࢆ Web࣓࡛࢝ࣛ┤⥺ⓗࢫ࢟ࣕࣥࡍࡿࠋࡑࡢࡓࡵ LQRไᚚࡼࡿ㸪┤⥺㐠ືࡢࢩ࣑࣮ࣗࣞࢩࣙࣥࢆ⾜ ࡗࡓࠋFig.14 ࡣ㸪┠ᶆ್ࡲ࡛ࡢ㣕⾜⯪ࡢ㐠ືࡢᵝ Ꮚࢆ♧ࡍࠋỈᖹ㐠ືᑐࡋไᚚࢆ㑅ࢇࡔࡢࡣ㸪ไᚚ ධຊᩘࡀ2࡛࠶ࡿࡓࡵ࡛࠶ࡿࠋ ┤㐍๓㐍㐠ືࡢ LQR ไᚚࡢホ౯㛵ᩘタᐃࡋࡓ 㔜ࡳ⾜ิࡣ௨ୗࡢ㏻ࡾ࡛࠶ࡿࠋ
dt
Ru
u
x
Q
x
J
t e e e t e)
(
0³
f (15)Q
= diag[1 500 500 10 500 500]
R
= diag[1000 1000]
LQR ไᚚჾࡣ Fig.14 ࡢǼx㸪Ǽy ࠾ࡼࡧǼȟࡀ 0 ࡞ࡿࡼ࠺ไᚚࢆ⾜࠺ࠋ MATLAB R2014a ࢆ⏝࠸࡚ࣇ࣮ࢻࣂࢵࢡࢤFig. 14: Airship motion control for horizontal direction
上昇速度と前進速度の変化を示す。 それぞれデューティ比によって摩擦係数が変動 するので，上昇時の摩擦係数の変動範囲は 3.51× 10-2から12.03×10-2 [kg/s]であった。変動対応でき るようにするため7.77×10-2 [kg/s]を基準とした。 4.1 上昇運動のシミュレーションと実測値との比較 4.2 水平運動のシミュレーションと実測値の比較
Fig. 15: Simulation results of optimal regulator for
horizontal motion ࣥ . ࢆタᐃࡋࡓࠋ㓄ิ $H%H&' ࢆධຊࡋ㸪 MATLAB ࡢ OTU 㛵ᩘࢆࡗ࡚≧ែࣇ࣮ࢻࣂࢵࢡ ࢤࣥ.ࢆồࡵࡓ⤖ᯝࡣ௨ୗࡢ㏻ࡾ࡛࠶ࡿࠋ LQR ไᚚჾࡢࢳ࣮ࣗࢽࣥࢢࡣࢩ࣑࣮ࣗࣞࢩࣙࣥࡢ ⤖ᯝࢆࡳ࡞ࡀࡽุ᩿ࡋࡓࠋ ࢳ࣮ࣗࢽࣥࢢᚋࡢࢩ࣑࣮ࣗࣞࢩࣙࣥ⤖ᯝࡢ ࢆ Fig.15 ♧ࡍࠋᶓ㍈ࡣ㛫࡛㸪⦪㍈ࡣྛ೫ᕪࢆ ⾲ࡍࠋᅗࡽ㸪┠ᶆᆅⅬ
[
UHI㐩ࡍࡿᩚᐃ㛫ࡣ ⣙ ⛊࡛࠶ࡿࡇࡀࢃࡿࠋ Ỉᖹ᪉ྥ┤⥺㐠ືࡢࢩ࣑࣮ࣗࣞࢩࣙࣥ⤖ᯝࢆ㋃ ࡲ࠼࡚㸪ᐇ㝿ࡢ㣕⾜⯪ LQR ไᚚჾࢆ⤌ࡳ㎸ࢇ࡛ ᐇ㦂ࢆ⾜ࡗࡓࠋࡑࡢࣈࣟࢵࢡ⥺ᅗࢆ Fig.16 ♧ࡍࠋ ➨ ẁ㝵ࡋ࡚㸪㢼࡞ࡢእࢆ㜵ࡄࡓࡵ㸪 㣕⾜⯪ࡢไᚚ⨨ࢆᦚ㍕ࡋࡓ࢝ࢦࢆᡭ࡛ᣢࡗ࡚ ⨨᪉ྥࢆኚ࠼࡚ไᚚࢩࢫࢸ࣒ࡢືసࢆ↓⥺㏻ಙ ࡼࡾእ㒊 3& ࡛☜ㄆࡍࡿᐇ㦂ࢆ⾜ࡗࡓࠋᐇ㦂࡛ࡣ GPS ࢭࣥࢧࡗ࡚㸪ࢫࢱ࣮ࢺᆅⅬ $ ┠ᶆᆅ Ⅼ % ࡢ⤒ᗘ⦋ᗘࢆྲྀᚓࡋࡓࠋලయⓗࡣ $ Ⅼ ࡽ % Ⅼࡢ㊥㞳ィ⟬ࡋ㸪ࡑࡢ㛫┤⥺ࢆᘬࡁ㸪┤⥺ ࡽࡢࡎࢀ߂\ࢆฟຊࡍࡿࠋฟຊࡉࢀࡓ߂\ࢆ࡞ࡃࡍ ࡓࡵไᚚࢩࢫࢸ࣒ࡀ㣕⾜⯪ࡢྑᕥࡢ࣮ࣔࢱ ฟຊࢆࡾᙜ࡚ࡿࠋไᚚࢩࢫࢸ࣒ࡢືస☜ㄆࡢࡓ ࡵ㣕⾜⯪ࡢ࢝ࢦࢆ┤⥺ࡽࡎࡽࡋ࡚┤⥺ࡽࡢ ࡎࢀ߂\ࢆ ᐃࡍࡿᐇ㦂ࢆ⾜ࡗࡓࠋ ḟࡢẁ㝵࡛㸪ᒇእ࡛Ỉᖹ᪉ྥࡢ┤㐍㐠ືࡢ㣕⾜ ᐇ㦂ࢆ⾜ࡗࡓࠋࡑࡢ⤖ᯝࢆ Fig.17 ♧ࡍࠋ㣕⾜⯪ ࡣ㢼ࡢᙳ㡪ࢆཷࡅࡸࡍ࠸ࡇ㸪⏝ࡋࡓ *36 ࡢ ᐃ⢭ᗘ㸦Ỉᖹ᪉ྥࡢᖹᆒㄗᕪ P ⛬ᗘ㸧ࡢ㛵ಀ ࡼࡾࢹ࣮ࢱࡤࡽࡘࡁࡀ࠶ࡿࡀ㸪߂[㸪߂\㸪߂ȟࢆ ࡍࡿࡼ࠺ LQR ไᚚჾࡀാ࠸࡚࠸ࡿࡇࡀㄆࡵ ࡽࢀࡓࠋ㢼ࡢᙳ㡪ࡼࡾ߂ȟࡢಟṇࡀ࠺ࡲࡃ࠸ࡗ࡚ ࠸࡞࠸ᵝᏊࡀぢࡽࢀࡿࠋFig. 16: Simulation diagram for LQR control system
Fig. 17: An experimental result of horizontal motion
using LQR controller system
Fig. 18: 180-degree rotation of the airship 㐼㞴⪅࡞ࢆ⏬ീㄆ㆑ࡼࡾ᥈⣴ࡍࡿࡇࡀ᭱ ⤊ⓗ࡞┠ᶆ࡛࠶ࡿࡀ㸪ᮏ✏࡛ࡣࡑࡢ┠ᶆࢆ༢⣧ࡋ ࡚㸪㣕⾜⯪ᆺࣟ࣎ࢵࢺࡀAR࣐࣮࢝ࢆ⮬ᚊⓗ᥈⣴ ࡍࡿᐇ㦂ࢆ⾜ࡗࡓࠋ Fig.4♧ࡋࡓࡼ࠺㸪᥈⣴ࡢᇶᮏࡣ࠶ࡿ㡿ᇦࢆ :HE࣓࡛࢝ࣛ┤⥺ⓗࢫ࢟ࣕࣥࡋ㸪᥈⣴ᑐ㇟ࡢ$5࣐ ࣮࢝ࡀ᳨ฟࡉࢀࢀࡤ㸪᥈⣴ࢆ⤊ࡍࡿࠋࡑࡇ࡛㸪㣕 ⾜⯪ࡣ⮬ᚊⓗṆࡋ㸪╔㝣ࡍࡿࣝࢦࣜࢬ࣒࡛࠶ ࡿࠋ᥈⣴㡿ᇦࡢ➃ࡲ࡛㐍⾜ࡋ࡚ࡶ᥈⣴࡛ࡁ࡞࠸ࡁ ࡣ㸪Fig.18ࡢࡼ࠺rᅇ㌿ࡋ࡚ࢫ࢟ࣕࣥࢰ࣮ࣥ ࢆࡎࡽࡋࡓࡢࡕ㸪ࡧ㏫᪉ྥ᥈⣴ࡢࢫ࢟ࣕࣥࢆ⾜ ࠺᪉ᘧ࡛࠶ࡿࠋ㣕⾜⯪ࡢᅇ㌿㐠ືࡢᩘᏛࣔࢹࣝࡣ⏝ ᐇ㦂ࢆ⡆༢ࡍࡿࡓࡵ㸪࠶ࡽࡌࡵ$5࣐࣮࢝ࢆ㣕 ⾜⯪ࡢ๓᪉ࡢᗋୖタ⨨ࡋ࡚࠾ࡃࠋ᥈⣴ࣉࣟࢢࣛ ࣒ࢆᐇ⾜ࡍࡿⅬ࡛㣕⾜⯪ᆺࣟ࣎ࢵࢺࡀ㞳㝣ࡋ㸪㧗 ᗘࢆ୍ᐃ㸦P㸧ಖࡗࡓ≧ែ࡛┤㐍㐠ືࢆࡍࡿࠋࡑ
５．ARマーカの探索実験
いず，方向を変えて次の初期位置に移動する。Fig. 19: Snapshot of control experiment exploring an AR marker ࡢ㝿㸪⯪యࡢඛ➃㒊ศタ⨨ࡋࡓ↓⥺࣓࢝ࣛࡢ⏬ീ ࣓࣮ࢪࢆฎ⌮ࡋ㸪ᑐ㇟≀࡛࠶ࡿAR࣐࣮࢝ࢆㄆ㆑ ࡋࡓⅬ࡛㣕⾜⯪ࡀ╔㝣ࡍࡿືసࢆ⾜࠺ࡼ࠺࡞ࢩ ࢫࢸ࣒࡛࠶ࡿࠋࡇࡇ࡛ࡣ㸪⮬ᚊᆺ㣕⾜⯪ࣟ࣎ࢵࢺ ⏬ീㄆ㆑ࢆ⏝ࡋ࡚AR࣐࣮࢝᥈⣴ࢩࢫࢸ࣒ࡢᐇ㦂 ࢆ⾜ࡗࡓࠋAR࣐࣮࢝ࡀ⏬ീࡢ୰ኸ࡞ࡿࡼ࠺ࢺ ࣛࢵ࢟ࣥࢢࡋ࡚࠸ࡿࠋ ึᮇẁ㝵࡛࠶ࡿࡓࡵ㸪⏬ീㄆ㆑ࡣ▷↔ⅬࡢWeb ࣓࢝ࣛࢆ⏝ࡋ㸪ᐇ⌧ྍ⬟ᛶࢆ☜ㄆࡋࡓࠋᑗ᮶ࡣ㣕 ⾜ᆺࣟ࣎ࢵࢺࡢ࣓࢝ࣛࢬ࣮࣒ࢵࣉ࡛ࡁࡿࣞࣥ ࢬࢆྲྀࡾࡅࡿࡇ࡛㸪㐲㊥㞳ࡽࡶ≉ᐃࣃࢱ࣮ࣥ㸪 㢦㸪ேయ➼ࢆㄆ㆑ࡍࡿࡇࡀ࡛ࡁࡿ⪃࠼࡚࠸ࡿࠋ ARࢩࢫࢸ࣒ࢆ࣐࣮࣮࢝ࣞࢫARࢩࢫࢸ࣒௦࠼ ࢀࡤ㸪ᑐ㇟≀ࡢᙧ࣭㑇㊧➼ࡢ᳨ฟ⧅ࡀࡾ㸪㢦ㄆ㆑ ࢩࢫࢸ࣒ࢆ㢦ㄆドࢩࢫࢸ࣒௦࠼ࢀࡤ㸪ᑐ㇟≀ࡀ≉ ᐃࡢ᥈⣴ᑐ㇟ே≀࡛࠶ࡿࡇࡢ㆑ูᙺ❧ࡘ⪃ ࠼࡚࠸ࡿࠋ㣕⾜⯪ᆺࣟ࣎ࢵࢺࡢAR࣐࣮࢝᥈⣴ᐇ㦂 ࡢᵝᏊࢆFig.19 ♧ࡍࠋ ᮏ✏࡛ࡣ㸪◁₍࡞ࡢᗈ࠸㡿ᇦ࡛㐼㞴⪅ࡸ⿕⅏⪅ ࢆ᥈⣴ࡍࡿࢣ࣮ࢫ࡛㸪㐼㞴⪅ࡸᑐ㇟≀᥈⣴ࢆ࡛ࡁࡿ ࡔࡅࢫ࣒࣮ࢬ⾜࠺ࢩࢫࢸ࣒㛤Ⓨࡢࡓࡵࡢึᮇẁ 㝵ࡢ◊✲࡛࠶ࡿࠋ⏬ീㄆ㆑ᶵ⬟ࡢຍࡼࡿ᥈⣴ື సࡢ☜ㄆࢆ┠ᣦࡋࡓࡶࡢ࡛࠶ࡿࠋࡑࡢࡓࡵ㸪㛗 㛫㣕⾜ྍ⬟࡞㣕⩧యࡋ࡚㣕⾜⯪ࢆ㑅ᐃࡋ㸪ࡑࡢ⮬ ᚊᆺࡢࣉࣟࢺࢱࣉࡢ㛤Ⓨࢆヨࡳࡓ㸬 ࡲࡎ㸪GPS⨨ሗࢆ⏝ࡋࡓ㡿ᇦࢫ࢟ࣕࣥไᚚ ࣝࢦࣜࢬ࣒ࢆ⪃࠼㸪㣕⾜⯪ࡢୖ᪼┤㐍㐠ືศ ࡅ࡚ᩘᏛࣔࢹࣝࢆᑟฟࡋ㸪ࢩࢫࢸ࣒ྠᐃࢆ⾜ࡗࡓࠋ ୖ᪼㸩࣍ࣂࣜࣥࢢࡘ࠸࡚ࡣ㸪PIDไᚚ⣔ࡋ㸪Ỉ ᖹ᪉ྥࡢ㐠ື㛵ࡋ࡚ࡣ㸪ྍไᚚᛶࢆ‶ࡓࡍࡼ࠺㸪 ┠ᶆ್ᑐࡍࡿ೫ᕪ⣔ࣔࢹࣝࢆ⪃࠼㸪LQRไᚚࡢᐇ ࢆྍ⬟ࡋࡓࠋ ḟ㸪MATLABࢆ⏝࠸࡚PIDୖ᪼ไᚚ᭱㐺ࣞࢠ ࣮ࣗࣞࢱࡼࡿ┤㐍ไᚚࡢࢩ࣑࣮ࣗࣞࢩࣙࣥࢆ⾜ ࠸㸪GPSࢆ⏝࠸ࡓᐇ㦂ࡢ↷ྜࢆ⾜ࡗࡓࠋ ไᚚ⣔ࡢᐇࡣArduinoࢆά⏝ࡋ㸪ࡇࢀࡽࡢไᚚ
Fig. 20: Image of three types of robots exploring cooperatively ⣔ࢆᐇࡋ㸪 180rᅇ㌿ࡍࡿ㐠ື⏬ീㄆ㆑ࡽᑐ ㇟≀㏆࡙ࡃไᚚࡢᐇ㦂ࢆ⾜ࡗ࡚㸪⮬ᚊⓗ࡞㡿ᇦࢫ ࢟ࣕࣥࡢぢ㏻ࡋࢆᚓࡓࠋ 㣕⾜⯪ࡢඛ➃Web࣓࢝ࣛࢆྲྀࡾࡅ㸪㐼㞴⪅࡞ ࡢᑐ㇟≀ࢆ᥈⣴ࡍࡿࡓࡵࡢ⏬ീㄆ㆑ᶵ⬟ࢆຍ ࡋ࡚᥈⣴ᐇ㦂ࢆ⾜࠸㸪ࣉࣟࢺࢱࣉࡋ࡚㸪ᐇ⌧ྍ ⬟ᛶࢆ᳨ドࡋࡓࠋ ⏬ീㄆ㆑ᶵ⬟㛵ࡋ࡚ࡣ㸪◊✲ࡢ⌧ẁ㝵࡛ࡣ㐼㞴 ⪅ࡸ⿕ᐖ⪅ࡢ௦ࢃࡾAR࣐࣮࢝ㄆ㆑ࢆ⾜ࡗࡓࡀ㸪 ᑗ᮶AR࣐࣮࢝ࢆ࣐࣮࣮࢝ࣞࢫAR௦࠼ࢀࡤ㸪ᑐ㇟ ≀ࡢࣃࢱ࣮࣐ࣥࢵࢳࣥࢢ࣭㑇㊧ࡢ᳨ฟ➼⧅ࡀࡿ ⪃࠼ࡿࠋࡉࡽ㸪Kinect࡞ࡢ㐺ษ࡞ࢭ ࡍࢀࡤ㸪ேయࡸேᕤ≀ࡢ᳨ฟࡶᐇ⌧ྍ⬟࡛࠶ࢁ࠺ࠋ ᮏࢩࢫࢸ࣒ࡢᐇ⏝㝿ࡋ࡚ࡣ㸪Fig.20 ♧ࡍࡼ ࠺㸪ࢱࣉࡢ␗࡞ࡿ」ᩘࡢ⮬ᚊᆺࣟ࣎ࢵࢺࡀ㐃ᦠ ࡍࡿࢩࢫࢸ࣒ࡶᐃࡉࢀࡿࠋ ཧ⪃ᩥ⊩
1) B. Siciliano and O. Khatib (Eds.): ͆Springer Handbook of Robotics͇㸪Springer-Verlag (2008).
2) 㒯ᚰ▱࣭㸸ࠕே㛫ࡢ௦᭰ࢆ┠ᣦࡍࢧ࣮ࢳࣟ࣎ࢵࢺ ࡢࡓࡵࡢࣄ࣮࣐࣭࣐ࣗࣥࢩࣥࣥࢱࣇ࢙࣮ࢫࠖ, ➨ 3 ᅇィ ⮬ືไᚚᏛࢩࢫࢸ࣒ࣥࢸࢢ࣮ࣞࢩࣙࣥ 㒊㛛ㅮ₇, Vol.I, pp.261-262 (2002).
3) E.Feron, E.N.Johnson: ͆ Handbook of Robotics Chapter44 : Aerial Robotics ͇ , Springer-Verlag, pp.1009-1029 (2008).
4) R.E.Murphy: “Handbook of Robotics Chapter 50: Search and Rescue Robotics”, Springer-Verlag, pp.1151-1173 (2008).
5) Khoury, G.A. and Gillett, J. D, eds, Airship Technology, Cambridge Aerospace Series #10, Cambridge University Press (1999). 6) ▼⏣࣭㸸ࠕᒇෆ⏝࣮ࣜࣔࢺࢭࣥࢩࣥࢢ㣕⾜⯪ࡢ㛤 Ⓨࠖ㸪᪥ᮏ⯟✵ᏱᐂᏛ୰㒊࣭㛵すᨭ㒊ྜྠ⛅ᮇ ㅮ₇㞟㸪42 ᕳ㸪pp.106-109 (2005). 7) ᶫᮏ㸸ࠕAR ࣉࣟࢢ࣑ࣛࣥࢢ̿processing ࡛స ࡿᣑᙇ⌧ᐇឤࡢࣞࢩࣆࠖ㸪࣮࣒࢜♫ (2009). 8) ▼ᕝ⦾ᶞ:ࣔࢹࣝண ไᚚࡼࡿ㣕⾜⯪ࢩࢫࢸ࣒ࡢ ไᚚ⣔タィ㛵ࡍࡿ◊✲, http://www.fl.ctrl.titech.ac.jp/paper/2005/Thesis/M_shig eki.pdf(2005) 原稿受付日 平成 28 年 1 月 18 日
![Fig. 5: Side view of the Airship ヨసࡋࡓ㣕⾜⯪࡛ࡣ T z | mg ࡛࠶ࡿࡢ࡛㸪≧ែ᪉⛬ ᘧࡣ BuAxx (3) ⾲ࡍࡇࡀ࡛ࡁࡿࠋࡲࡓ㸪ฟຊ᪉⛬ᘧࡣ DuCxy (4) ࡓࡔࡋ㸪 C =[1 0]㸪 D = 0 ࡍࡿࠋ ௨ୖࡢᩘᏛࣔࢹࣝࡘ࠸࡚ྍไᚚᛶࢆࢳ࢙ࢵࢡ ࡍࡿ㸪ྍไᚚ⾜ิࡣ Co== » ¼«º¬ª1.1790.390.30 (5) ࡑࡢࣛࣥࢡࡣ rank](https://thumb-ap.123doks.com/thumbv2/123deta/9947235.1394283/4.892.99.417.87.272/FigSideヨసࡋ㣕⾜⯪࡛ࡣTࡣࡲࡓ㸪Ꮫࣔࡿ㸪¼ࡣ.webp)
