Local Planner

1) Simple Algorithm
1) Go straight to the goal if direction of the goal is opened
2) When the direction of the goal is blocked, the robot follow the obstacle’s
wall by same direction. While following the wall, if goal direction is opened,
then move to the goal. Goto 1)

2) Many other heuristic algorithms

Greedy policy,….

Local Planner

Sensor-Based Motion Planning

POINT; It is very hard to get “complete global information” of obstacles in advance

Bug type algorithms Bug-1, Bug-2, Tangent-Bug, Algo-1, Algo-2,…

Guarantee convergence

Metric algorithm
Use of ultrasonic sensor Topological algorithm
Use of touch sensor
or laser range sensor

Other ideas (related studies)

Behavior based algorithm Such as subsumption archtecture

Learning-based algorithm
SLAM=Simultaneous Localization and Mapping
 Bug-1 algorithm
１） Go straight to the goal if direction of the goal
is opened. When the point is the goal then stop.
2) When the direction of the goal is blocked by
obstacle, the robot follows the obstacle’s wall by
same direction (predefined same direction).
The blocked point (hit point) is recorded.
When tracing the wall, all the position of the path
is recorded.
3) When returning to the hit point, find the
closest point to the goal on the followed path
and return the closest point using the shorter path.

4) Leave for the goal at the closest point. (leave point)

go to 1) (Repeat this manner)
5) When the robot cannot go to goal direction
because of obstacle, then answer “No solution” and stop.

Bug-2 algorithm
0) Draw a straight line from start to goal
1) Move to goal direction. When the current position is
the goal, then stop
2) When the direction of the goal is blocked by obstacle,
the robot follows the obstacle’s wall by same direction
(predefined same direction). The blocked point (hit point) is
recorded. When tracing the wall, all the position of the path
is recorded.

3) If detecting the straight line and the goal distance

is shorter than the case of the previous hit point, and the
goal direction is opened, then goto 1) (leave point).
Else continue to follow the wall.

4)While following the wall, if reaching the previous hit point,

then answer “no solution” and stop.
 Tangent-bug, Visi-bug algorithm
（considering distance sensor’s detecting range）
0-1) Move to goal direction. When the current position is
the goal, then stop
0-2) When detecting the obstacle wall by the range sensor,
Pi is the detected points. (Pi may be multiple points)
1)The distance from the current position x to the detected
point Pi is d1. The distance from point Pi to goal is d2.
Calculate Di=d1+d2. Go to the Pi which is the shortest Di.
When goal, then stop.
2) When reaching the obstacle’s wall, follow
the obstacle’s wall by same direction (predefined
same direction). The reached point (hit point) is recorded.
When following the wall, all the position of the path and
distance d(x, goal) are recorded.
3) When following the wall, if the goal direction is opened,
and the distance d(x, goal) is shorter than the
previous d(x, goal), then leave the wall (leave point)
and move to goal. Go to 0-1)

4) When following the wall, the robot reaches the same

point x where the shortest distance d(x, goal) ever,
With zero sensor range case then answer “no solution” and stop.

Global Planner
Roadmap method
(C-obstacleが多面体で表現されているとする）

Visibility Graph （可視グラフ）

 Global Planner
Roadmap method
(C-obstacleが多面体で表現されているとする）

Visibility Graph （可視グラフ）

Reduced visibility graph

Path generation by visibility graph (global method)

0) representing obstacles by polygons (convex or concave are both OK) in 2D-space
1) Give the sequential number from the start point to goal point including all vertices of the polygons
2) Connect the all points (vertices) with straight lines with the numbered order. If the line segment
does not penetrate the obstacle, then give the number for the line segment with the order.
(This is called initial visibility graph. Note that the edges off the polygons are also the line segments)
3) When each edge is tangent to one polygon or two polygons, then keep the line segment (edge).
Otherwise, any extension from an edge would penetrate a obstacle (see following figure).
Remove such edges (not the tangent but penetrating one) When the tangent line is extended
then penetrating, such line segment (edge) is not removed (case A). =>
This is called a reduced visibility graph.

ＮＧ
ＯＫ ＮＧ
penetrating
tangent tangent ＮＧ
ＯＫ OK (case A)
tangent
4) Add a cost for each line segment (edge) such as distance, then, we get weighted
non-directed graph (we can search the minimum cost path by such as Dijkstra method. )
5) Give the minimum cost from the start to each edge with above order (from the start to goal)
 start 2 start 2 start
1
1 1
2
7 6 4 5
3
remove

4 3 10 4 8 3
5 9
5
goal goal goal

2
start 2
1
start
1 1
1 2
2 7 6 5
7 6 5
penetrating=>remove

10 4 8 3
10 4 8 3 5 9
5 9

goal goal reduced visibility graph

Initial visibility graph

start
start Edge 2 Cost=2
2 Cost=2 start Edge 2 Cost=2 cost=2
cost=2 1
1 1 Cost=3
Cost=2 Cost=2
Cost=3 Cost=3 Cost=1
Cost=2 Cost=1 Cost=2 Cost=2 Cost=1 Cost=2
Edge
cost=3 Cost=2
Cost=2 Cost=2 4 3 Edge
4 3 4 3 5 cost=2
5 5 Cost=3
Edge Edge
Cost=3 Cost=3
cost=4
goal cost=2
 graph search 11月18日ここから
graph search
Search method for minimum traveling cost problem
for weighted non-directed graph Search method for minimum traveling cost problem
for weighted non-directed graph
Efficient search methods are based on
“Principle of Dynamic Programming” that is
Bellman Ford method
for any point (vertex) Vi on the optimum path,
the cost (start, Vi) is minimum and the cost (Vi, goal) An extended method of Daijkstra method when there is negative cost of edge.
is minimum. (Daijkstra method can not apply, because of infinite loop for the negative cost)

Daijkstra method A* method

Basically same with Daijkstra method using a priori knowledge (heuristic function)
1) Write 0 (zero) for the start vertex as the initial cost of vertices in Dijkstra's algorithm.
(other vertex set ∞ temporary) (Example of the heuristic function: direct distance to the goal)
2) Visit each vertex (with depth first or width first) and write
the current minimum cost to the vertex. When the cost at the vertex
is less than the previous cost to the edge, then update the cost with the less value. F(n): the evaluated cost at n
The updated path (less cost path) to the vertex is selected currently. Repeat this manner. g(n): actual cost from start to n
h(n): expected cost from n to goal
3) Visit all the vertices and adjacent edges, the final cost is the current cost at the goal
and the final minimum path is the path for the cost. h(n) is called “heuristic function”
For example

Roadmap method (how to make a graph) Global Planner

ボロノイ図作成方法
Voronoi diagram method How to draw Voronoi diagram
＝＞most safe path for robots For the point set , Euclidian distance d
Voronoi diagram is a partition of a plane into
regions close to each of a given set of
objects (points). For the case of points, the
The region set， is called Voronoi diagram
boundaries are perpendicular bisectors for two points.
0-1) When two points case, the Voronoi diagram is two regions
divided by a perpendicular bisector
0-２) When the three points case (not on a same line), the perpendicular bisectors
for the points divide three regions.
Vor（3)

Point 1 perpendicular bisector Poin 1

p1 remove
Voronoi region
remove Point 3
(Voronoi cell) d(p,p1) Voronoi diagram
p Voronoi diagram

Point 2 remove
d(p,p2) Point 2
Generalized Voronoi diagram p2 The cross point of the Voronoi diagram is
Voronoi diagram the center of same circle (same distance r from each point)
 繰り返し母点追加法によるボロノイ図作成
How to draw Voronoi diagram Iterative adding method to make Voronoi diagram
1) Draw the Voronoi diagram for any three points (This is Vor(3)). Set i=4
定義： 空間内の有限点集合（母点） に対して，距離をdとするとき
2) Find the Voronoi cell V(q) (=V(pi)) where the cell includes the adding point pi.
3) Draw the perpendicular bisector for pi and q. Find the first cross point of
the bisector and the Voronoi diagram.
で構成される領域Vをボロノイ領域，V をボロノイ図と言う．
4) Find the adjacent Voronoi V(q) cell for the boundary
0-1) 母点が2個の場合はボロノイ図は母点の垂直2等分線で2領域に分割
5) Set new q, repeat 3),4) until q=pi
0-２) 母点が３個（一直線上にない場合），各点の垂直2等分線は一点で交わる３領域に分割
（一直線上にある場合は，平行な２直線） V(1) Vor（3)
3母点でのボロノイ図交点は，母点から等距離＝円の中心 Vor（3) V(3)
p1 V(3) V(q)
p1 母点1
Vertical bisector 母点1 q
母点1 トル p3 p3
V(1) 母点3
トル 母点3 Adding point p4
ボロノイ図 pi
p4 pi
ボロノイ図 母点4
q p2
母点2 p2 母点2
トル V(pi)=V(2)=v(q)
母点2
1) 任意の3点（母点）p1,p2,p3によりボロノイ図を作成する（これをVor(3)とする）
i=4とおく

5) Set new q, repeat 3),4) until q=pi.

When return to the first V(p_i)=V(q) then the stage i is end.
8) i:=i+1. goto 2)
6) The loop by the new perpendicular bisectors is added to the Voronoi diagram Vor(3)
=Vor(i-1)
7) Delete the Voronoi diagram inside the loop, then we get Vor(i)
Vor（3) Vor（４)
p1
p1
V(q)=V(3) p3
母点3
p1
q 母点1 ｑ
p3 p4 p4
p1 母点3 p2
p4 pi p3 p2
母点4 p4 pi
母点4 q
p2
V(r) If the additional point q is outside the closed loop (red region),
母点2
p2 since the new loop will be larger than the region,
It becomes difficult to draw by hand drawing the diagram.
Therefore, it is easier to draw
if all the other generating points should be included in the area surrounded
8) i:=i+1. Rpeat the above by the first three points.
 Vor（3)
Idea to draw the generalized Voronoi diagram
Vor（４)
Center point 1 p1

Center point 3 p3
Expansion method
Center point4 ｑ smallest
p4
Gradually expand the walls and points
with constant speed (let's call this the
p2
expansion line). The point when the
Center point 2
Expansion lines intersecting, the points set
Hint 1) When you are not sure which line segments should be deleted, is the Voronoi diagram.
you can find out it by examining that any point on the line segment is
same distance to the closest points or not.

Hint 2) The final Voronoi diagrams are all convex regions.

Hint 3）For any point in a convex region in the Voronoi diagram,

the distance to the center point contained in the region is always
the smallest among the distances to other center points.
一般化ボロノイ図 generalized Voronoi diagram

GVD (Generalized Voronoi Diagram) ボロノイ図と可視グラフによる経路生成と最短経路探索の例

Minimum cost path by Voronoi diagram and visibility graph
Voronoi diagram Visibility graph

Safety first Performance first

Voronoi diagram Visibility graph
Voronoi diagram Visibility graph
 Cell Decomposition Method （セル分割法）
ARProb6
・Hierarchical cell decomposition method
1) 与えられた５個の母点でのボロノイ図を作成せよ （階層セル分割法）

Global Planner

Exact cell decomposition method （完全セル分割法） Probabilistic Roadmap Method (PRM)

確率ロードマップ法

ドロネー図
 Graph theory
Advanced Robotics 2022 A set of edges and vertices is called a graph
（先端ロボット工学） non-directed graph directed graph

vertex

Motion Planning edge (start edge, goal edge)

Intelligence of Robot Systems

connected graph and non-connected graph

No.3 the number of edges

starting from the
vertex v

Motoji Yamamoto
yama@mech.Kyushu-u.ac.jp Euler graph (one-stroke writing for all edge)
A non-directed graph is called an “Euler trail” if there
exists a path that has every edge just once.
Such graph is called “Euler graph”.
Kyushu University Königsberg bridge problem
Leonhard Euler (1707-1783)

Theorem of Euler graph (Euler trail) Proof: Theorem of Euler graph

All the order of vertex deg(v) are even, then Euler graph (necessary and sufficient condition)
or just two odd order of vertex, then quasi-Euler graph
1) All the order of vertex deg(v) are even <= Euler graph
(means one-stroke writing is possible)
(when Euler graph, start vertex and end vertex are same) For all the vertex of the Euler trail, when encountering the edge then
(when quasi-Euler graph, starting or end edge must be odd order vertex) always leave the edge, which means (always) even order of the vertex.
If starting edge case, the trail is stopped at the same edge, which means
even order of the edge.

2) All the order of vertex deg(v) are even => Euler graph

thus “not Euler graph”

Leonhard Euler (1707-1783)

Euler graph Quasi-Euler graph
 Proof: Theorem of Euler graph
Graph theory
Similar ideas with the Euler graph
(necessary and sufficient condition)
2) All the order of vertex deg(v) are even => Euler graph
Step1) Remove a closed loop trail C from G (described by G C)
Hamilton graph (Hamilton trail) (one-stroke writing for all vertex)

Where the order of vertex keeps even. The closed trail is an Euler trail. A trail (path) that passes through every vertex on a graph once
is called a Hamiltonian trail (Hamilton graph).
Step2)Repeat this until no closed trail
Step3) Starting with one vertex of the first closed trail.
Visit every edge of the trail (just trace the loop),
when appears other closed trail, repeat this manner, until no edges or
Two edges. Which is the route of Euler trail.
C G C
How to make a Euler trail
No sufficient and necessity condition to check the Hamilton graph,
Is known, however some sufficient conditions are known;
Euler trail
Ore’s theorem (Sufficient condition of Hamilton graph)
Euler trail v,w are no-adjacent vertices
deg(v)+deg(w) >= n
n is the number of vertex
Euler trail
Dirac’s theorem (Sufficient condition of Hamilton graph)
2×min{deg(v)}>= n

Graph theory
Ore’s theorem (Sufficient condition of Hamilton graph) Similar ideas with the Euler graph

deg(v)+deg(w) >= n v,w are no-adjacent vertices TSP (travelling salesman problem)
n is the number of vertex
Finding the path that pass through every vertex
Dirac’s theorem (Sufficient condition of Hamilton graph) once where the traveling cost is minimum one
2×min{deg(v)}>= n (minimum cost Hamilton trail)

Typical combinatorial optimization problem

Chinese postman problem

Finding the path that pass through
every edge (not always once) where
the traveling cost is minimum one
(minimum cost Euler similar trail)
The graph is not an
Euler graph generally
 Check each graph Euler graph or not Check each graph Hamilton graph or not

When Euler graph show an Euler trail

by adding the order number for each edge

グラフのデータ表現（データ構造） data structure for a graph

木（tree) １．隣接行列 adjacent matrix
A connected non-directed without any closed loops is called a tree グラフＧ=(V, E)に関わる諸問題を計算機を用いて解決するためにはその解決
のためのアルゴリズムとともに如何に効率よく計算機にＧ=(V, E)を格納する
かという点が肝要となる。
directed graph
non-directed graph

非対称行列
Binary tree Ternary tree 対称行列 No diagonal
diagonal
Case of weighted graph
data structure for a graph 2

２．隣接リスト表現 Adjacent list

（頂点ごとに接続された頂点を示す）
For each vertex, represent adjacent vertex by vertex label
無向グラフの

有向グラフの隣接リスト
演習）
１）上記の隣接リストで表現される有向グラフGを描け
有向グラフの隣接リスト ２）上記Gの隣接行列を求めよ
 (zot:1711
´
∠av¨ ce1 23っ市 c ARP,"Lto
I孔ιttt ι=「 ナl t t lt t 乃 ′二 ●′ リ
`
′
ル ,(_ ,,で
"
J/st"*"f 4 (st-s2,e,ts) "ら pz,p.t.pr
´
( r; Mok'e ta
l"l,lcJ
M.1で 1に・JI
fsz -,(.r yり
ヽ

κ′・.2,
γ
2リ
3) M4ttで c“ 11,%ハ 琢
(

(
(sr-s9) P′
′

( 4),ぃよ 誡 ,Iレ r′ 』′
ゴ′ r , ;t;^ 卜ヽ
′ ■ ,■ ,́^ b.l1 t) )>'lr),(t )
l
/ ""p
( し)IⅥ 碑ct(ttヤ シけ OOS'rC
“ rリ
Cac l" r"L1 S“
ヽ

fa。
ツ づ
●

OaL.。 Ll」 w γ S I´ 。yC V,'w=l*fi "r d ,

′

小 ,∝
`・
1.ヽ ■

t ,l,4r ,t.
` .

,
. , "rr ￨ '
1に k.
R3並 t́ rS3 。
マ
ヽ

` 1tt loセ
",(5メ
ι(LK嶺 :偽ム
R t ,,o,,

レ , lM
C
ヽ￨
Rl、 tt 4
prr".( I
t.
ι)
で
イ
1夕 ).
´ 122 0°
｀ ●

′
2220以
￨

12

7 2 "r｀ F [ レ
t
■， ●

R3,● ( セ ,′
ロ年 く P,
′

0つ
ハ ′ ´ ￨ ′2
ν

12,
・
じ ．
 
    Ｃ    

S■1
．
 

C ト P争
 

2,
 

︱      ヽ
 
 
 
 
ｌ

2
′

￨,,
に,
R,Ooよ
`
￨￨
〔 P2 l つ ,
t

eマ . staペ 142曖
畔 ^ ´, P′
ヽ
F kOst´｀

↓P3秒 裁
・ 牛
(`'・ h)
s2′ S3 ,Sf CL'r )

ｒ、
ｏ        ４

p“ ・
￨｀ ,

↑ヽチ ht'if!hL
０ レ

2
●

(Rt ,",*l r" *rf )

2

t 2 ヽ

`
ク ` ゝ11ト
「

s〕 ′Sダ

,3
“

42
￨

,S
3 / tL ¨ヽ
．  ′

'(t
？
３●

_1,).
Y.￨
︱︱

力■ ，

!年
、
一

1 nt,10
ク
:￨′
'1.Pzrfft
(h.&ri)
4 0
ヽ ヽ ―
￨.,
子 ゝ 02 り
 ２
，
Sy
てヽ
〉 喘t′ r、 t ,十 fィ

´ 
. ●

ヽ
一
`千 `
一
︶，    ３

( ev w",r' d 2 -*1" trc )

ヽ
Ｚ

, . ・ .

‐
. . . イ: .

・
′ ,. ■ .
´５ ‘ ．

_
1 (* ) 0 2 3 ● メ
ぐ

,メ
`ト )
 ｒ   ４   ３     ２

Pl ゝ′
●′

ヽ
，

    ヽ
・
 
．
ヽ     ３     ２

′
/
ノ
´ ′
´
●
´ .1 `・
/ 1_
/
(ヽ / ■

)
/ ∠
・―_`―
Y｀
●―
｀ ｀ n tf")〔 23V"
,
0 : ぉ =
ιっ
´
t
υ
k■ 〉 `,``
P2 ｀年
Ｌ・

,
‘    ︶   ３     ２

1
，

    ヽ   コ   ー
一
      ヽ

′― >
1-
/ ′
,

′ /
ヽ
′ `

● 2
3 .や ゞ ( `￨ `
し
r rr,. l(.1 t
ヽset′ ヽ

∫
〔′
`
 ヽo、 1

,224に 隔Ь
l
。けf聞 h,k=10腋 Ⅲ●ピIC)
P。

1)
ヽ

2リ

3)
ル

蹴嘴
P蹴
れPlole
1磁 ∝ ftc ktt И
颯 a ur"l l*wl
￨
,

い el、 ●1も

`ひ
`,
′
'

r
,

・´
^011。 7り iれ
″ ″´ ,1, ‖ ど . ´
ヽ
`

滋
n)Vq"A ,t,t) ,!"lh-うtC heり に
fば ″`で
￨し .

I Yro t ret
rar^-fi ド sPa∝ Ma'い aイ

P?r^tt'ut r€d lihtr ( Carフ で∫ ) akレ hイ l夕出

ヽ
ヽ
｀ ＼ ＼、＼ ＼ヽ ヽ＼

ヽ ヽ ヽ
ヽ

ヽ    ・・
出 小ム

       ヽ
￨―

●   
︲
ヽ

、
●  ︱

ヽ       ヽ
￨

 
 
ヽ
￨

´
´´
´ ′
´
ン

`
∫
ヽ  ヽ ＼

′
ヽ
_^ ― ヽ ・ .‐

ヽ｀ ヽ
ヽヽ
ヽ
￨ ヽ
ヽヽ

｀ ・ ヽ ●●
ヽ ヽ ヽ
ヽ、
ヽ
ヽ

・ ヽヽ

ヽヽ
ヽ●

ヽヽ

ヽ
ヽ

.
ヽヽ
ヽ
＼
ヽ
ヽ ＼ ゝ ＼ ヽ｀
 ′

ヽo.2
2022 41RI)oι 7
ζヽ鈴o"tぞ P,ulれ に いdヽ 断ξ′
a仏

ツο″
ιい力で , ` ′

6)Lh"l e,c供 ん oF。 レJ II卜 く T,に いお

武弘
″格〔J,端.c(セ ロ で
)=>4, "ハ ￨,ぃ
￨

tle n'tl,* t,
_
n
,″
´

￨′ eocL verle\ l,r lt ″九

′‐■
1 1ι γい
` 、
￨

n.( p,;t\ ヘ
,

￨ sa I
￨
,

ゞo ,“ι ″
jaれ μ
ι n*e 4^e
″
ヽ ´
r :も nμ
8, ●/ Mdiサ γ
` . `l
,.,“で しぃ
`

￨フ

り
`

″
●
_

“ k/
´

Sh 。 ′
弓 “
′ ′
●
C
″
欧
● ′ 7
“
`か

／
ヽ
ヽ
ty
_ヽ
・
ヽ

／
,, ‐ さ

．
．  
= ヽ

 
●●

 
 
 
 
 
 
／
///
・
１
ヽ  ´．
●

り く ｏ‘
′ / /
′

●
´
′ ヽ ´

//
′ ′

/
´一
一  一

●′,
．

′ ′ ―
ヽ

企
―
く
'

‐■■
・ ヽヽ
´に
ヽ  ‘
ヽ ０

ヽヽ

ゆ
´′
´
´
′

′
′ /
． ノ

′ ●,
′            ′             ′

/
′
ヽ
／

● ●■
／
 

////
′

/ / /
・ ´ ′
 ，
■
∼

′,
ゝ′
「
,・
′●

2
3
3

μ°
″"ヽ
,
■

ヽ
ヽ

′
0

‐ ´ ´ .

″
,ぃ よ 颯
′
'マ
ス ・
x)→ ο→のく>)a>つで> ― ・ ・
´ _フ ア
_´ _′
い

′ ￨
′̀ /「 /
￨
イ

′
う
ι
I
12
,
ヽ

〉 ￨ ￨

-0
´ ｀
´ ヽ ´

(3レ ,(Gl θ′
今ι)
￨

)()) r AP Pr"L " 1 it',t "- fuo 2

-\ S,,'.v th? res,lt,,t rnJ '.hl f"tl^

6) lor,*d exl hoF o"d 0o,r-fs t,y ow lr;/6e I 'l^?
cl*tct l;rfasl ) > na
"t
t:,< $e n,nlt f, eocL veylex f,r lt t\
(

● )
P1 リ
ヽ
o l;"" ;, o/t o
)v,\ ;a^
t) 87 ".lr",t i, /aL r^t,
11
E

I
ヽ,

′
● e ,-t ee hsk
′
^
A

／ ／
′
′
′
/
′

ヽ
︱ ︲ ，︱
t'
，
︶

ヽ ′
ヽ^′
._′
￨
´
■
ヽ

a
一  ︱

′
\)
・
′

″
‐
● ‐
●

■
口
ヽ
"

ヽ
／ ′
＾
︱

一  ●  一
ー

ヽ ０
ヽ 、，

￨
´

／／

`′
ノ
 ヽ

,
5
″ て
,.

△ ヽ

` (
) ￨ハ 12■
・

￨ 0 ´′'
→

・
1ヽ
′リ ー 、

0
ヽ
 
 
 
 
．

_ `

●()β し
^
の← 」

´
π:"5 F.メ んt
‘ ●

Sで ′晰
ι
ι

ていC〉→
ヽ

5 CO→ ζ), √ヽつ ´

´

■2)→ `10 で >'の→ ‐

―
′´
ク
′

‐ヽ
´ ヽ
︱  ´

ぅの 4И L
_

―
 stLrdentlD Major

I .Dive into sub convex rcgions ylung the idea of exact ceL decompositlon method
(ve.ti..l inet
2-Assign the nurnber to ea.h middle point ofthe bolndary ofthe convex regiofsfrom (aft (0)to goal.

sta rt
´´
外 「1 ￨′
つ

ヽ
．

ヽ
￨●

′ ●０  
︱

,
カ
ー
・

ノ22
●
．

/
′ ●
￨
ン●
￨ゞ
` ´
´
“ 7‐ ,′

`t ′
0
Ａ，
”ス

0/ ノ ￨
/
7・

1
￨

ヽ ′
タリ
t,

ι｀
(1) ● (1, 堂
―
σ つも 3
γや
イ
,

` υ)
/
′ ヾ
ν´
`
2
沌γ
●

―
]
ヽ
ン① 央
2・

凛
′
ヽ
, ′
1 ヽ
〔リ 2
ノ´っ
,
ヽ  

′
″ 2
_´ ´
，

r ′
喝
2 ′ つ
t

い
，

'l
ノク
′︺ ●

` と
ヽ 3
︱ 〓

そ ￨
り

´
︰

/
● ・

1
︱

“
ノ
ヽ

f,.I tlafer't frlh 15

GFじX)ο イ ン のく
)GX'て )で〕― ´。
o‐ ′′ 2 , 、
`は
 イ73計 'S
ハルa〆 α ,ヽ ヽ1,

22 レ ￨_0 ―て(ヽ
2ヽ
1'Dl ′

0,O

４︶
ヽ １．
︲
１ ﹂

Ь
tlrow Vornn", d,n

´
ン
● ,
´
´ ヽ
ヽ
´
ν
‐ ￨
ワ (_3ノ
●

/
・
一

―
・
ヽ

―
―
1,
Ｃ、

●
ヽ

ヽ
′

ヽ
ヽ、
ヽ
い
′た

ｌｄ

①
♭
い

③ヤ《
①、
∝
ヽ４ ｕ
ヽ

)
ｈ
０

,ヽ
」 ヽ
フ r∝w
・  ノ
、

い ´

① Q.ν ￨
tl´
' ●

'
"
′

、1
ヽ′
￨′

ρ
ヽ

/
´２ 一
ヽ ・ン
く

― ■
 ■¨
沖n?￨。 し] -(. ) ￨■ ￨￨

´ ∼
´

／
Ｌ レ  ．

ヽ
ヽ２ ︶
わ０
＞
ｒ

‘ヽ

．
(3)ν いW

１
︶／
Vrro r',", Jt 十

′
●

一
ヽ

‘ゝ
ヽ  ヽ
θ
＞
, ｀
千
″

0
■

●●
ι ノ
´■ ■

,
／ヽ
ヽ    一  ´
ヽ ︱︱′′

ヽ
0
υ
︱が
げい

w Q ,① の00θ
´

。
仄
ヽイ ヽ

Sへ
ヽ一

´ ′o′no.

"￨ ^丈 "、 Clarレ

´ヽうじ
ヽ り／
〆 ヽ、
ｒ ヽ
ヽ ′′
´Ｓ
ヽ

● ￨

(ワ
￨

/
￨
・

．
一

,￨ ., ⅢⅢ _
Ⅲ ■ . ￨ ￨
︺ヾ ミ ゞ 、 ■ξ り
Ｏ① 〓００① ﹂０い﹂００″
ＣＤ⊂ ﹂①Ｏ﹂０ ０ 〓︺ ＯＣ一
〇ＯＣ ＞０
り                     ヽ
①

コШ
つ一
′      一
０〓∽ 〓ＱＯ﹂０ ﹂０一
Ｃ ∞ ≧′ コШ一⊆①〓＞＞
∞
L_

一      一
Ｌ
＋  
Ｉ ●Ｌ ヽ●
しヽや‘・ヽし 、ｃ■
︲︱
こ
一︶０ 一     
Ｌヽ
,t」
C

■
ヽ‘

●
マ
ゝ、づ︸
／
イ一
一

ヽ
_

●
ヽ
Ш

●. 、

●
´ ‘‘ ｒ ・

‐
●
∼
Ｌ どヽ

―
′U

●●●●●

´
一き ●

:
●０
￨
ヽｏ
ヽ

ぬ
‐
● ＝ロ

0
￨ヽ
︶
●●

●´ ●
， ご︶
ｉヽ震だ
.￨

口.
●●
●[〕
.`1
●
“
‐ヽ
, ︱ ︱

‐て,ヽ
ヽヽ
︱

‐‐●

ヾ
ヽ

`
一
ヽ一卜

￨
ヽ■ ヽヽ
，
“●ｓ﹂¨

ヽ
ミ、・
ヽ ヽＳ

ヽ
ヾ ５︺ コ ー

t
ヽ‐
、′
一ｏ 一 ．
﹄
く︱ ■“
τ‘６
ヽ

￨
ヽ

`
´ ´
・ ´ｑ

‐
`
一〇 Ｃ ﹂０ 〓Ｑ∞﹂０ ﹂①一
コШ一〓Ｑ∞﹂０ 〓００① 〓００〓︵
︺
′
，′
Ｔ︵
倒ＯＯＮ︶∞Ｏｏ﹂﹂∝＜
 二‘工 、‘‘、ヽ＾Ｔ
′
‘手
,
ぃ×＾
戯
―
″

ヽ ︲
ρ

、フ
′
でいｔ一 一 ヨ ‘ｒ ●・

｀ヽ
︶‘﹄´つ
ヾヘミく ‘二 一 ヽ〓 ‘こ ∵‘ 傘 ゝし、
、ィ一
．
．，
ヽ
え“サ
。・べ
ヽい
に

0
ヽ ヽ ︺
・
Ｘ ヽ︶ Ｎ マ
ベ い ︶ゝヽ
´0

r
︺も
ミ一
ヽ
ヽ
ゝ．
  ヽ︶ ヽいへヽ ︵

,
りヽヽ ０

nヽ
′ヽ

チ
・ I,
∼

´
￨

∼
‘工 ︵
〓 こｙ ・
“ 午 ‘◎ ０ ↑一

③
■‘エ ハ
二一 “
ヽ一
︶＾ ＋ κ Ｆｆ■ ヽ ェη
一
ヽハ ＾＾い ∫ヽ
‘ヽ
ヽ
n
′

ヽ２バ一
′り
´

.十
´し
,

三〇

yい o
１ェｎ
ｓ■ 一

O`￨
ヽ`
,ヽ

′・
∫

′ツ.
●´
■５。こ′
・

‐
一い
ｄゝ＾
″＋ ゛ ′１２■ ‘
ヽ
L‐

十
∼

′^
∽ヽゝ０ ど．、 〓 ｃ メ

rl
一 ヽゝ ＾ ≠
，

ヽ
ヽ
●ヽ
￨'

tヽ

ヽ
ゝ

０   ヽ ・ ´  １ 一ヽ
︱  ・ ´
‘． ︲´一
ヽ  一 一

￢
○
`
かシ
´″ヽ
︺ＯＣ ﹂０ 〓Ｑ●﹂Ｏ ＣＯ〓ＴＣＣＴ一〓ＱＯ﹂０ こ００① 〓００〓︵
︱ ．︲・︱ ︱ ． ．●           ︱ ︱
∪
︱︲︱
・                                              ●           
，
︵０四〇銀︶ΦＯｏ﹂Ｌに く
∞，
 ■ ＯＦ 〓 ＣＯ●０ す ０ ０ こ一 〓 Ｏ Ｃの ︵Ｎ
〓 いＯ Ｘ一
Ｏｏ ｏ 〓¨ ゝＯ Ｃ〓 ｏ 〓∽
一∽〓 ´ＣＯＯｏ一 “Ｃ κ 押尽 ｔ 輩 選 ｅ ｏ朧 ＝ ︵Ｎ
０ ‘Ｏｏ﹂Ｏ つｏ一 つ ｏ二¨〓ｏ﹂０ ︵す
ｏｏ﹂一 〓 埋 掏０ヽいヽ Ｅ 毎 Ю某 拘縣 騰 ︶工К ヽ騨 蓋 ｅ 朧 ＝ ︵一
¨
Ｅ ０５９ Ｌ ●‐‐‐‐‐‐
︵脚娯
●

`￨.
エ
C:ヽ

中
ヽ

￨｀
‐
O O

e9o
つ

U01●

￨
OO

●
崎
っ
0

い
う

3e

く ∞

中
0

=
ふ 0

‐
Cヽ
し,も

ヽ
・

t.
一  合
て

●●
￨

凸
￨ ●

″
`_L

0
●
●

、ヽ
′ ―‐‐ヽ

て
× 短壕ッ
〓 ヽ●
?
,
C
■■
ー

=♪
‐‐‐‐‐■●
￨

___」

山
口
＼
ヽ  ●
・
﹁＝ コ、１ー ︲
  ヨコ
﹁︱︱ ︱︱コ
ＦＬ Ｌ
   
  一  一 ヽ

‐1
    
¨
¨ 、
´一
， ヽ    ´

￨●
一  
Ｆ●■１ ︱
 
，，
ユＫヽ
・螂 彗 Ｃヽいヽ Ｅ 〓αｏ﹂
ぃっｏも９一
０﹂ 〓ｃｏｏ。τｏｃく
ｏち一
， ∞人倒ＮＯ銀︶ΦＯｏ﹂
Ｌにく
1ユK〕 b5of Advanced Robotics 2022
StudentlD Name

Make initial reduced visibility graph

1) connect two vertices for all the combinations excluding
the path of penetrating obstacle
2) write the number of qach edge in the order of combination
3) find the reduced edges by marking the circle for the number

START
E

OAL

total number of initial edges:? I Y

“
ψ
一︸

′ ′
一

total number of reduced edges:? rP

，

１
一
リ
 b5-2 ofAdvanced Robotics 2022
り
StudentlD Name

Starting with the reduced visibility graph

1)write the cost of edge (rounded distance with cm)
2) write the cost for node in the order of width first
3) find the minimum cost path by Daijkstra method

[41)
A
ヽ
ヽ
ヽ
〃
、
、

START ｀ __′/
(12♪
`κ
′
σ ｌ ｌ′

/
じ
′
E
7

/
つ
0
D B ´
、 、
ヽ︱

ヽ
“
/′

しノ
′
︲′

、 (
ti ヽ

ι ＶЧ，
ヽ ヽ
、

Ｂ ヽ

又
GOAL
:

ф)
ヽ ヽ
︱

―F 3)
′ ヽヽ

(ノ
つ

『
ヽ1ミ
′ヽ
い    ´

・ ＼ )

(_｀ ち) ＼

S→ 八 っ ざ → 694た
VVhich one is the shortest path and
how the traveling distance (eac d ge ls
rounded cm)? ′ (ミ
: :
' :
 Name Major
1

r、

W以 Simple Algorithm (Left Rule)

●
GOAL

‐
ヽl
r2?Uヽ
Fth
SaV'｀

tふ1121)

START

tr hit point
indicate O leave point
 ヽ

Natne Ⅳl可 Or
口７ ︲

BUGI Algorithm (Left Rule)

tゎ
Y`
メ脚ι
, キi
Ч
徹 1外

/. ∂Й七 ノ
_∠1lori`::i',ヽ
(ぐ
/
・■
′

START
，
０１ １

)
―

・ ■
・

lndlcate
・ tr hit point
O leave Point
７、
・    ン

°″∫丁
,

´ヽ t、

Sて lο
″
準 格折ム "ρ
1メ

△ '
 Name Major

BUG2 Algorithm (Left Rule)

GOAL

START

indicatc
tr hit point
O leave point
 Name Major

tangent BUG Alg orithm Let Rule)

with sensor range 5R=15mm

るヾ
●４。

*
︱′

STAR|

indtate
tr hit point
O leave Point
X localminima Point
i'd; raTr Pn'
 マ人 ″´
ヽ １
し
、3 イ ′
[{ +0.) a"s,

︲ ︲ ︲ ︲ ′ ′
￨フ ￨も 2`)22リ
1,r。 ￨[

０
ル
ヽ

︱
A [/t0,)si"0,

オー
‘
no+
dl ,ru -t[e

，
c-01じ:仏cF´ (,,f )
It.t 0,-0. SPaCC
o(0,( l8o"
(-c*,1,b ぃr〆 :oぃ o:0, 13 o p,^,,
7に ノ
ilbs/ut(t ( lia-a" )
."hX

,-j 0 P
1
1il
2

ヽ
ヽ
z
{
/´

-lo -zo 3o (。 ro fo

ヘーノ
l0 B (rn t∼ ,

0.
(rr")
3o
I
￨

十
7.1

10

fit
00 10" l8o'
 人
、ノ
／

ヽ
イ
c,oo, '

′＝

︱ ︲ ︲ ︲ ︲ ′
{{t0,)

■‘
＾′ ″

´
イ

ー ーー
ヽげ

′′

Ｍで
〓
n',bl<*-)

ー ー
ル 
ヽ ︱イ Ｊ
A

ヽ
flt0.,)s,u0,

ー
flo+
s t,. o u^l 孔くC― obt(lfr (t'f )
o!9,( l&0"
rh 0,-0. sro.e
′ ′ ´

//ん °
/ / /
/
010,!J0-.-
M
r-tu,f,tr y6l"- syur
′

, CIbsfrンヽし'｀, i+ ( A=n-"
"ht .qr-f. )
6,
′ ノ
0 o
/-1 2 ヽ ‐
′
´t
′
多 ヽ

一 ¨
ゝ、
ラ _

一
   
ア

_-^ -20 -rc 01 :o 4o 5-o 6e

ヽ
(m

︱り
´
,0 つ

0.
力´ ￨
{r',")
′

., ヽ
,,\
/‐
′
ヽ

"t
´
    ヽ  一
‘・
︱
ＶＩ 、 １
、 ︲

/ At
４ １、

igo1o ttr 1pU. l80'

Ｐ     ′ ′

l\ t
，       ｉ ′﹃

b, "lt
Ｃ
●ｒ

00
，
ヽ
″ ヽ´
′
 r`ヽ
￢

一  一
一  ヽ
．  ヽ
ヽ
ギ

‖
■Ч

7 ■
｀

ツ
「

３︼  ａ
P

ｏ Ｏ
ｔ８﹁
εヨＯ﹁ ヽ
―

∵
″

ｒむ ＋ く
︵
︶
＝ヽ
ヽ

、
‐

ゝイ            
Ｄ
￨￨
ヽ

一ＣＯ ｃＥＯ Ｆミ ００﹂Ｏい ０ ⊆一

つ ００ 〇 三一 ０一
Ｏ ｆ Ｏ︺
レ

Φぅ０﹂
Ｏく００﹂
０﹂一
Ｃ一〇一Φ〓一０一
０一コ０一
〇０︶︵
寸︶
ヽ
.

一ロゥ︻ Ｃ一ゥリ﹁ｏ引 Ы︻
〓●︻ 〓もヽ に
′
﹂お Ｅ ＣＧ一
Ｘ一 〇〇Ｏｏ ﹁ Ｏ ｆ一 ３ ０ ｆ∽ ︵∞ ︶
お一 ｏ＞ Ｑ〓①三︸①︺
〇〇一 ｏ一⊃０一
ｏＯ ︵
倒︶
０ ﹂Ｏ ＥＯ 〓０⊂コい ０ 三︸
∽０ ´
∩︶．
＞．Ｘ︶Ｑ ＣＯ〓０一
Ｃ〇一
﹂Ｏヽ のＯＱ Ｑ一
ＣＯコ一 一０三一０一
０一⊃０一
０︵︶´▼︶
．
Ｅ ﹂０ 一ｏＯｏ﹂ ﹂ ○ ∩ ︲∞ σ⊂一
３ ｏ 〓ｏ﹂ ｏ ①＞ｏ 三 ①≧ ′
Ｆ〓ｏ 一０■０﹄いｏ いＵ〓”” ︼ ”
にｏＣ¨ ︵ 〓０一
ＮＮＯＮ︶Ｎ ″ Ｏｏ﹄﹂
 [r.b{- z
tl)
マ
/ヽ
イ

４θ 鮮︶  ・χ ・ ４ “ ︶ ．Ｏ Ｖ
_
‐
υ′ 十‰)¨ 11† ダ、 oar ($,t$, )
一一
      ヽヽ

プ(イ 12),)仄 Ъt十 ノιsi'.. ($, + tr )

T
のク
ａ﹁

イ
ο

。 ４ ，′ ｑ Ｄ
． ４

ヽ
ＳＤ

ヶ ち1でぁ)ィ +3

４Ｄ
９
Й

．， ．， ′
′ ・ｏも
、， ノ

一¨    一一    ヽ＾

′             ′
´
２

レρ :′
2 (′ tイ t)2) i麟

，
 
．９

ヽ ︱′′
(
‐′か(1′ で
ｑＤ

0
 

り卜 3')(5,十
／

Cι`‐ 132)
●
01o ″■
０
ａ

{‐
 

．０
．９
 

`
■
′Ｄ
 
０

１０
ｒ

(p) ′
響・ 亀1プ ヽ 九6tも ) 姉 t -1,iu,0,''L

κり０
′ ＝︱

1、 )S・
?
し
。も ィ
′ sirly '!,vo (1,rlt1

フレ
((1う (イ /2い 33)
(′
0̀
0 ヤ
￨
“
′
￨

丁f■ )υ う

´
し ャ ,丁 rν ′移
)3「ヽ。:.h
て ￨ヮ

′
dイ 'd′r'J`タ ノγ
リ
り夜
,(,.イ
′ /ヽ
0
マlη ,シ 1)レ 々
1∠ +-1 Tlt : :
`,′ ν
｀1 ノ
ずマー^ Jlr7″7.。
「
仰
_1 ` I
ι ´￨ ￨ ￨
t″
イ
ヒ
7へ 」
Ч r,J化 ,ど γ
、
ヽ

↓
ヽ
耐
プ

Ｌσ

ぃ
グルo(
dV Z lJ彬
ｌυｒ ｌ

／″

つ‐

n(= t
′
嗽

レヽ
d )7
識 ＼
ヽ１ ７ ＝ Ｖ

‐ (1
′
´ワ

こ= J ら/ 例
´ 」
(いu , ヽ

ラアリ ￨
r
κ
ヽ
CI
0
こ η
‐
■ ヽ

ヽ に
︲／

{P,utllX,
ｒレ“ Ｍ

{,-t5)rt{,4 "rt({,.1) 11,*1,")c.ofli ) '(

υ

わら ,い 3
1

p.,,'0,t
2`れ 、 イ
-′ ι 3F) 11′
?, )
 ↑ ＯＱゝ︺Ｏ＞一 〇〇ｆ≧ｆ一ＣＯ﹂﹂
﹂０ 一
‐

〓 ｏ三一、
﹂０﹂ ＣＯ〓０コ００ ００コＯＣ﹂０⊆一 ５０Ｔ一´銀︶
０＞０一
Φ ＯＯ Ｅ 〇 一
ヽ︱ 
ｍ
︱
一
ヽ

¨
︵Ｏ Ｏ︶＝︵ゝ．
Ｘ︶ ・
■︰︰︰︲︰︰︲
▼＝ Ａ︺   〇＝ 〇Ｑ ΦＯＣＯ﹂①一０﹂
一Ｃ一
▼＝ Ｖ一＝ ＶＯ ︺＝＞

い ｀0
→
●ヽ
´ｒ

%
Ｅ ︶＞ ゝ〓〇〇一
∽ス Ｏ①のヽ ①＞ ＯＣ∞

い、0
︵
ｔ 。Ｃ∞ 。Ｅ ①①ギ
∽Ｃ８ ︶ Юヾ ＝ Ｏ ①一 ∽ ①﹂①〓≧
Ｏ
， 鎚ゝ
０００︶︵
︵ ︶＝Ｖ一＝ＶＯ︶
︵
︵じ① バじゝバじＸ︶ ゝ﹂０ち ①一
ｏ﹂︸①Ｃ一①︺
ｏ一 ｏＯ ︵︶︶
⊃ｏ一
５％ヽ
０ ︵〇ぃ ¨ ︵ メぃ ０三⊂一
の一
〇＝︺ＯＥ 軍 ︸ ＝ ① Ｘ︶ の一①一
Ｏ Ｏ︶ ０一
︼ “ｒ ”
ＥｏＥ¨
︺ｏ一ｏ﹄●〓〓︰﹄●Ｕい。 ぃｕ〓 暉Ｆ 一０﹄﹂
〓０一
鰯倒 ｏい
 frlv* 6′
`ズ
′ク́
ncs r2o22´ )
bι /ζ ン
bノ ′

L ′
メ 幅
l'retn 1で
・
ah‐

lw*'dx= ltt.) , lSi,^x'A^ = Srtt) one ‐

′′
′ `レ
/ `
′
｀ノ ヽ
´´
ヽ

4クrく
レ
リ
αレノ o[co `ん `し

f ctxlAt = CCtu) t I Sft)tt)(" Slrx) arc alco t)^(*oU*u)

[ila.;*yt;+ oaqnfr be i',f"yntel ustT e[aafiry. fluctr*" )
グ・
ち
‘θ

そ2́
φ′

° /・
′
Ｚ

″=ル
プ

φ=￠∫ う ヽ

ニ
′ノ
′
′ヽ

θ
` 0ノ
︱

ウ
`
rυ
ソ =し ′

ア・ ひι
′
いけ
″

｀
ヽ

`"焼 t
″

t ‐,, ‐1
そ

りz,[Z′ ノご々私
´
ヽ ′ ′

2い が
一′

。Zι r2ノ ″z
イ ′

り
/ヽ 2ノ
ノ0

'2/7(ε のィのノ2/〆
つ一 ,)￨ /。

‐
ｒヽ
デ  ハＶ

“
ｆ ｌ ノ
＾‘

′ ``漁
レー

ダ
、 １ ，ノ

´・
ー
ー イ′ ′

ι 〉
__上
オ
n】 Iぞ =[7メ
り ￨『 Sr2)`イ ?
「 2′ 5'イ
‐
l:て
′υ
2/′ ´
￨
￨

´●
ιてs/4ノ づんノ
リ 9t (ssrg s.t1o,)
2´
￨
ヽ
′
′
/ ノ
f

に ク
つ〃′ル ●γ 多 ′ヽ & .

/
ぐ
a/′｀ ヽ J
ノ
dri/c c"sg )
\rg'tv ( fr"<
′ ヽ

に
2リ
0
'hJ
uf( cft )
へ 、
ｒ

( rear
４ Ｕ ヽ′

dtAw c
，

',rlrJ-
ン

チ
∫てた′つπ ′力θ
γ崚
／
て

∝ S 91″
慶ル

へ♪
′ニ

′
ル

︱竹
９／
′
，
ハＮ

イ＾
´一
Ｖ