1,a) 2,b) k=1 k
1.
[1]
2015
6
1
[2]
1
Graduate School of Systems Information Science, Future University Hakodate 2
School of Systems Information Science, Future University Hakodate a) [email protected] b) [email protected]
[4]
[5]
[6]
[7][8]
[7]
[8]
1
1
2.
2.1
[9][10][11][12]
[13][14][15]
[9][10]
[11]
[12]
[13][14][15]
[14]
[15]
2.2
2.2.1
[12]
[10][13][14]
[12]
[10][14]
2.2.2
[10][13][14][15][16]
[10][14]
[13][15]
[16]
3.
[7][8]
3.1
1
1
2
1
2
2
3.2
3
挙動データ収集 挙動データ収集 特徴量抽出 特徴量抽出 学習データDB 挙動分類 回避挙動検出障害物 ：登録ユーザ ：検出ユーザ 学習フェーズ 検出フェーズ 正解ラベル 特徴量 長距離走行時の 挙動データ 回避区間候補を個別抽出した 挙動データ 挙動データ収集 回避区間候補 抽出 障害物情報 障害物回避挙動 検出GPS
100Hz GPS
5Hz
GPS
4
1Hz
GPS
GPS
3
4
3.3
3.2
5
(a) (b) (c)
3.3.1
5 (b)
5(c)
6
6
最大値
最小値
平均値
(a)
(b)
(c)
5
7
7
(1) (3)
7
!"##
$%_'$(= +
$%_'$(– #
$%(.
$%_'$()
(1)
!"##
$%_'01= +
$%_'01– #
$%(.
$%_'01)
(2)
.
$%_012= 3.
$%_'01− .
$%_'$(3
(3)
Diffaz_max
faz(x)
Taz_max
yaz_max
faz(Taz_max)
Diffaz_min
Taz_min
yaz_min
faz(Taz_min)
Diffaz_max
Diffaz_min
Taz_int
Taz_max
Taz_min
8
8
3.3.2
5(b)
fsp(x)
Tsp_max
ysp_max
fsp(Tsp_max)
Diffsp_max
Tsp_min
ysp_min
fsp(Tsp_min)
Diffsp_min
spex
Diffacc/dec_sp
(4) (6)
!"##
56_'$(= +
56_'$(– #
56(.
56_'$()
(4)
!"##
56_'01= +56_'01 – #
56(.56_'01)(5)
!"##
56_$77/9:7= ;<
:(_0– ;<
:(_0=>(6)
9
fsp(x)
Diffsp_max
Diffsp_min
Diffsp_acc/dec
3.4
SVM
4.
3
4 0 4 12 4 86 6 近似直線(faz(x)) 最大値(yaz_max) 最小値(yaz_min) 所要時間(Taz_int) Taz_max Taz_min Diffaz_min Diffaz_max [ ] [sec] 最大値(ysp_max) Diffsp_max 最小値(ysp_min) Diffsp_min 極値(spex_i-1) 極値(spex_i)fsp
(x)
Diffsp_acc/dec 極値(spex_i+1)Tsp_max
Tsp_min
4.1
4.2
4.2.1
Z
L
k = 1
k
1.5
10
10 Z
4.2.2
L
Lmode
Lmode
Lmode
Z
Lmode
Cl
Lconst
t
t
2t
t
1
11
11
-. t t ِ٦ؙٔحس騃ꨄ 80 80 Cl 630 4125.
[7][8]
5.1
1
3.2
600m
800m
15
18
5.2
5.1
Lconst = ClLmode
Cl
Cl = 1.5, 2.0, 2.5
5.2.1
1
1
(a) Cl = 1.5
13 4 17 2 15 16 1 17 2 18 29 5 34 4 33(b) Cl = 2.0
13 3 16 2 15 16 0 16 2 18 29 3 32 4 33(c) Cl = 2.5
10 4 14 5 15 16 0 16 2 18 26 4 30 7 338
Cl
Cl = 2.5
Cl = 1.5
Cl = 2.0
Cl = 2.0
5.2.2
Cl
Z
Z
5.3
5.2
Cl =
2.0
SVM
90
t Ct
4m/s
30m
7
6
2
3
2
t 2t 1 3 6 2 2.5 5 3 2 45.3.1
3
3
(a)
1 t = 3
4 11 0 15 16 2 0 18 20 13 0 33(a)
2 t = 2.5
4 11 2 15 16 2 0 18 20 13 2 33(a)
3 t = 2
2 13 2 15 16 2 0 18 18 15 2 335 6
2
8
1
3
5.3.2
1 2 3
1
6
2 3
3.3.1
5.2.2
12
30
10
12
5FTU 5SBJOJOH "WPJEBODF6.
[7]
[8]
1
1
8
6
JSPS
JP17K00128
[1] 27 http://www.e-stat.go.jp/SG1/estat/Pdfdl.do?sinfid=00003140011 2 (accessed 2017-1-14) [2] https://www.mlit.go.jp/common/001085121.pdf (accessed 2017-10-23) [3] http://www.bicyclemap.net/ (accessed 2017-10-23) [4] J. Burke, D. Estrin, M.Hansen, A Parker, N. Ramanathan,S.Reddy and M. B. Srivastava, “Participatory Sensing,” World-Sensor-Web (WSW ’06) at SenSys ’06, pp.1-6 (2006). [5] “ ” 2014 pp. 966-974 (2014) [6] “ ” ° (MBL) Vol.72 No.19 pp. 1-8 (2014) [7] “ ” 2017 pp.83-90 (2017) [8] “ ”
° (ITS) Vol.71 No.6 pp.1-8 (2017) [9] “ ” (EMB) Vol.36 No.51 pp.1-6 (2015) [10] “ ”
(ITS) Vol.113 No.491 pp.1-6 (2014)
[11] A. Zhan, M. Chang, Y. Chen and A. Terzis, “Accurate Caloric Expenditure of Bicyclists Using Cellphones,” Proceedings of the 10th ACM Conference on Embedded Network Sensor Systems, pp.71–84 (2012). [12] “ ” ° (ITS) Vol.68 No.6 pp.1-9 (2017) [13] “SVM ” 2014 pp.44-52 (2014) [14] “ ”
(KBSE) Vol.113 No.475 pp.73-78 (2014) [15] “sBike ” Vol.53 No.2 pp.770-782 (2012). [16] “ ” Vol.136 No.3 pp.363-372 (2016).
