-
Notifications
You must be signed in to change notification settings - Fork 0
/
regression9.py
executable file
·134 lines (115 loc) · 5.32 KB
/
regression9.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Mon Sep 21 02:18:15 2020
@author: suhaib
"""
import math
import numpy as np
import pickle
line_dist = 25
num_points = 351 # This should be an odd number
mid_ind = int(num_points/2.0)
with open('reg8.4_data', 'rb') as f:
#8.3 is track A new
#8.2 is track B new
#8.4 is fixed track B new
nbrs_right, rightLane, nbrs_left, leftLane, midLane, center_nbrs = pickle.load(f)
def generate_rays(pos_x, pos_y, yaw):
angle = yaw
angle3 = angle
angle2 = angle+35
angle1 = angle+70
angle4 = angle-35
angle5 = angle-70
ep1_ray1_x = pos_x + line_dist*math.cos(angle1*math.pi/180)
ep1_ray1_y = pos_y + line_dist*math.sin(angle1*math.pi/180)
ep2_ray1_x = pos_x# - line_dist*math.cos(angle1*math.pi/180)
ep2_ray1_y = pos_y# - line_dist*math.sin(angle1*math.pi/180)
x1 = np.linspace(ep1_ray1_x, ep2_ray1_x, num_points)
y1 = np.linspace(ep1_ray1_y, ep2_ray1_y, num_points)
ep1_ray2_x = pos_x + line_dist*math.cos(angle2*math.pi/180)
ep1_ray2_y = pos_y + line_dist*math.sin(angle2*math.pi/180)
ep2_ray2_x = pos_x# - line_dist*math.cos(angle2*math.pi/180)
ep2_ray2_y = pos_y# - line_dist*math.sin(angle2*math.pi/180)
x2 = np.linspace(ep1_ray2_x, ep2_ray2_x, num_points)
y2 = np.linspace(ep1_ray2_y, ep2_ray2_y, num_points)
ep1_ray3_x = pos_x + line_dist*math.cos(angle3*math.pi/180)
ep1_ray3_y = pos_y + line_dist*math.sin(angle3*math.pi/180)
ep2_ray3_x = pos_x# - line_dist*math.cos(angle3*math.pi/180)
ep2_ray3_y = pos_y# - line_dist*math.sin(angle3*math.pi/180)
x3 = np.linspace(ep1_ray3_x, ep2_ray3_x, num_points)
y3 = np.linspace(ep1_ray3_y, ep2_ray3_y, num_points)
ep1_ray4_x = pos_x + line_dist*math.cos(angle4*math.pi/180)
ep1_ray4_y = pos_y + line_dist*math.sin(angle4*math.pi/180)
ep2_ray4_x = pos_x# - line_dist*math.cos(angle4*math.pi/180)
ep2_ray4_y = pos_y# - line_dist*math.sin(angle4*math.pi/180)
x4 = np.linspace(ep1_ray4_x, ep2_ray4_x, num_points)
y4 = np.linspace(ep1_ray4_y, ep2_ray4_y, num_points)
ep1_ray5_x = pos_x + line_dist*math.cos(angle5*math.pi/180)
ep1_ray5_y = pos_y + line_dist*math.sin(angle5*math.pi/180)
ep2_ray5_x = pos_x# - line_dist*math.cos(angle5*math.pi/180)
ep2_ray5_y = pos_y# - line_dist*math.sin(angle5*math.pi/180)
x5 = np.linspace(ep1_ray5_x, ep2_ray5_x, num_points)
y5 = np.linspace(ep1_ray5_y, ep2_ray5_y, num_points)
return np.array([[x1[i], y1[i]] for i in range(num_points)]),\
np.array([[x2[i], y2[i]] for i in range(num_points)]),\
np.array([[x3[i], y3[i]] for i in range(num_points)]),\
np.array([[x4[i], y4[i]] for i in range(num_points)]),\
np.array([[x5[i], y5[i]] for i in range(num_points)])
def find_intersect_dist(pos_x, pos_y, ray):
distances_right, _ = nbrs_right.kneighbors(ray)
distances_left, _ = nbrs_left.kneighbors(ray)
ind_r = np.argmin(distances_right)
ind_l = np.argmin(distances_left)
if min(distances_left[ind_l], distances_right[ind_r]) > 1.2 and classify_left(pos_x, pos_y) != classify_right(pos_x, pos_y):
return math.sqrt((ray[0, 0] - pos_x)**2 + (ray[0, 1] - pos_y)**2)
return max(math.sqrt((pos_x-ray[ind_r][0])**2+(pos_y-ray[ind_r][1])**2), math.sqrt((pos_x-ray[ind_l][0])**2+(pos_y-ray[ind_l][1])**2))
def find_intersect_point(pos_x, pos_y, ray):
distances_right, _ = nbrs_right.kneighbors(ray)
distances_left, _ = nbrs_left.kneighbors(ray)
ind_r = np.argmin(distances_right)
ind_l = np.argmin(distances_left)
if min(distances_left[ind_l], distances_right[ind_r]) > 1.2 and classify_left(pos_x, pos_y) != classify_right(pos_x, pos_y):
return ray[0, 0], ray[0, 1]
if math.sqrt((pos_x-ray[ind_r][0])**2+(pos_y-ray[ind_r][1])**2) > math.sqrt((pos_x-ray[ind_l][0])**2+(pos_y-ray[ind_l][1])**2):
return ray[ind_r][0], ray[ind_r][1]
return ray[ind_l][0], ray[ind_l][1]
def get_distance(test_x, test_y):
dist, indices = center_nbrs.kneighbors(np.array([[test_x, test_y]]))
index = indices[0][0]
nearestX = midLane[index][0]
nearestY = midLane[index][1]
deltaX = 0
deltaY = 0
if index > 2:
deltaX = nearestX-midLane[index-2][0]
deltaY = nearestY-midLane[index-2][1]
else:
deltaX = midLane[1][0]-midLane[0][0]
deltaY = midLane[1][1]-midLane[0][1]
otherDeltaX = test_x-nearestX
otherDeltaY = test_y-nearestY
cross_result = np.cross([deltaX, deltaY], [otherDeltaX, otherDeltaY])
return (cross_result/abs(cross_result))*dist[0][0]
def classify(test_x, test_y, train_x, train_y):
_, indices = nbrs_right.kneighbors(np.array([[test_x, test_y]]))
index = indices[0][0]
nearestX = train_x[index]
nearestY = train_y[index]
deltaX = 0
deltaY = 0
if index > 2:
deltaX = nearestX-train_x[index-2]
deltaY = nearestY-train_y[index-2]
else:
deltaX = train_x[2]-train_x[0]
deltaY = train_y[2]-train_y[0]
otherDeltaX = test_x-nearestX
otherDeltaY = test_y-nearestY
cross_result = np.cross([deltaX, deltaY], [otherDeltaX, otherDeltaY])
return cross_result > 0
def classify_left(test_x, test_y):
return classify(test_x, test_y, leftLane[:, 0], leftLane[:, 1])
def classify_right(test_x, test_y):
return classify(test_x, test_y, rightLane[:, 0], rightLane[:, 1])