Self-driving car is a program, which implement PSO/GA/fuzzysystem algorithm by Java. GUI implement by using JavaFX.
Self driving car is automobile's trend nowadays, and this program use three algorithms to accomplish the task. In this program, you can load another map files, the auto will drive to destination by itself.
the button introduction is following:
- Open Map: Only you want to load another map, clicked the button. It willopen maps file, and draw the maps.
- Open Solution: Open solution file(4D, 6D are accepted), auto will drive accroding to solution file and draw the driving path.
- Run :
CAUTION: before clicked this button, you need to train model or load the solution. Auto will drive by itself. You can cliked it many time for one map, but it will be broken when you click button during car is driving. - Train: open training dataset, and train the RBFN (if use RBFN)
- Load Traing Result: load the RBFN parameter.
- Save training Result: save the RBFN parameter.
0,0,90
18,40
30,37
-6,-3
-6,22
18,22
18,50
30,50
30,10
6,10
6,-3
-6,-3
first line is beginning（x,y,φdegree）
2,3 line indicate destination position
line 2 is left-higher point（x,y） in the region
line 3 is right-lower corner（x,y） in the region
the point (x,y) of rail will show in line 4 to the last line
last line will be same to line 4 to form a enclourse rail
the beginning right-higher corner is (-6,-3);left-lower corner is (6,-3)
beginning is (-6,0) -> (6,0)
Choose the self driving algorithms, and download the folder, then use Java compiler to compile.
I used fuzzy system to decide the turning degree and direction of the auto, and implement 4 different algorithms to decide the turning degree by input the three detect distance of right, front, left.
Design 5 fuzzy rule to decide the decide the degree.
Used RBFN to decide the turning degree, input 3 detect distance, output the turining degree, and train the RBFN by using genetic algorithm.
Used RBFN to decide the turning degree, input 3 detect distance, output the turining degree, and train the RBFN by using PSO algorithm.
Used RBFN to decide the turning degree, input 3 detect distance, output the turining degree, and train the RBFN by using PSO-GN algorithm.
NCU CSIE | 105082015 | 陳昱瑋